Doctor of Philosophy IInn In the Faculty ofIn the …shodhganga.inflibnet.ac.in/bitstream/10603/69120/1/24.pdforal epithelial dysplasia as ‘risky’ and ‘non-risky’ using the

I

“VALIDATION OF A NEW CLASSIFICATION SYSTEM

FOR CATEGORIZING ORAL EPITHELIAL DYSPLASIA”

Thesis Submitted to

THE KLE ACADEMY OF HIGHER EDUCATION AND RESEARCH, BELGAUM

(KLE DEEMED UNIVERSITY)

[Declared as Deemed-to-be-University u/s 3 of the UGC Act, 1956 vide Govt. of India Notification No.F.9-19/2000-U.3 (A)]

(Accredited ’A’ Grade by NAAC & Placed in ‘A’ Category of MHRD by GoI)

For the award of the Degree of

Doctor of PhilosophyDoctor of PhilosophyDoctor of PhilosophyDoctor of Philosophy

In In In In the Faculty ofthe Faculty ofthe Faculty ofthe Faculty of Dentistry Dentistry Dentistry Dentistry

((((Oral PathologyOral PathologyOral PathologyOral Pathology & Microbiology& Microbiology& Microbiology& Microbiology))))

by

Dr Alka Kale (Registration No: KLEU/Ph.D/07-08/ IZUN07001)

Under the Guidance of Prof. Dr Prakash V. Patil M.D, Ph.D. Dean, Mayo Institute of Medical Sciences

Gadia, Barabanki, UP.

2014

II

KLE ACADEMY OF HIGHER EDUCATION AND RESEARCH,

(KLE DEEMED UNIVERSITY) BELGAUM



Declaration Declaration Declaration Declaration bybybyby tttthe Candidatehe Candidatehe Candidatehe Candidate I hereby declare that the thesis entitled “Validation of A New

Classification System for Categorizing Oral Epithelial Dysplasia” is a bonafide

and original research carried out by me under the guidance of

Dr Prakash V. Patil, Ex-Professor & Head of Pathology, JNMC, Belgaum and

presently Professor of Pathology and Dean of Mayo Institute of Medical Sciences,

Barabanki. The thesis or any part thereof has not formed the basis for the award of

any degree/fellowship or similar title of any candidate of any University.

Place : Belgaum

Date : Dr Alka Kale M.D.S.

KLE VK Institute of Dental Sciences

Belgaum.

III

KLE ACADEMY OF HIGHER EDUCATION AND RESEARCH




Certificate by the GuideCertificate by the GuideCertificate by the GuideCertificate by the Guide

This is to certify that the thesis titled “Validation of a New

Classification System for Categorizing Oral Epithelial Dysplasia” is a

bonafide research work done by Dr Alka Kale MDS

Place: Belgaum

Date :

Dr Prakash V. Patil MD, Ph.D. Professor of Pathology

Dean, Mayo Institute Medical Sciences

Gadia, Barabanki.

IV





CertificateCertificateCertificateCertificate bybybyby the Deanthe Deanthe Deanthe Dean

This is to certify that the thesis entitled “Validation of a New

Classification System for Categorizing Oral Epithelial Dysplasia” is a

bonafide and genuine research carried out by Dr Alka Kale M.D.S. under the

guidance of Dr Prakash V. Patil, Professor of Pathology & Dean, Mayo Institute

Medical Sciences, Barabanki, UP.

Dr S. G. Kardesai MD Director Academic Affairs KLE University

Belgaum 590010.

Prof. (Dr) V. D. Patil MD, DCH.

Registrar

KLE University

Belgaum 590010.

Date: Date:

V





Copyright DeclarationCopyright DeclarationCopyright DeclarationCopyright Declaration

We hereby declare that KLE Academy of Higher Education and

Research, Belgaum (Karnataka) shall have the rights to preserve, use and

disseminate this thesis entitled “Validation of a New Classification System

for Categorizing Oral Epithelial Dysplasia” in print or electronic format

for academic and or research purpose.

Place: Belgaum

Date :

Guide :

Dr Prakash V. Patil MD, Ph.D. Professor of Pathology

Dean, Mayo Institute Medical Sciences

Gadia, Barabanki.

Research Scholar :

Dr Alka Kale MDS

KLE VK Institute of Dental

Sciences, Belgaum.

© KLE ACADEMY OF HIGHER EDUCATION AND RESEARCH, BELGAUM

VI

ACKNOWLEDGEMENT

I express my sincere thanks to Dr Prabhakar Kore Hon. Chancellor, KLEUniversity, who with his great vision and undaunted effort created this empire ofEducation to enable plenty of scholars pursue their further studies and research work. Ihumbly thank Prof. (Dr) C. K. Kokate, Hon. Vice-Chancellor who has shaped KLEUniversity in a well acclaimed global university. He has always been an inspiration and arole model who has instilled in me the quest to improve by learning and research.

The work presented here is due to the encouragement and motivation given byProf. (Dr) Prakash V. Patil, (MD Pathology & Microbiology, Ph.D) Ex-Prof. & Head,Department of Pathology, J. N. Medical College, Belgaum and presently Professor ofPathology and Dean of Mayo Institute of Medical Sciences, Barabanki, for his constantguidance to me. Without his stewardship the completion of this research work andpresentation in the form of thesis would not have been possible. I am grateful to him.

I am thankful to Prof. Warnakulasuriya S, King’s College, London who had senthis valuable opinion and criticism on our earlier work, and to my teacher Prof. V K.Hazarey, Dean, Government Dental College, Nagpur for reviewing my ideas and givingvaluable inputs, suggestions and guidance in grading dysplasia. I am indeed obliged tohim.

I am indebted to the constant help and everlasting enthusiasm of my co ll ea gu esDr Seema Hallikerimath, Dr Punnya Angadi and Dr Deepa Mane throughout my study andalso to be the co-observers for the analysis of 100 cases. I thank them for constantlysu pp or ti ng me with their timely help & co ns tr uc ti ve su gg es ti on s . A special thanksto Dr Deepa Mane for helping with IHC procedures. I thank Dr Pushpak Shah for thephotomicrographs, Dr Chetan & all my department colleagues and past & presentpostgraduate students for their support and help. My thanks are also due to Dr MinalChoudhary, Prof. Oral Pathology, Datta Me gh e De nt al Co ll eg e , Wardha, Dr AlkaDive, Prof and Head, Or al Pa th ol og y & Dr Devendra from Lata Mangeshkar DentalCollege, Nagpur for the pilot study. I am es pe ci al ly ob li ga te d to the help rendered byDr Kunal Sah, Vinuth DP, Gaurav Sapra, Santosh Hiremath, Dipika Shukla & NiteshNaresh.

I am truly indebted to Prof. S. S. Mallapur, Faculty, Department of Statistics, KLEUniversity, Belgaum for his endurance to patiently evaluate my data and explain to me theimplication of the statistical analysis. I also extend my sincere thanks to Dr MamataHebbal and Dr Vaishali Keluskar for contributions made, Dr Anjana Bagewadi, Vice-Principal and Dr Anil Ankola, for sharing my administrative responsibilities. I amextremely thankful to Mr. Mahesh Desai for his constant help in preparing the entiredocument. Mr S. D. Pandit histopathology technician and all the non-teaching staff fortheir contribution.

Dr Alka Kale

VII

LIST OF ABBREVIATIONS

AgNOR : Argyrophilic Nuclear Organizing Region

APES : Aminopropylethoxysilane Solution

CK /K : Cytokeratin

CK-19 : Cytokeratin – 19

DFC : Dense Fibrillar Components

EGFR : Epidermal Growth Factor Receptor

FC : Fibrillar Centres

GC : Granular Components

H&E : Hematoxylin and Eosin

HOIN : High grade oral intraepithelial neoplasia

HUGO : Human Genome Organization

IHC : Immunohistochemistry

Ki-67 : Kiel, Germany 67th well on the plate

Ks : Simple Kappa

Kw : Weighted Kappa

LOIN : Low grade oral intraepithelial neoplasia

MMP-9 : Matrix metalloproteinase -9

NORs : Nucleolar Organising Regions

OED : Oral Epithelial Dysplasia

OIN : Oral Intraepithelial Neoplasia

OL : Oral Leukoplakia

OSCC : Oral Squamous Cell Carcinoma

PCNA : Proliferating Cell Nuclear Antigen

PMD : Potential Malignant Disorder

PVL : Proliferative Verrucous Leukoplakia

SCC : Squamous Cell Carcinoma

SIN : Squamous Intraepithelial Neoplasia

TAM : Tumor Associated Mucosa

WHO : World Health Organization

VIII

ABSTRACT

Introduction: The mortality of oral cancer has increased seven times in the last 50 years

in-spite of millions worth research going on globally. India has one of the highest rates in

oral cancer in the world partly attributed to high prevalence of tobacco chewing.

Alarmingly 194 million people aged 15 years and older have tobacco habit. In parallel to

the increase in Oral Cancer, border-line malignant lesions which range from epithelial

dysplasia to intra epithelial carcinoma have also increased in numbers. These lesions are

significant part of routine oral pathology service. Not all precancerous lesions and

conditions transform to cancer. Morphological alterations amongst some may have an

increased potential for malignant transformation.

The current histopathological grading of oral dysplasia lesions is notably unreliable

mainly due to lack of objective and validated grading features. Many studies have shown

wide variability in the diagnosis and grading of oral epithelial dysplasia (OED) with

results demonstrating only poor to moderate agreement. The customary grading system of

oral epithelial dysplasia into mild, moderate and severe does not allow accurate prediction

of which cases may eventually transform into malignancy. Many studies have remarked

that precancers with epithelial dysplasia have shown to develop into cancer more readily

than lesions without. Nevertheless all precancer or epithelial dysplasia do not develop into

cancer, whereas, some have even shown to regress with time.

Objectives: The present study was undertaken to assess the different grading systems to

determine the inter observer and intra observer variation and to study the individual

features defining atypia and general disturbance in the epithelium of oral epithelial lesions;

dysplasia. An attempt is also made to correlate the histopathological features with

Immunohistochemistry markers for cytokeratin (CK-19), basement membrane

enhancement protein (Ki-67) and extracellular matrix change (MMP-9) so as to

substantiate the observation of atypia observed in H&E sections. The study will also try to

validate a new binary classification which will objectively define each feature and classify

oral epithelial dysplasia as ‘risky’ and ‘non-risky’ using the H&E stains as gold standard.

Methodology: Pilot study to evaluate inter and intra observer variability was done on 25

precancerous cases. 100 cases of potentially malignant disorders were subjected to inter

and intra observer variability using two classifications: Smith and Pindborg (1969), Binary

Classification of Kujan (2006) and WHO (2005). Evaluation was doneto validatea new binary

IX

classification based on scores for individual features that was also substantiated by IHC.

The data was statistically evaluated by Kappa analysis and Fisher’s Exact Test.

Results: Almost perfect agreement was obtained by the new binary classification (Non-

risky & Risky) (k = 0.790) in comparison with the other classification where k value =

0.27 and kw =0.483. Amongst the 100 cases evaluated we found 8 cases to be of high risk

and were finally graded as ‘risky’. Immunoexpression of three markers was found to be

maximum in these 8 cases. The two features of epithelial dysplasia identified and found to

be statistically significantly present in the risky cases were abnormal and superficial

mitosis and marked cellular and nuclear pleomorphism and anisonucleosis.

Conclusions: Binary classification with scores for individual features of Epithelial

dysplasia was found to be the better method of grading which was statistically significant.

Abnormal and superficial mitosis and marked cellular and nuclear pleomorphism and

anisonucleosis were identified as features to be present in a given H&E stained

histological section of oral dysplasia to be graded in the ‘risky’ category suggesting their

chances of malignant transformation.

Keywords: Binary Grading, Potentially Malignant Disorders, Oral Epithelial dysplasia,

Tobacco, Leukoplakia, Erythroplakia.

X

CONTENTS

Sl. No. Sections Page No.

1. INTRODUCTION 1

2. AIM AND OBJECTIVES 6

3. REVIEW OF LITERATURE 7

i. General

a. The Epidemiology of Head and Neck Precancers

b. Epithelial Dysplasia

c. Clinical features of Potentially Malignant Disorders

d. Histopathology of Potentially Malignant Disorders

e. Grading Systems for Oral Epithelial Dysplasia

f. Subjectivity / Variability in Oral Epithelial Dysplasia

g. Various Methods of Evaluation of Oral Epithelial Dysplasia

h. Molecular Aspects of Potentially Malignant Disorders

i. Correlation of Dysplasia with Eventual MalignantTransformation

ii. Ki-67 Protein

iii. Matrixmetalloproteinase

iv. Cytokeratin

7

7

9

14

18

22

39

45

56

60

64

66

74

4. METHODOLOGY 79

5. RESULTS 89

6. DISCUSSION 105

7. CONCLUSIONS 127

8. BIBLOGRAPHY 130

9. ANNEXURES – 1 to 9

XI

LIST OF TABLES

Table No. Particulars Pages

1 Statistical analysis of 25 cases for Pilot Study 89

2 Statistical analysis of 100 study cases for Inter observervariability

90

3 Statistical analysis of 100 study cases for Intra observervariability

90

4 Scoring for atypical architectural and cytological features forthe proposed classification

91

5 Distribution of cases into ‘non-risky’ (total score ranged from1-15 ) and ‘risky’ (total score ranged from 16 to 84) groups

92

6a Calculation for risk categories using median and inter quartilefor 100 cases

92

6b Classification of risk categories for evaluation of 59 riskycases

92

7 Risk categorization of Risky and Non-risky cases 93

8 Ten Important Features with higher scores considered in‘Risky group’ are segregated with a total score of 61

94

9a Calculation for risk categories using median and inter quartilefor study group for 10 important features

95

9b Classification of risk categories for evaluation of cases for 10features

95

10 Risk categorization of risky and non risky groups for the 10features

96

11a CK-19 Area of expression in Epithelium of the study groups(Fisher Exact Test)

97

11b CK-19 Expression in Epithelium of the study groups (FisherExact Test)

97

12a Ki-67 Expression of area in the study groups (Fisher ExactTest)

98

12b Ki-67 Expression in study groups (Fisher Exact Test) 98

13a MMP-9 Expression in Epithelium of the study groups (FisherExact Test)

99

13b MMP-9 Expression in stroma of the study groups (FisherExact Test)

99

XII

LIST OF FIGURES

Figure No. Particulars Pages

1 ‘Non-risky’ dysplasia (a - 10x, b - 40x, H&E stain) 100

2 ‘Risky’ dysplasia (a - 10x, b & c - 40x, H&E stain) 100

3 ‘Non-risky’ dysplasia with minimal MMP-9 Expression (a- 10x,b- 40x, IHC stain)

101

4 ‘Risky’ dysplasia with enhanced MMP-9 Expression (a & b - 10x,c- 40x, IHC stain)

101

5 ‘Non-risky’ dysplasia. Restricted Ki-67 expression in basal celllayer. (a- 10x, b- 40x, IHC stain)

102

6 ‘Risky’ dysplasia. Enhanced Ki-67 expression in suprabasal layer.(a- 10x, b & c- 40x, IHC stain)

102

7 ‘Non-risky’ dysplasia. Restricted CK-19 expression in basal celllayer. (a -10x, b- 40x IHC stain)

103

8 ‘Risky’ dysplasia. Enhanced CK-19 expression in basal andsuprabasal layers. (a & b -10x, c- 40x, IHC stain)

103

9 Individual Features of Epithelial Atypia (H&E stain) 104

Introduction

1

INTRODUCTION

A disease will affect the seething mass of activity within a living cell and reflect

dysfunction. Corten et al1 stated that the cells tend to maintain a relatively normal range of

structure and function designated as normal, until the cellular stresses become serious. The

specialty of pathology begins with the examination of disease at cellular level which

modifies its structure and function in response to changing demands and advances.

Cancer evokes the strongest emotions in mankind. It connotes pain, protracted

suffering, hideous growth and death. An array of malignant neoplasms can arise from oral

tissues but squamous cell carcinomas of the oral mucosa constitute 90% of oral

malignancies worldwide. Being one of the most mutilating diseases, it is the sixth most

common malignant neoplasms in the world. However, in Indian sub-continent, it is the most

common cancer accounting for 1/3rd of all malignancies.

The significance of oral cancer involves its mortality and the disfigurement and

dysfunction associated with treatment. The rate of curability of cancer depends on the stage

and site. Small cancers for example on anterior tongue, floor of the mouth and buccal

mucosa have a local control rate upto 90%. Moderately advanced lesions without evidence

of spread to cervical lymph nodes have control rates upto 80-90%, whereas the intra oral

carcinomas with cervical lymph node metastasis, the survival rate is less than 50%.2

The mortality of oral cancer has increased seven times in the last 50 years in-spite of

millions worth research going on globally. India has one of the highest rates in oral cancer in

the world partly attributed to high prevalence of tobacco chewing. Forms of tobacco

chewing include pan (piper betel leaf with sliced areca nut, lime, catechu and other spices

chewed with or without tobacco), pan masala or gutkha (a chewable tobacco containing

areca nut) and mishri ( a powdered tobacco rubbed on the gums as toothpaste). According to

WHO 65% of men and 33% woman consume tobacco. Out of this 29.3% men and 2.3%

women smoke tobacco; 28.1% men and 12.0% women chew tobacco, whereas 46.5% men

and 13.8% women either smoked or chewed. Alarmingly 194 million people aged 15 years

and older have tobacco habit.79% of people who consumed some form of tobacco lived in

rural areas.

Introduction

2

In India, the prevalence of smoking and chewing varies. Chewing tobacco / pan

masala is common in Central, Eastern, Western (except Goa) & Northeastern states (except

Meghalaya) compared to Northern and Southern states, whereas, smoking is high in

Northern states except Punjab.3,4 Byakodi et al reported 1.12% (112 cases) prevalence rate

of oral cancer from 35,122 cases screened.5

R. Shankarnarayan in his research communication reported that out of 1151oral

biopsies received from 2003 – 2007, 344 were potentially malignant lesions and 422 were

oral cancers. He reports that though there is no clear cut trend in prevalence, a gradual

increase in number of cases was noted.6

In parallel to the increase in Oral Cancer, border-line malignant lesions which range

from epithelial dysplasia to intra epithelial carcinoma have also increased in numbers. These

lesions are significant part of routine oral pathology service. This is due to increased

awareness, improved consciousness towards health, availability of health care centers at

primary, secondary and tertiary levels, which results in early detection of oral borderline cases.

Over 90% of all oral cancers are squamous cell carcinomas (SCC) arising from the

lining mucosa. SCC may arise de novo in clinically normal mucosa, though a majority is

preceded by visible changes of the mucosa. There are several histologically distinct lesions

of oral mucosa which have malignant potential. Among these are leukoplakia, erythroplakia,

proliferative verrucous leukoplakia, oral submucous fibrosis etc. These vary in their

malignant potential and genetic background. The lesions may be found in association with

and / or preceding SCC.

Bouquot JE et al (2006) have rated the malignant transformation potential of

precancers, with proliferation verrucous leukoplakia and erythroplakia topping the list with

maximum potential whereas smokeless tobacco keratosis and lichen planus to have the least

malignant potential. 7

The term ‘Precancer’ has assumed newer terminologies as “Potentially malignant

lesions, oral precursor lesions, premalignant conditions etc”. These lesions present either as

white macule with excess surface keratin or as red, non-blanching macules with minimal

surface keratin. Histologically these precancerous lesions present with epithelial dysplasia,

which is described by Stedman's Medical Dictionary as a disorder of differentiation of

Introduction

3

epithelial cells which may regress, remain stable, or progress to invasive carcinoma. The

features of epithelial dysplasia include: Increased nuclear to cytoplasmic ratio, acanthosis,

atypical mitosis, pleomorphism, nuclear hyperchromatism, enlarged darkly stained nucleoli,

loss of cellular adhesion and abnormal keratinization. These features vary with individual

tissue specimens and assessment varies with pathologists viewing the same tissue specimen.

Prognostic significance of a lesion is difficult to determine because of the above factors.7,8

Leukoplakia is the most commonly encountered potentially malignant disorders. 2%

of leukoplakia may undergo malignant transformation per year. In lesions with epithelial

dysplasia this rate could vary between 1.1% -17.5%.9 Longitudinal studies of rural

populations in India revealed 80.6% of oral cancers being preceded by precancerous lesions

or conditions. 10,11

Though erythroplakia is uncommon, it often presents with dysplasia or intra

epithelial carcinoma when compared to leukoplakia which generally show no dysplasia or

may just present with mild to moderate dysplasia. Similarly the malignant transformation

potential for smokeless tobacco keratosis has been questionable. It is known that the tobacco

habit itself is said to carry a risk four times greater than normal mucosa. About 16% of

biopsied lesion showed mild dysplasia, and a few chewers showed severe dysplasia. Some

authorities have used clinical grading scale for smokeless tobacco keratosis based on

intensity of whiteness, erythema and fissuring. Unfortunately none of these grading systems

have successfully correlated severe or high grade lesions with an increased risk of

malignancy.

Not all precancerous lesions and conditions transform to cancer. Morphological

alterations amongst some may have an increased potential for malignant transformation.

These are also indicators of risk of likely future malignancies elsewhere in clinically normal

appearing oral mucosa and not only site specific predictors.12

There is always a challenge for oral pathologists to assess the degree of dysplasia in

potentially premalignant oral lesions with accuracy, so that the burden of malignant

transformation can be predicted and treatment modalities can ensure reduced morbidity.4


mainly due to lack of objective and validated grading features. Many studies have shown

Introduction

4

wide variability in the diagnosis and grading of oral epithelial dysplasia (OED) with results

demonstrating only poor to moderate agreement.13

The customary grading system of oral epithelial dysplasia into mild, moderate and

severe does not allow accurate prediction of which cases may eventually transform into

malignancy. Many studies have remarked that precancers with epithelial dysplasia have

shown to develop into cancer more readily than lesions without.14 Nevertheless all precancer

or epithelial dysplasia do not develop into cancer, whereas, some have even shown to

regress with time.15

The limitations inherent in the histologic grading, suggest the necessity to improve

histologic assessment and the need to identify more reliable makers. In the last decade

approach to understanding of molecular carcinogensis has revealed newer directions. Many

molecular markers functioning as elements of cell cycle control have been recognized.

These have been tested individually and in combinations to develop diagnostic markers for

identification of dysplasia and their ability to predict malignant potential.16

Most of the authors have concluded that though the progress in molecular oncology

has significantly advanced the knowledge on tumorigenesis, yet the practical applications of

these genetic makers remain unresolved in detecting oral dysplasia. None of the markers, the

authors studied like EGF, TFG-α, EGFR, Ras, Myc, Cyclins, p53, p21, pRb including DNA

ploidy, loss of heterozygosity, apoptosis, Cox1& 2, cell adhesion molecules, angiongensis,

proliferation and differentiation markers and viruses, either singly or in combination appear

to be ready for use in routine clinical diagnostic practices. Not many studies have focused

correlating the findings at molecular level with the clinical attributes of terminal outcome.

From the observation of Gayani P et al it can be concluded that the molecular studies

are not only nonspecific but also require expertise, armamentarium and facilities which may

not be available for day to day diagnosis at clinics and primary or secondary health

centres.16

A variety of precancers are successfully evaluated and managed despite ongoing

controversies. Bouquot JE states that one of the major advances in this field has been the

simple recognition that a premalignancy is not guaranteed to eventually transform into

cancer. Thus, he defines premalignant lesion as an identifiable, benign tissue or cellular

Introduction

5

alteration associated with greater than normal risk of malignant transformation or

“degeneration”. Much of contemporary oral oncology research is focused on the

identification of clinical, histopathologic and molecular biological parameters of oral

precancers and early cancers which might aid in the recognition of those lesions with the

higher risk of transformation.17

The present study was undertaken to assess the different grading systems to

determine the inter observer and intra observer variation and to study the individual features

defining atypia and general disturbance in the epithelium of oral epithelial lesions; dysplasia.

An attempt is also made to correlate the histopathological features with

Immunohistochemistry markers for cytokeratin (CK-19) basement membrane enhancement

protein (Ki-67) and extracellular matrix change (MMP-9) so as to substantiate the

observation of atypia observed in H&E stained tissue sections. The study will also try to

validate a new binary classification which will objectively define each feature and classify

OED as ‘risky’ and ‘non-risky’ using the Hematoxylin and Eosin stains as gold standard.

Aim & Objectives

6

AIM AND OBJECTIVES

Primary Objective (Aim) :

To propose a new classification for dysplasia to categories dysplasia into ‘non risky’

and ‘risky’ groups

Secondary Objective :

1. To histopathologically evaluate 100 cases of potentially malignant disorders (PMD)

for degree of dysplasia based on the classification system given by (a) Smith and

Pindborg’s classification (1969) as mild, moderate and severe, using photographic

standardization and scores (b) WHO classification (2005) as mild, moderate &

severe dysplasia and (c) Binary classification as low risk and high risk cases given

by Kujan Omar (2006).

2. To assess the inter and intra observer variability and agreement for the above

mentioned 3 classification systems.

3. To histopathologically evaluate the degree of dysplasia for the same 100 cases of

potential malignant disorders (PMD) using the proposed classification.

4. To assess the inter and intra observer variability and agreement for the proposed

classification.

5. To identify individual histopathological features within the proposed classification

that would aid in categorization of risk behavior.

6. To correlate the proposed classification for degree of dysplasia with

immunohistochemical markers for : CK-19 for epithelial differentiation, Ki-67 as

proliferative marker and MMP-9 as marker for epithelial - connective tissue

interaction.

7. To correlate the observations of IHC with histopathological features analyzed with

the new proposed classification.

Review of Literature

7

REVIEW OF LITERATURE

Oral Cancer is one of the top three cancers in India. Out of which 30% is oral

squamous cell carcinoma (OSCC). The variation in incidence and pattern of oral cancer is

due to regional differences in the prevalence of risk factors. Literature review demonstrates

that nationwide incidence varied considerably based on study designs, sampling

methodology as well as by age, gender and location. Variations in age specific incidence

rates also increased with age, which drops at the age of 70 years and is consistent with

multiple studies. The diagnosis of oral cancer also depends on the clinical examination

conducted which can lead to variations.

a. THE EPIDEMIOLOGY OF HEAD AND NECK PRECANCERS

Prevalence and Incidence : Investigations in Western population have reported that

leukoplakia affects approximately 30-40 of every 1000 adults (3-4%) whereas prevalence

rates in India is about 170/1000 in some communities, though this may varies with other

population. Oral lichen planus, affects one of every 1000 adults in Western populations; in

certain subpopulations it is seen as 1/32 persons. No data are available for erythroplakia,

but incidence of oral/pharyngeal carcinoma in situ, which represents erythroplakias, is one

per million persons each year.17

Patient Age and Gender : Precancers are usually found more frequently in men than in

women. More than 60% and 80% of leukoplakia and erythroplakia patients are male and

the proportion of males increases with increasing age at diagnosis. More than 95% of

smokeless tobacco keratosis patients are men, except in populations wherein the chewing of

tobacco is popular among women. Only 30-50% of oral lichen planus patients are males.

Two-thirds of oral submucous fibrosis patients in India are females, but this proportion

varies as the men in some cultures use areca nut more than do the women.

Smokeless tobacco keratosis prevalence varies with the popularity of the smokeless

tobacco habit in a given population. In Western investigations it has been diagnosed in as

many as 9% of adult males. As a general rule, up to half of daily users will show at least a

mild degree of keratosis. In some Indian communities the prevalence is higher but is

confused with leukoplakia in reported epidemiologic investigations. Oral submucous

fibrosis is rare in Western cultures but occurs in as many as four per 1000 persons in rural

communities of the Indian subcontinent and 11/1000 among chewers of betel quid.17


8

Smokeless tobacco keratosis is typically visible within a few years of the onset of the

smokeless tobacco habit and remains indefinitely unless the habit is discontinued. The

average age at diagnosis of smokeless tobacco keratosis is usually thought to be 45-60 years.

These lesions are initiated much earlier, however, and population-based prevalence studies

in young adults and teenagers have found keratotic changes in almost a third of young

smokeless tobacco users. Western investigation of smokeless tobacco keratosis to provide

age-specific prevalence rates identified a biphasic pattern with increased prevalence in

young adult males and in males older than 65 years of age.17

Coelho KR18 reported that 57% men and 11% women between 15-59 years use some

form of tobacco. Use of smokeless tobacco is more prevalent as it is accepted socially and

culturally in India. Gutkha, Zardha, Kharra, Mawa and Khani are all dry mixtures of lime,

areca nut flakes and condiments with or without powdered tobacco, custom mixed and are

commercially available and sold by vendors. Pan or guktha is kept in the cheek and chewed

or sucked for 10-15 min, sometimes kept overnight.

High incidence of oral cancer is attributed to the use of tobacco as quoted in several

studies : - Regression of leukoplakia with reduced tobacco consumption (Mehta et al);

Cessation of tobacco habits dropped the incidence of leukoplakia (Gupta et al) ; Tobacco

chewing (betel quid or Khaini) and smoking (Bidi & Cigarette) are common cause of oral

cancer – 48% cases stage III and IV (Khandekar et al) ; Significant association between oral

cancer incidence and daily frequency of tobacco chewing (p<0.001), which increased 9.2

fold among women chewing tobacco 10 times a day or more, with the highest risk during the

first twenty years of chewing (Jayalakmi et al). As quoted in the study by Coelho et al.18

Coelho K R.18 noted the oral cancer incidence from 1990 to 2005 and stated, an

increase in incidence of oral cancer among young adults especially tongue cancer. Age

related variability in the incidence of oral cancer in different region of India has increased

over the years.

500 students were selected and subjected to questionnaire and urine rapid nicotine

test in a study conducted by Saddichha S. et al19 55.6% confirmed tobacco use. Out of which

78% reported to be cigarette smoking, 20% used khaini and 2% of the college students

consumed gutkha. They further reported that 60% - 80% patients present with advanced

disease as compared to 40% in developed countries. In India 2.5 million people are living


9

with disease; there are 0.8 million new cases and 0.55 million deaths due to oral cancer

every year.

Makwana NR et al20 reported from the questionnaire based study on consumption of

tobacco containing products that 28.41% of school children were of 10-13 years of age,

33.56% were 14-16 years and 36.26% were 17-19 years old. They stated that the reason for

addiction was due to peer pressure i.e. friends in 4.4% of children, 1.4% stated that movies

influence them and 94.2% did not respond.

Hazarey VK et al21 in a hospital-based cross-sectional study on 1000 OSF cases

reported the male-to-female ratio as 4.9:1. Occurrence of OSF was seen <30 years among

men when compared with women (OR = 4.62, 3.22–6.63, P = 0.0001). Reduced mouth

opening, altered salivation and taste sensation was noted. Exclusive areca nut chewing habit

was significantly more prevalent in women (OR = 44.5, 25.4–79.8, P = 0.0001). Whereas

significant increase for gutkha (Areca quid with tobacco) (OR = 2.33, 1.56–

3.54, P = 0.0001) and kharra / Mawa (crude combination of areca nut and tobacco)

(OR = 6.8, 4.36–11.06, P = 0.0001) chewing was found in men when compared with

women.

b. EPITHELIAL DYSPLASIA

Baillie and Royal Society Committee in Edinburgh first proposed the concept of

premalignancy in 1806. Sir James Paget in 1851 was first to apply this concept to pipe

smokers with “Smokers Patch”. Alterations in oral mucosa have been closely involved in the

advancement of our understanding of precancers or potentially malignant lesions. An

association between benign oral mucosal lesion and its subsequent malignant transformation

was clearly implicated by Sir James Paget in 1870. Oral “ichthyasis” (white keratotic

plaque) was a significant precursor to lingual carcinoma. In 1877 Schwimmer coined the

term “Leukoplakia” for white tongue changes seen prior to lingual cancer development in

tertiary syphilis. The term : “erythroplakia” was first used by Queyrat in 1911 to describe a

red macule of the genital organ in a syphilitic male patient. In 1924 Darier, a French

dermatologist had applied the term “erythroplakia of Queyrat” to a lesion of the oral

mucosa.17

The term dysplasia was introduced by Reagon in 1958 where he described the

features of dysplasia in relation to exfoliated cells from cervical lesion.22 In Medical


10

terminology the word dysplasia means an abnormality in development while

histomorphologically dysplasia expresses the cellular and structural changes in the

epithelium.

Pindborg (1977) defined epithelial dysplasia as lesions in which part of the thickness

of the epithelium is replaced by cells showing varying degree of cellular atypia. The term

intra epithelial neoplasia and atypical epithelial hyperplasia were being synonymously used.

When the appearance of cancer is preceded by some other lesion in the epithelium

that lesion may show cellular changes from that point towards the possible subsequent

development of malignancy. The individual cellular changes are referred as ‘atypia’ and the

general disturbance in the epithelium is designated as ‘dysplasia’. Some features of

epithelial dysplasia are also represented in normal growth, proliferation, maturation and

organization of cells.23

Cellular responses which involve stimulation of cell growth such as hyperplasia or

change in cell type as in metaplasia do not by themselves predispose to malignancy.

Abnormal cell proliferation shows itself as visible changes as dysplasia in the microscope

which literally means “disordered growth”.24

Dysplastic cells are part of the multistep pathway to overt malignancy. Until these

cells get the ability to invade the connective tissue and metastasize, they are not defined as

‘malignant’. If these dysplastic cells are studied by tissue culture, they will show some

features of a “transformed cell”. The progressive changes of mild to severe dysplasia and

then the development of invasive activity are best seen in epithelial tissue. The dysplastic

cells invade the lower, middle and then the full thickness of the epithelium and eventually

invade the basement membrane.24

Oral epithelial dysplasia : a histopathologic marker of premalignancy, is a diagnostic

term used to describe the histopathological changes seen in a progressive chronic and

premalignant disorders of the oral mucosa. Clinically it may present as leukoplakia,

erythroplakia or leukoerythroplakia. It may also be seen in verrucous or papillary

leukoplakias or in the margins of chronic mucosal ulcers. It is also consistently seen in the

apparently normal mucosa adjacent to the squamous cell carcinoma.25

To develop a better understanding of the biology of oral precancer and to achieve

consistency in diagnosis an international working group comprising specialists in the fields


11

of epidemiology, oral medicine and pathology and molecular biology met in London in May

2005 for a workshop co-ordinated by WHO collaborating center for oral cancer and

precancers. They suggested the following recommendations : -

Concept of Precancer : -

1. In longitudinal studies, area of tissue with certain alterations in clinical appearances

identified at the first assessment as “precancers” have undergone malignant change

during follow-up.

2. Some of these alterations, particularly red and white patches, are seen to co-exist at

the margins of overt oral squamous cell carcinomas.

3. A proportion of these may share morphological and cytological changes observed in

epithelial malignancies, but with frank invasion.

4. Some of the chromosomal, genomic and molecular alterations found in clearly

invasive oral cancers are detected in these presumptive “precancer” or “pre-

malignant” phases.

The terms “pre-cancer’, ‘precursor lesions’, ‘premalignant’, ‘intra epithelia

neoplasia’ and ‘potentially malignant’ have been internationally used broadly to describe

clinical presentations that may have potential to become malignant. The 2005 WHO

monograph on Head and Neck tumors, uses the term ‘Epithelial precursor lesions’.

A much earlier working group of WHO in 1978 classified the lesion into two broad

groups as (i) Lesions and (ii) Conditions with following definitions: -

A precancerous lesion is “a morphological altered tissue in which oral cancer is

more likely to occur than its apparently normal counterpart”.

A precancerous condition is “a generalized state associated with a significantly

increased risk of cancer”.

It is now known that even the clinically ‘normal’ appearing mucosa in a patient

harboring a precancerous lesion may have dysplasia on the contralateral anatomic site or

molecular aberrations in other oral mucosal sites suggestive of a pathway to malignant

transformation.

The consensus of the 2007 working group recommended the term “potentially

malignant disorders”. This describes all those lesions and conditions which may transform to

cancer. Potentially malignant disorders are indicators of risk of likely future malignancies


12

elsewhere in clinically normal appearing oral mucosa and not only the site specific

predictors. This working group did not favour subdividing precancers to lesions and

conditions and had the consensus view to refer to all clinical presentations that carry the risk

of cancer under the term ‘potentially malignant disorders’ to reflect the wide spread

anatomical distribution.12

Three major problems are attached to the importance of epithelial dysplasia in

predicting malignant development. (i) The diagnosis is essentially subjective and (ii) not all

lesions exhibiting dysplasia will eventually become malignant and some may regress (iii)

carcinoma can develop from lesions in which epithelial dysplasia was not diagnosed in

previous biopsy. As epithelial dysplasia does not seem to be invariably associated with or

may not be a necessary pre-requisite for malignant development there is a need to develop

other methods for predicting the malignant potential of precancerous lesions.14

The features that are predominantly caused by alterations of cell kinetics are found in

the proliferative compartments of the epithelium. This relates to an increase growth or cell

division. Those features that relate to disturbed maturation manifest in the form of irregular

stratification and increased keratinization of individual cells beneath the keratin layer.26

Most pathologists grade oral epithelial dysplasia according to a combination of

cellular and tissue changes. Low grade lesions are considered to be less ‘severe’ than high

grade lesions but this assessment is subjective. It is not unusual for different pathologist to

place emphasis on different aspects of dysplastic alteration, thereby arriving at varying

diagnostic conclusions. Regardless of individual emphasis, certain oral precancer grading

criteria are in general use :

The cellular atypia or dysplasia is similar to that seen in squamous cell carcinoma.

There is no evidence of invasion into underlying stroma

The epithelium with the greatest proportion of atypical cells has the greatest risk of

being or becoming a carcinoma.

The epithelium with the most extreme atypia of cells has the greatest risk of being or

becoming a carcinoma.

The final grading or diagnosis should be based on the most severely involved area of

change, even if that area includes no more than a few rete processes.17


13

Epidemiological data suggests that a proportion of many oral mucosal lesions

including idiopathic white patch (leukoplakia), speckled leukoplakia, erythroplakia, chronic

hyperplastic candidosis, oral submucous fibrosis, discoid lupus erythematosus and

dyskeratosis congenita undergo malignant transformation. At present, it is not possible to

predict which dysplastic oral lesion will proceed to malignancy and which will regress. A

molecular study found identical genetic changes in the dysplastic oral lesions and oral

cancers in the same patients, suggesting that the progeny of cells in a dysplastic oral lesion

may, with further modification, ultimately form a malignant lesion. This is the first evidence

that a clone of oral keratinocytes proceeds through stages of transformation from normal, to

dysplastic, to malignant. The findings suggests that oral keratinocytes suffer genetic damage

at each stage boundary until the accumulated genetic damage is sufficient for autonomous

proliferation and tumor formation.2

The New WHO classification (2005) applies the term dysplasia for “presence of

architectural disturbances accompanied by cytologic atypia.” The degree of dysplasia is

determined as a measure of tissue and cellular deviation from normal.

The features of epithelial dysplasia are summarized below as quoted by (Soames JV

and Southam JC 27:

1. Nuclear and cellular pleomorphism: Nuclei and cells show variation in size and

shape- anisocytosis and anisonucleosis.

2. Reversal of nuclear cytoplasmic ratio: Increase in size of nucleus with respect to

the cytoplasm.

3. Nuclear hyperchromatism: Abnormally intense nuclear staining.

4. Prominent nucleoli.

5. Increased and abnormal mitoses: Normal mitoses are increased in number and

found higher up in the epithelium than is usual i.e. away from basal layer :

suprabasal mitoses. These may have an abnormal form.

6. Disturbed polarity of the basal cells or of cellular orientation. The cells in the

basal layer have no definable long axis and the nuclei have no polarity.

7. Basal cell hyperplasia: The presence of several layers of cells having basaloid

appearance. It is often associated with drop shaped retepegs.


14

8. Drop shaped retepegs: The retepegs are wider at their deeper part than they are

more superficially.

9. Irregular epithelial stratification or disturbed maturation: The cells no longer

show a proper sequence of morphological and maturational changes as they pass

from the basal layer to the surface.

10. Abnormal keratinization: Keratinization occurs below the normal keratin layer,

either as individual cell keratinization within the stratum spinosum or as disturbed

maturation of group of cells resulting in the formation of intraepithelial keratin

pearls.

11. Loss or reduction of intercellular adhesion (or cohesion) : Cells of the stratum

spinosum remain no longer adherent to each other.

It should be noted that not all these changes are necessarily seen in any one case.27

c. CLINICAL FEATURES OF POTENTIALLY MALIGNANT DISORDERS

The clinical features and prognosis of oral precancers (PMD) will depend on the

exact precancer involved. As a general rule they present as white lesions with excess surface

keratin or as red lesions with minimal surface keratin.

Leukoplakia : is defined as a chronic white mucosal macule which cannot be scraped off,

cannot be given another specific diagnostic name, and does not typically disappear with

removal of known etiologic factors, excepting tobacco. Oral leukoplakia occurs most

frequently on the lip vermilion, buccal mucosa and gingival mucosa, but many may be

misdiagnosed with frictional keratosis, though the tongue, lip vermilion and oral floor are

the sites most likely to present with leukoplakia containing dysplastic cells. Leukoplakia

("white patch") is the most common and well-researched oral precancer.

It should be emphasized that oral leukoplakia is a diagnosis of exclusion which

requires the clinician to be well acquainted with all other white oral lesions so as to be able

to rule them out prior to using the term leukoplakia for a particular keratosis in any patient.

It must be emphasized that the term leukoplakia is a clinical one. Even though biopsy and

microscopic evaluation is frequently required in order to identify dysplastic or malignant

cells, the presence of such cells does not alter the clinical diagnosis.

Case-control studies have strongly implicated smoked & smokeless tobacco and

alcohol abuse in the etiology of oral cancer, but have not further correlated white mucosal


15

plaques with malignant transformation. It is not known whether oral cancers in the smokers,

identified in epidemiologic studies arise from leukoplakia or from clinically normal

mucosa. However various investigations have found leukoplakia adjacent to a third (range:

10-100%) of oral squamous cell carcinomas.17

Numerous follow-up studies have established malignant transformation rates for

leukoplakia, ranging from 1-28%, with an average of 4% overall. The few investigations

which have examined the latter feature have found that 7% (range: 0.1%-40%) of oral

leukoplakia will have at least one focus of squamous carcinoma in the initial biopsy. Cancer

arising from a leukoplakic lesion may be diagnosed two to four years after the leukoplakia is

diagnosed, but transformation may occur within months or may occur after decades.

Leukoplakia is associated with progression to cancer when its surface becomes

thickened and rough or granular and when its surface becomes verruciform, or when red

areas of minimal keratin production are interspersed amongst the background of thickened

keratin (erythroleukoplakia, speckled leukoplakia).

Erythroplakia : Is a clinical term for a chronic red mucosal macule which cannot be given

another specific diagnostic name and cannot be attributed to traumatic, vascular or

inflammatory causes. Such lesions are less common than white precancers but very careful

observation will reveal erythroplakia in association with an early invasive oral carcinomas.

Erythroplakia may also be associated with leukoplakia (erythroleukoplakia), and in mixed

red and white lesions it is the red portion that is more worrisome than the white. Mucosa of

the lateral and ventral tongue, the oral floor and the soft palate is the common site for

erythroplakia reported in most of the cases.17

This lesion has been referred as "the dangerous oral mucosa" because under the

microscope, it typically presents as carcinoma in situ, severe epithelial dysplasia or

superficially invasive carcinoma. In high risk clinical situation, such as oral floor lesions in

heavy smokers and alcohol abusers, 80% of these red patches may contain focal areas of

micro invasive cancer at the time of initial biopsy.

Though follow-up studies are not available for erythroplakia, its usual microscopic

counterpart, carcinoma in situ, has been shown to transform into invasive carcinoma in

approximately one of every four cases. This malignant transformation occurs despite the

fact that carcinoma in situ lesions are routinely treated by conservative surgical removal or


16

laser ablation. Without therapy this disease transforms into invasive carcinoma in 60-90%

of the cases within 5-10 years after initial diagnosis.17

Smokeless Tobacco Keratosis: Also referred to as snuff pouch, snuff dipper's lesion, or

tobacco pouch is a chronic white or gray/translucent mucosal macule localized in areas of

direct contact with smokeless tobacco. This lesion cannot be scraped off, may disappear

with cessation of the tobacco habit, and is poorly demarcated from surrounding mucosa.

The altered mucosa is soft, velvety to feel and further on palpation, the tobacco chewer's

cheek will usually reveal a distinct "pouch". That results due to flaccidity in the chronically

stretched muscles in the area of tobacco placement. During a clinical examination, this

usually stretched mucosa appears folded or fissured. Induration, ulceration and pain are not

associated but occasional inflammatory erythema may be noted.17

It usually takes 5-10 years of tobacco habit for smokeless tobacco keratosis to

become apparent, but it may be present after less than a year. Once it occurs, it typically

remains unchanged indefinitely unless the daily smokeless tobacco contact time increases.

In such a case, it will gradually become thickened to the point of appearing as a distinctly

white, leathery or nodular plaque and become clinically indistinguishable from leukoplakia.

Development of smokeless tobacco keratosis in users is dependent on the type of

habit practiced in a society. For example Snuff is reported to produce more keratosis than

chewing tobacco, and people who keep the quid in a single selective site are more prone to

keratosis than those keeping in multiple sites. A specific brand of tobacco used, duration of

the habit, excessive daily contact-hours of tobacco on oral mucous membranes, an

increased amount of tobacco consumed daily, and a deficiency of beta-keratin or vitamin A

are other factors leading to high risk of keratosis.17

The malignant transformation potential of smokeless tobacco keratosis has not been

determined, but the tobacco habit itself is said to carry a risk four times greater than normal

mucosa, as observed in case-control studies of oral cancer patients. Few if any keratotic

lesions with serious dysplasia are reported in investigations using large numbers of tobacco

chewers, 16% of biopsied cases show at least mildly dysplastic cells in smaller

investigations.

Clinical grading scales for smokeless tobacco keratosis according to intensity of

whiteness, erythema and fissuring have been provided by some authorities. Unfortunately,


17

none of these has successfully correlated severe or high-grade lesions with an increased risk

of malignancy.

Oral Submucous Fibrosis: Patients complain of burning sensation of the oral mucosa,

occasional mucosal ulceration, marble-like blanching of the mucosa, and palpable fibrous

bands of the buccal mucosa, soft palate and lips. Leukoplakic lesions are commonly seen

along with oral submucous fibrosis and oral carcinoma development has been shown to

occur in 5% of users during a 15 year period of follow-up. This irreversible precancerous

condition is strongly associated with the habit of chewing areca nuts. Whether cancers

develop from leukoplakia or from the non-leukoplakic mucosa is as yet not established, but

it is presumed that leukoplakia behave in the same precancerous fashion as in persons

without oral submucous fibrosis. The increased risk of oral cancer in betel quid chewers is

mostly relevant to those who include tobacco in the quid, although all other features of the

disease remain the same.

Proliferative Verrucous Leukoplakia (PVL) : First described in 1985, it is a white

mucosal plaque or discoloration which virtually always develops nodular, papillary or

verruciform surface projections and which gradually, sometimes rapidly, spreads laterally to

encompass a large mucosal area. It is a very special form of oral precancer. Four of every

five affected patients are female and the mean age at diagnosis is 62 years (range: 22-89

years). The usual site of female involvement is the buccal mucosa (63% of cases), while the

tongue is most frequently affected in men (82% of cases). Etiologic factors are elusive.

Two-thirds of patients do not have a tobacco habit, but there are especially strong

associations with human papillomavirus (89%) and candida albicans (50%).

Almost half of PVL samples will demonstrate epithelial dysplasia at initial diagnosis

and few cases are spared this change eventually. Oral carcinoma will develop in more than

70% of affected patients during the decade following this diagnosis. Aggressive and

frequent surgical interventions and very close follow-up has been identified thus far as the

best treatment option.

Lichen planus : Is an autoimmune disorder that presents on the bilateral buccal mucosa as

intertwining, thin strands or streaks of white keratosis known as Wickham's striae. Lesions

tend to vary in intensity and are known to be induced by a variety of medications and other

antigens, stress, overwork, including betel quid or tobacco habit. The excess keratin


18

production may be the result of reduced proliferation of basal keratinocytes, with enhanced

maturation or differentiation of the epithelium. Skin involvement as purple to brownish,

pruritic papules are seen in some patients, but only the oral lesions are considered to be

precancerous. The more severe oral involvement demonstrates an erythematous background

(erosive lichen planus) and may present with bullae or ulcers (bullous lichen planus,

ulcerative lichen planus).

The erosive or atrophic forms of the disease, with its red mucosal background, are

considered to be more worrisome than the typical case of white striae on otherwise normal

mucosa. The malignant transformation rate for erosive lichen planus is probably less than

1%, according to follow-up studies. There are many who claim no malignant transformation

potential for oral lichen planus.

Malignant transformation in Lichenoid dysplasia too is debatable which presumes

that lichenoid microscopic changes associated with dysplastic cells are simply an immune-

mediated response to an otherwise unremarkable epithelial dysplasia. According to this

theory such lesions are not true lichen planus at all.

Recurring oral melanotic macule : Though oral melanoma is a rare cancer, it has been

estimated that one-third of these are preceded by an apparently innocuous mucosal

melanosis. A recent example of recurring oral melanotic macule has been reported to have

eventuated after seven years in melanoma. None of the six early biopsies demonstrated

histopathologic characteristics of melanoma; they were classic examples of oral melanotic

macule. DNA flow cytometry, however, showed obvious aneuploidy of cells in the early

biopsies, as well as in the final melanoma suggesting that histologically benign-appearing

oral melanosis may actually be the radial growth phase of oral melanoma in some cases. 17

d. HISTOPATHOLOGY OF POTENTIALLY MALIGNANT DISORDERS

Histopathologic features of oral precancers

Experience has taught us that certain cellular and tissue alterations are associated with

premalignancy and malignancy. Altered cells appear to be more primitive than normal and

so these changes are presumed to be examples of immature or inappropriate differentiation,

although pathologists typically refer to them as dysplasia or atypia. Animal models for oral

carcinogenesis have reliably demonstrated that normal epithelium passes through stages of

more and more severe dysplasia prior to the onset of invasive neoplasia. As a general rule,


19

fewer than 20% of oral leukoplakia will demonstrate dysplasia where 8% will have severe

dysplasia. On the other hand, erythroplakia show severe changes more than 90% of the

time.14

Cellular changes in dysplasia : Specific alterations of individual epithelial cells are

important in the determination of epithelial dysplasia. Cells and nuclei take on a more

primitive appearance, similar to those of basal cells with enlarged nuclei (nuclear

hyperplasia), dark-staining nuclei (hyperchromatism), enlarged, often eosinophilic nucleoli

(prominent nucleoli), and with an increased nuclear-to-cytoplasmic ratio. These cells also

appear to be placed closely together than normal keratinocytes (increased cellular density).

Such changes are not exclusive to carcinogenesis, as they may be seen in reactive epithelium

or epithelium influenced by a variety of systemic alterations. Flow cytometry and image

cytometry add significantly to the pathologist's ability to assess nuclear changes associated

with eventual cancer development.

There is often an increase in mitotic activity in dysplastic epithelium, but this also is

seen in many reactive lesions. Enlarged, tripolar or star-shaped mitotic figures (abnormal

mitoses), however, are much more indicative of precancerous changes. Abnormal mitosis

may also be defined as mitotic figures found in unusual locations above the basal cell layer.

A key alteration of dysplastic epithelial cells is variation in the shape of the cells and

nuclei. This pleomorphism is unusual outside of cancers and precancers. Premature

production of keratin below the surface layer is another important alteration, but it is much

more commonly seen in oral carcinomas than in oral premalignancy. This dyskeratosis may

be represented by individually keratinized cells or by tight concentric rings of flattened

keratinocytes (epithelial pearls). The pathologist must be careful not to misinterpret a

keratin-filled surface indentation cut tangentially as true intraepithelial keratosis. Not all

dyskeratosis is related to malignancy or premalignancy. Individually keratinized cells, for

example, are also characteristic of hereditary benign intraepithelial dyskeratosis (Witkop

disease).

Cellular necrosis and loss of cellular cohesiveness (acantholysis) are major signs of

poorly differentiated carcinoma but are extremely rare in the epithelial dysplasia of oral

precancer. When present, these features must be distinguished from intercellular edema,

intraepithelial inflammatory cells and degenerating cells with pyknotic nuclei and


20

vacuolated cytoplasm. Virus-induced koilocytes may be seen but are not necessary to the

diagnosis, nor does their appearance appear to be of prognostic significance.

When dysplasia is seen in epithelium which otherwise has the microscopic features

of lichen planus like liquifaction degeneration of basal cells, saw-toothed rete processes,

hyperkeratosis, sub-epithelial band of inflammatory/immune cells, etc, the lichen planus

features are essentially ignored and the lesion is graded according to the above mentioned

criteria for epithelial dysplasia, although the term lichenoid dysplasia may be applied to the

case.

Tissue (morphological) changes : Many oral precancers show excess surface keratin

(hyperkeratosis, hyperparakeratosis, hyperorthokeratosis) and most show hyperplasia of the

spindle cell layer (acanthosis), but both changes are common to a number of mucosal

lesions without any cancer transformation potential and neither is necessary for the

diagnosis of dysplasia.

Smokeless tobacco keratosis is characterized by a somewhat unique intracellular

vacuolization or "edema" of superficial layers, perhaps interspersed with streaks of

parakeratinized cells. This change most likely results from a low-grade chemical burn or

from the alkaline tobacco used. It has been referred to as surface etching and has no

implications to the risk of malignant transformation.12

Basal cell hyperplasia is of major importance to the diagnosis as well as to the

grading of dysplasia. Other tissue changes, however, are important features of oral

epithelial dysplasia, especially certain alterations of the rete processes.

Extremely elongated rete processes with minimal cellular atypia are of little

concern, as they are characteristic of a variety of hyperplastic conditions, including

papillomavirus infections, frictional keratosis, psoriasis and pseudoepitheliomatous

hyperplasia. Rete processes with a bulbous enlargement of the lowermost region (drop-

shaped rete processes), however, are worrisome regardless of their size, especially if

secondary projections or nodules are seen to arise from the basal layer and branch in

different angles into the lamina propria and connective tissue papillae. There is no

physiological explanation for secondary nodules extending laterally from a rete peg of the

oral mucosa. To many, these alterations are ominous enough to alter the histopathologic

grading of a lesion to a higher level.


21

Dysplastic epithelium may be atrophic as well as acanthotic and some experts

believe that atrophic forms have a higher risk of malignant transformation. Atrophic

epithelium often lacks rete processes and may be ulcerated, thereby mimicking a traumatic

or inflammatory lesion with thin, regenerating epithelium creeping in from the margins.

Regenerating epithelium often has granulation tissue beneath it to distinguish it from

precancerous dysplasia in atrophic epithelium. The very nature of thin, atrophic epithelium

presents a diagnostic grading dilemma because basilar hyperplasia very quickly extends to

the surface. There are no standards for this situation but Speight et al (1996) have

recommended that these be regarded as severe dysplasia. Leukoplakia of the vermilion

border of the lower lip in actinic cheilitis (actinic cheilosis, farmer's lip, sailor's lip) are

especially prone to these difficulties.14

An alarming morphological alteration of dysplastic epithelium is loss of

stratification (loss of polarization) due to an apparent inability to properly differentiate and

mature from basal cells to prickle cells to flattened keratinocytes. Cells high in the

epithelium have the same immature appearance as those in the basal layers. This feature is

especially pronounced in severe epithelial dysplasia and carcinoma in situ.

The pathologist must pay particular attention to the appearance of the epithelial cells

at the lateral surgical margins, and the presence or absence of dysplasia should be

mentioned in the histopathologic description of the lesion. The presence of squamous

metaplasia of the excretory ducts of the minor salivary glands should also be mentioned,

especially when cellular atypia is evident. Many treatments for oral precancers remove or

destroy only the most superficial portions of the submucosal tissues, thus leaving behind the

salivary glands and their ducts.17

As stated by Odell EW28 cellular pleomorphism, loss of basal cell polarity and

irregular stratification reflect a loss of the structural organization that is found in normal or

hyperplastic epithelia with a well defined basal layer, a structured prickle cell compartment

and a uniform superficial zone which may or may not be keratinized. These features also

need to be considered along with the subtle changes by more obvious features such as

premature or individual cell keratinization or deep squamous pearl formation. Cellular

pleomorphism implies differences from the cell size and shape appropriate for a given layer

in the epithelium and the presence of new or bizarre forms. Altered polarity, sometimes with


22

the appearance of spindle shaped cells and increased size unaccompanied by cell division

are all components of pleomorphism. Crowding of cells (‘hypercellularity’), usually basal or

in elongate or bulbous rete processes and either generalized or focal, is seen more frequently

than reduced inter cellular cohesion although these features may co-exist in the same

dysplastic lesion.

An increase in the proportion of the epithelium occupied by the proliferative

compartment is a common feature of dysplasia. Multiple layers of cells with basaloid

features, a high nuclear cytoplasmic ratio and mitoses high in the epithelium are

manifestations of this. Dysplasia in which these features predominate may have a

deceptively uniform appearance, as in atrophic epithelium such an expanded proliferation

compartment may occupy almost the full thickness. Bizarre mitoses and extremes of

anisonucleosis are rare in oral epithelia and should perhaps be reconsidered as a component

of dysplasia. Hyperchromatism of normal sized and shaped nuclei is common in

hyperplastic conditions.28

e. GRADING SYSTEMS FOR ORAL EPITHELIAL DYSPLASIA

Histopathological Grading of Oral Epithelial Dysplasia (OED)

Of all disciplines in clinical medicine, histopatholgy is often credited with being the

most scientific. There is no doubt that pathological examination has led to many of the

currently used disease classifications and that morphological observations and correlation of

the observations with clinical parameters has provided a sound basis for clinical medicine as

it is today. With the increasing demands for ‘evidence based medicine’, notoriously

subjective histopathological approaches do need to be redefined as concepts as well as

diagnostic criteria, scientifically validated. Grading, in fact, is an attempt to impose discrete

categories on what is in effect a continuous grey scale. By assigning a grade one artificially

creates discrete sub-entities in a biological continuum.29

There is a generally held view that assessment of dysplasia in premalignant lesions is

important because dysplastic lesions are more likely to undergo malignant change and

because it is felt that the chances of malignant transformation increases with increasing

severity of dysplasia. There is a broad measure of agreement among pathologists about the

altered epithelial features that comprise dysplasia, but there is great variation in interpreting

the severity of dysplasia and not all investigators describe the criteria they use.30


23

Pindborg JJ et al13 in 1985 during the II meeting the International Association of

Oral Pathologists (Amsterdam) held an exercise to study subjectivity in evaluating epithelial

dysplasia. A surprisingly wide spectrum of diagnostic suggestions was found indicating the

need for an internationally accepted set of criteria for OED.

The diagnosis of epithelial dysplasia has caused considerable distress for

pathologists because of ambiguous diagnostic criteria and difference of opinion among

pathologists about what constitutes “epithelial dysplasia”. Therefore Abbey LM et al31

studied the intra examiner and inter examiner reliability in diagnosis of OED. The exact

agreement (interobserver) with sign-out diagnosis was 50.5%, whereas examiners agreed

with their own diagnosis (intraobserver) 50.8%, of the time. This lead to the conclusion that

accurate reproducible agreement among experienced board certified oral pathologists is

difficult to achieve.

Karabulut A et al32 investigated observer’s variability in grading epithelial dysplasia

between two general pathologists and two oral pathologists for 100 sections of oral

leukoplakia to grade from no dysplasia to carcinoma in situ. The inter observer agreement

rates were in the range of 49-69%. The calculated kappa values 27%-45% showing poor to

moderate agreement between the pathologists. These values were also similar when

comparing the two pairs of pathologists with the same education, indicating that the inter

observer variability was due to individual differences rather than to their educational

background.

The ideal situation would be to create a system for gauging epithelial dysplasia, then

to follow the patients with dysplasia and study how often and how soon the dysplasia turns

into carcinoma. However, ethical considerations do not allow the follow-up, for longer

periods of time for moderate and severe dysplasia and carcinoma in situ. “We are in fact

faced by an insoluble problem : although we may be able to produce objective system to

gauge epithelial dysplasia, we will not be able to test it” stated Pindborg JJ et al.13

The severity of dysplastic features is designated as grade of epithelial dysplasia.

Many dysplastic features in varying combinations have been used for grading. However,

difficulties have been encountered in assessing and standardizing the different degrees of

epithelial dysplasia. Many systems of grading oral epithelial dysplasia have been proposed

in order to standardize the severity of dysplastic features and the relationship of epithelial


24

dysplasia in its varying grades to the subsequent development of cancers has still to be

worked out.33

Any grading system is said to be clinically useful if they are reproducible between

different observers. In addition, the parameters considered in the histological assessment

should be biologically meaningful, reflecting the malignant potential of the lesion.34

The various grading systems put forth by different authors are as follows:- 1. Smith

and Pindborgs ’ Photographic method (1969), 2. Banoczy and Csiba (1976), 3.WHO (1978),

4. Burkhardt and Maerkar (1981), 5. Lumermann H et al (1995), 6. Neville et al (1995), 7.

Speight P M et al (1996), 8. Bouquot JE (1997), 9.Soames JV (1998), 10.Kuffer and

Lombardi (2002), 11. Brothwell DJ (2003), 12. Takasi Saku Japanese society of Oral

Pathology 2004, 13. WHO (2005), 14. Binary System by Kujan O (2006).

1. Smith and Pindborg (1969)35 attempted to standardize the grading of dysplasia by

photographic method. They placed the diagnosis of epithelial dysplasia on an objective and

semi quantitative level by: (i) Concentrating the observer’s attention on one

photographically standardized microscopic feature at a time and (ii) Enabling the observer to

assess each feature individually and allocate a weighed score to each one.

The system was subjective involving the comparison of histological sections with a

series of standardized photographs. They used 13 histologic features which were

standardized by a set of photographs. Each feature was graded Absent, Slight and Marked as

follows:

Histologic features for Smith & Pindborg (1969) classification of oral epithelial dysplasia.35

Type of change Severity of dysplasia

1. Drop shaped retepegs None Slight Marked

2. Irregular epithelial stratification None Slight Marked

3. Keratinization of cells below keratinized layer None Slight Marked

4. Basal cell hyperplasia None Slight Marked

5. Loss of intercellular adherence None Slight Marked

6. Loss of polarity None Slight Marked

7. Hyperchromatic nuclei None Slight Marked

8. Increased nucleo-cytoplasmic ratio in basal andprickle cell layers

None SlightIncrease

MarkedIncrease


25

9. Anisocytosis and anisonucleosis None Slight Marked

10 Pleomorphic cells and nuclei None Slight Marked

11 Mitotic activity Normal SlightIncrease

MarkedIncrease

12 Level of mitotic activity Normal Slight Marked

13 Presence of bizarre mitoses None Slight Marked

Scoring: A grading of ‘none’ was scored – 0 (zero). Grading of ‘slight’ or ‘marked’ was

scored – 1 to 10. The total score of all the features was taken as epithelial dysplasia index

(EDI) which was a semi objective score. It could vary from 0-75. The grading finally was

done as follows: Total score (EDI) 0-10 : Not dysplastic, 11-25 : Mild dysplasia, 26-45 :

Moderate dysplasia and 46-75 - Severe dysplasia.35

Fishman SL et al36 studied clinico-pathologic correlation of oral premalignancy.

They scored 153 cases of premalignant lesions for various histologic characteristics

following the criteria suggested by Smith and Pindborg (1969). According to them, certain

histologic changes were found to be more prominent in lesions felt to be premalignant.

These changes progressed as the severity of diagnosis increased. These changes were :-

Irregular epithelial stratification, Intraepithelial keratinization, Loss of polarity of basal cells,

Hyperchromatic nuclei, Increased nucleo-cytoplasmic ratio and Anisocytosis and

anisonucleosis.

Thus using the criteria suggested by Smith and Pindborg (1969), the authors scored

histologic changes and developed a meaningful index wherein the histologic changes were

described in a quantitative manner. They concluded that this photographic method can be

used as routine process.36

Katz HC et al37 evaluated 214 cases of epithelial dysplasia using Smith and Pindborg

(1969) method of standardization. They found the system to be of considerable value for

purposes of standardization and eliminated observer bias by the use of standardized

photographs. But they questioned the accuracy of the weightage given to each of the

histological characteristics. They suggested testing it further, as to which histological criteria

was of greatest value in predicting the potential for malignant change, though they found

that 98% of cases graded as severe dysplasia showed presence of superficial mitosis

suggesting that it could be an ominous feature.


26

This system despite providing more objective data, could not find general favor

among pathologists and in routine diagnostic pathology as quantification was tedious.

Warnakulasuriya S (2001)17 commented on this system and noted that even inflammatory or

reactive lesions which are considered non-neoplastic may show some features of dysplasia.23

2. Banoczy and Sciba (1976)38 diagnosed epithelial dysplasia using the following criteria

suggested by Mehta et al (1971) : - Irregular epithelial stratification, Increased density of the

basal cell layer or prickle cell layer or both, Increased number of mitotic figures (a new

abnormal mitoses may be present), Dropped shaped retepegs, Increased nuclear cytoplasmic

ratio, Loss of intercellular adherence, Nuclear pleomorphism, Nuclear hyperchromatism,

Enlarged nucleoli, Keratinization of single cells or cell groups in the prickle cell layer, Loss

of intercellular adherence.

They graded epithelial dysplasia as:- Mild - When 2 of the above listed histological

changes were present. Moderate - When 2 to 4 changes were present. Severe - When 5 or

more of the changes were present. They found that out of 120 cases of epithelial dysplasia

analyzed, the most frequently appearing histological features were: a. Irregular epithelial

stratification. b. Drop shaped retepegs. c. Hyperplasia of the basal cell layer.

Based on their classification of dysplasia, they described the distribution of

leukoplakia lesions by age, sex, clinical types of leukoplakia, location and follow up results.

They correlated these factors with the grade of dysplasia. This system was based on

subjective interpretation of the features and didn’t take into account which factor could be

important in determining the malignant potential.38

3. WHO System (1978): In an attempt to standardize the criteria for oral precancer,

established a collaborating reference center in 1967. The center aimed to characterize and

define those lesions that should be considered as oral precancer and to determine, if possible

their relative risk of becoming malignant. In its report in 1978, WHO defined and listed the

12 histologic characteristics of epithelial dysplasia as follows :- 1) Loss of polarity of basal

cells. 2) The presence of more than one layer of cells having basaloid appearance-Basal cell

hyperplasia. 3) An increased nuclear-cytoplasmic ratio. 4) Drop shaped retepegs. 5)

Irregular epithelial stratification. 6) Increased number of mitotic figures. 7) The presence of

mitotic figures in the superficial half of the epithelium. 8) Cellular polymorphism. 9)

Nuclear hyperchromatism. 10) Enlarged nucleoli. 11) Reduction of cellular cohesion. 12)

Keratinization of single cells or cell groups in the prickle cell layer.37


27

WHO 1978 graded epithelial dysplasia as: Mild, Moderate, Severe : -

Mild dysplasia: Slight nuclear abnormalities, most marked in the basal third of the

epithelial thickness and minimal in the upper layers, where the cells show maturation and

stratification. A few, but no abnormal mitosis may be present, usually accompanied by

keratosis and chronic inflammation.

Moderate dysplasia: More marked nuclear abnormalities and nucleoli tend to be present,

with changes most marked in the basal 2/3rd of the epithelium, nuclear abnormalities may

persists up to the surface, but cell maturation and stratification are evident in the upper

layers. Mitoses are present in the parabasal and intermediate layers, but none is abnormal.

Severe dysplasia: Marked nuclear abnormalities and loss of maturation involving more than

2/3rds of the epithelium, with some stratification of the most superficial layers. Abnormal

mitoses may be present in the upper layers.

Severe grades of dysplasia may merge into the lesion customarily designated as

carcinoma in situ, in which the whole or almost the whole thickness of epithelium is

involved. In the present knowledge, it was not possible to say whether the presence of severe

dysplasia carried a different degree of risk of subsequent development of invasive carcinoma

than the presence of carcinoma in situ.

It was generally believed that mild degrees of epithelial dysplasia did not indicate

any great danger for the patient, although, special reference had to be made to certain high

risk sites such as the floor of the mouth and the ventral surface of the tongue. Moderate

dysplasia, however called for a more cautious approach and severe dysplasia indicated that

there was a considerable risk of the development of cancer. They suggested that generally

the degree of dysplasia could be linked to the degree of probability of development of

malignancy.37

4. Burkhardt and Maerkar (1981)39 listed 6 relevant histological and cytological

parameters, based on which diagnosis and classification of epithelial dysplasia could be

made: 1) Basal cell hyperplasia. 2) Loss of basal cell polarity. 3) Cellular pleomorphism. 4)

An increase in mitotic figures. 5) Dyskeratosis. 6) Abnormal and absent epithelial

stratification. Additional indicators for dysplasia were i) An increase in sub epithelial

lymphocytes, plasma cells ii) intraepithelial cells (stroma reaction) and iii) Presence of

Candida organisms.


28

They graded dysplastic criteria for classification according to degree of dysplasia and

characteristics of carcinoma in situ.

Degree Characteristics

Low Basal cell hyperplasia.

Basal cell polarity disrupted.

Medium Basal cell hyperplasia.

Loss of basal cell polarity.

Moderate degree of cellular polymorphism.

Slight increase in rate of mitosis.

Occasional dyskeratosis.

High Basal cell hyperplasia.

Basal cell polarity lost.

Marked cellular pleomorphism.

Increase in ratio of mitosis.

Numerous dyskeratosis

Abnormal epithelial stratification.

Ca-in

situ

Characteristics of high degree dysplasia more marked.

Epithelial stratification lost.

Stroma not yet invaded.

They classified leukoplakia lesions into following groups:

a) Leukoplakia without or with low degree dysplasia: These lesions showed no signs of

dysplasia. Those with minor dysplastic changes were included in this group as low degree

dysplasia could also represent a completely harmless reversible epithelial irritation. This

group of lesions was not considered as precancerous.

b) Leukoplakia with moderate degree of dysplasia: They held an intermediate position,

having more marked dysplastic criteria. They had to be considered as risk lesions when

planning treatment.


29

c) Leukoplakia with a high degree of dysplasia and carcinoma in situ : These had to be

considered as precancerous and were identified by the fact that there was a marked degree of

dysplasia. Endophytic growth and downward extension of epithelial retepegs were common

findings.

d) Carcinoma-in situ: Was characterized by the additional feature of complete loss of

epithelial stratification. It was regarded as early form of cancer not showing invasive

growth.

e) Candidal leukoplakia: The incidence of candidiasis increased with the degree of

dysplasia. Fungal growth in the lesion had to be regarded both as risk factor and risk

indicator.

In their study, the degree of dysplasia correlated with the clinical parameters that

were known to relate to relatively poor prognosis like age, sex, location, the dental status,

and exogenous irritants. Patients exhibiting moderate to high degrees of dysplasia developed

malignancy. The authors proposed a treatment and follow up plan for different degrees of

epithelial dysplasia.

For no/low degree of dysplasia : i. Aim at total excision, ii. If excision total, no follow up

observation and iii. If excision incomplete, further observation at longer intervals (3-6

months).

For medium degree dysplasia : i. aim for total excision, ii. if removal incomplete initially,

total excision to follow, iii. with total excision, no follow up observation, iv. if excision

incomplete, further observation at shorter interval (4-8 weeks).

For marked dysplasia (carcinoma in situ) : i. total excision in every case, ii. follow up as

for cancer patients (every 4 weeks during the first year) and iii. In addition to this, Candida

infection requires antifungal therapy.39

Girod SC et al (1999)40 in their study classified the epithelial dysplastic lesions

according to Burkhardt and Maerkar (1981)39 and studied the expression of p53, PCNA and

Ki-67 in different degrees of epithelial dysplasia. The expression pattern correlated better

with the different degrees of dysplasia.

5. Lumermann H. et al (1995)25 considered: Basal cell hyperplasia, -Nuclear enlargement

and hyperchromaticity, drop shaped retepegs as ‘minimal’ criteria for the diagnosis of oral

epithelial dysplasia. The dysplastic changes were graded as:-


30

1. Mild epithelial dysplasia: ‘Minimal’ dysplastic alterations confined to the lower third

of the epithelium.

2. Moderate epithelial dysplasia: Dysplastic changes seen in upto 2/3rds of the thickness

of the epithelium.

3. Severe epithelial dysplasia: Dysplastic cells fill more than 2/3rds but less than the

entire thickness of the epithelium.

4. Carcinoma in situ: The entire thickness of the epithelium contains less differentiated

basaloid or squamous epithelial cell with enlarged, hyperchromatic nuclei and a variable

number of typical and atypical mitotic figures with no invasion into the submucosa.

5. Verrucous hyperplasia with dysplasia: The epithelium exhibits considerable

thickening with surface papillations, hyperparakeratoses and parakeratin plugging and

occasional dysplastic cells confined to the lower 1/3 rd of the epithelium.

After the microscopic analysis and grading of the dysplasia, each case was correlated

with the clinical data. The authors described the distribution of the lesions according to age,

sex, site and follow up. 16% of patients with epithelial dysplasia showed invasive squamous

cell carcinoma within a mean period of 36.6 months. Interesting feature was, 2 of the 11

patients with histological verrucous hyperplasia with mild dysplasia developed squamous

cell carcinoma, indicating the serious nature of this variant of epithelial dysplasia.

Their study confirmed that the dysplastic changes are multicentric, with presence of

normal or hyperplastic epithelium in the skip areas of dysplastic epithelium. The grade of

dysplasia was different in various portions of the same biopsy specimen suggesting that the

incisional biopsy samples were not representative of the true nature of the lesion. This

implied that the examination of the entire lesion was necessary for accurate grading of

dysplasia.27

6. Neville BW et al (1995)41 graded dysplasia as:

Mild: Hyperchromatic and slightly pleomorphic nuclei are noted in the basal and suprabasal

cell layers of stratified squamous epithelium.

Moderate: Dysplastic changes extend from the basal layer to the mid portion of the spinous

layer and are characterized by nuclear hyperchromatism, pleomorphism, cellular crowding

and hyperkeratosis of the epithelial cell layer along with prominent granular cell layer.


31

Severe: Cellular crowding and disordered arrangement throughout most of the epithelial

thickness, although slight maturation and flattening of the cells appears to be present at the

epithelial surface. Epithelial cells are less matured as they progress toward the

hyperparakeratotic surface.

Carcinoma in situ: When the entire thickness of the epithelium is involved, the term

carcinoma in situ is used. Dysplastic cells extend from the basal layer to the surface of the

mucosa (top to bottom change) with no invasion into the underlying connective tissue.28

7. Speight PM et al (1996)42 considered the thickness (height) to which the cellular and

tissue changes may extend, as important in grading dysplasia. According to them : a. Mild

forms of dysplasia represented recognizable changes limited to the parabasal layers (lower

third), b. Moderate dysplasia represented recognizable changes extending to middle third,

and c. Severe dysplasia represented as recognizable changes extending to the upper layers.

However, Warnakulasuriya S. 2001 commented that there was wide variation in the

thickness of the covering epithelium in the oral cavity, with much undulation it led to

practical difficulties in using this grading system.

8. Soames JV (1998)27 The features used for evaluating the grades of dysplasia have been

summarized as follows : Mild dysplasia : Epithelial dysplasia involving the lower third of

the epithelium showing pleomorphism, suprabasal mitosis and loss of ordered stratification.

Moderate dysplasia : Epithelial dysplasia involving about 2/3rd of the epithelium Severe

dysplasia : Showing marked cellular atypia and disturbed maturation of the epithelium.

9. Kuffer and Lombardi (2002)43 felt that the choice of clinical rather than histological

criteria in the diagnosis and terminology of precancer is the cause of a disorderly mixture of

dysplastic and non-dysplastic lesions. Therefore, they proposed to dismember the classical

“oral precancerous lesions” to classify all cases which histologically do not show dysplasia

into the category of “risk lesions”(ex. Simple tobacco keratosis) and to place lesions with

dysplasia (i.e. already engaged in the process of malignant transformation ) into the category

of “precursors” of squamous cell carcinoma(ex; tobacco keratosis with dysplasia). This

“precursor” term seems to be the most accurate to characterize the limited but already

malignant intraepithelial alterations of dysplasia and carcinoma in situ, which herald the

onset of an invasive squamous cell carcinoma.


32

Drawbacks:

1. As there was considerable difference in potential for transformation between lesions

without dysplasia or with mild to moderate dysplasia and those with severe dysplasia, the

application of the term “risk lesion” to lesions without dysplasia which have a “zero risk” of

transformation (ex. Frictional keratosis) was inappropriate. It could lead dentist with less

expertise in this area to exaggerate the risks posed by the lesions with little or no possibility

of developing into cancer.

2. The use of the term “precursor of oral squamous cell carcinoma” to denominate dysplastic

lesions suggested that they were unequivocally associated with the future development of

cancer. This had no scientific evidence. On the contrary, as demonstrated by Mincer et al

(1972), 20 % of oral dysplasia regressed and 40% showed no modification in severity.

According to Gupta et al (1980), 13% of cases regressed and 40 % showed no modification

in severity.44

The concept of dysplasia had been criticized among gynecologists and pathologists

because of possibility of confusion between mild dysplasia and squamous metaplasia or

changes due to epithelial repairs and still worse because of the difficulty and lack of

consistency to differentiate between severe dysplasia and carcinoma in situ. The fact that

“dysplasia” considered as “premalignant” was usually treated conservatively and

“carcinoma in situ” considered as “malignant” was surgically treated, was criticized by

Richart RM and he demonstrated that dysplasia and carcinoma in situ were different aspects

of the same disease “Cervical intraepithelial neoplasia (CIN)” and treatment should be

same for both. This concept of CIN (one or more clones of transformed cells slowly

replacing normal keratinocytes starting from basal and parabasal layers to progressively

invade the whole epithelial height) has now replaced almost completely that of cervical

dysplasia. It has been extended with some modification to oral mucosa as “oral intra

epithelial neoplasia” (OIN) and in general as squamous intraepithelial neoplasia (SIN)

as described by Gnepp GM.45

As for CIN, there are 3 grades of OIN are as follows : - OIN 1: Mild dysplasia- less than

1/3rd involvement of the epithelium, OIN 2: Moderate dysplasia- 1/3rd to 2/3rd involvement

and OIN 3: Severe dysplasia – full thickness involvement or equivalent to carcinoma in

situ.


33

This system uses OIN terminology to emphasize that all levels of dysplasia are

actually different levels of true neoplasia, albeit prior to invasion. This system is largely

based on subjective interpretation and lacks consistency in diagnosis among pathologists.

The “Bethesda classification” for cervical cytopathology, includes only 2 grades : - 1.

“Low grade squamous intraepithelial lesion (LSIL)” corresponding to CIN 1 and 2. “High

grade squamous intraepithelial lesion (HSIL)” corresponding to CIN2, CIN3.

Based on this “Bethesda classification” the former system with 3 grades for OIN

was replaced by a 2 grade system, which helped in better stratifying patients for clinical

protocols. Accordingly they chose to report the diagnosis of oral dysplastic lesions as :

1. Low grade oral intraepithelial neoplasia (LOIN) - including OIN 1 (mild dysplasia) or as

2. High grade oral intraepithelial neoplasia (HOIN) - including both OIN 2(moderate) and

OIN 3 (severe dysplasia).

According to the authors, the regression of oral dysplasia without any further

treatments, other than biopsy and cessation of smoking, were difficult to prove. A few cases

of low grade OIN regressed but not high grade OIN.

Drawback: It was inappropriate to consider OIN2 as high grade precursor lesions alongside

severe dysplasia and carcinoma in situ, as they had markedly lower potential for malignant

change.44

10. Brothwell DJ et al (2003): In an attempt to determine the extent of observer agreement

in diagnosis of oral epithelial dysplasia, they graded 64 sections of epithelial dysplastic

lesions according to 5 point scale routinely utilized at their institution (Faculty of dentistry,

University of Toronto). The criteria were: 0 = No dysplasia, 1 = Mild dysplasia: Increased

number of cells in the basal and parabasal epithelial regions showing nuclear

hyperchromatism and pleomorphism, 2 = Moderate dysplasia: Bulbous retepegs with

increased number of cells showing nuclear hyperchromatism and pleomorphism,

involving the basal, parabasal and prickle cell layer. 3 = Severe dysplasia: Bulbous retepegs

with increased numbers of cells showing nuclear hyperchromatism and pleomorphism

involving the entire thickness of epithelium. 4 = Carcinoma in situ: Markedly atypical

changes showing nuclear hyperchromatism and pleomorphism and encompassing the entire

thickness of the epithelium, with the suggestion of early superficial connective tissue

invasion, but without convincing evidence.


34

Using this system, and with statistical analysis, the authors proved that intra and inter

observer agreement in grading the dysplastic lesions was substantial to an almost perfect

range. This study emphasized the use of appropriate statistical methods and calibration to

eliminate the bias in assessment of oral epithelial dysplasia.46

11. Takashi Saku et al47 proposed a new concept for histopathology of oral borderline

lesions. This group classified dysplasia as (i) Mild (ii) Moderate (true) dysplasia (iii)

Carcinoma in situ. (i) Mild epithelial dysplasia is applied to lesions with atypical features

whose potential for malignant transformation is not known. (ii) Moderate dysplasia is

regarded as true dysplasia which has carcinomatous potential. They also describe the term

epithelial hyperplasia as hyperplastic epithelial lesions with normal shape of rete processes

and normal cellular stratification. They suggested that basal cells are neither germinal cells

nor a source for squamous epithelial regeneration but are terminally differentiated cells.

Most of the cells with proliferative potential are present in the parabasal cell layer, which

may be regarded as the germinal center of the oral squamous epithelium.

When there is a focal proliferation of basaloid (parabasal) cells in the lower part of

rete processes, they replace not only the basal cells but also the lower prickle cells. Such

proliferating cell foci make a straight interface with parakeratinsed cells in the upper layer,

which shows distinct contrast between the upper and lower layers of the epithelium. This

sudden transition is caused by loss of prickle cells due to apoptosis. This phase is referred to

as “two phase” appearance. When any traces of the two phase appearance are recognized,

the lesion is diagnosed as moderate (true) dysplasia.

(iii) Carcinoma in situ is defined as true but not yet invasive neoplasm staying within an

epithelial layer : It is an advanced stage of moderate (true) dysplasia. In moderate dysplasia

the cells are only proliferating, whereas in carcinoma in situ the cells differentiated as

keratinocytes can be recognized. These are basal cell like cells referred to as ‘basal cell

minics’. Dyskeratotic cells in lower layers are important in diagnosing carcinoma in situ.

This working group classified carcinoma in situ into 4 histological types : basaloid,

verrucous, acantholtic and atrophic. In basaloid type most of the constituent cells are basloid

cells replacing the whole layers of epithelium. The verrucous type is characterized by round

shaped reteridges with gradual differentiation towards keratinization and enhanced

keratinsation in surface with keratin plugs. The acanthotic type is the most difficult subtype


35

because it resembles simple epithelial hyperplasia. There will be distinctive keratinization

on the surface , the reteridges shape are slightly indented, individual cell dyskeratosis or

sudden keratinsation occurs in the lower layers of rete processes, which has no distinct

alignment of basal cell mimics. In the atrophic type the surface is flat and lacks the

hyperkeratotic layer. Rete processes are small and basal cell mimics are clearly aligned

along the basement membrane. Cell cohension appears loosened within the rete processes

with dense band of lymphocytic infiltration.47

12. WHO Classification (2005)48 A working group that convened for an editorial and

consensus conference in Lyon, France in July 16-19-2005 put forward criteria for grading

epithelial dysplasia which is published as WHO classification(2005).They have used a

combination of architectural & cytological changes with more explicit consideration of

levels of changes within the epithelium. Criteria used were in two groups

Architecture Criteria : 1. Irregular epithelial stratification, 2. Loss of polarity of basal

cells, 3. Drop-shaped rete ridges, 4. Increased number of mitotic figures, 5. Abnormal

superficial mitoses, 6. Premature keratinisation in single cells (dyskeratosis), 7. Keratin

pearls within rete ridges,

Cytology Criteria : 1. Abnormal variation in nuclear size (Anisonucleosis), 2. Abnormal

variation in nuclear shape (nuclear pleomorphism), 3. Abnormal variation in cell size

(anisocytosis), 4. Abnormal variation in cell shape (cellular pleomorphism), 5. Increased

nuclear cytoplasmic ratio, 6. Increased nuclear size, 7. Atypical mitotic figures, 8. Increased

number and size of nucleoli, 9. Hyperchromatism.

Grading of dysplasia

Mild dysplasia : In general architectural disturbance limited to the lower third of the

epithelium accompanied by cytological atypia define the minimum criteria of dysplasia.

Moderate dysplasia : Architectural disturbance extending into the middle third of the

epithelium is the initial criterion for recognizing this category. However, consideration of

the degree of cytologic atypia may require upgrading i. e. those lesions that show marked

cytological alteration should be elevated to a higher grade level regardless of how

extensively the atypical cells fill the epithelium.

Severe dysplasia : Recognition of severe dysplasia starts with greater than two thirds of the

epithelium showing architectural disturbance with associated cytologic atypia. However,


36

architectural disturbance extending into the middle 3rd of the epithelium with sufficient

cytologic atypia may be up graded from moderate to sever dysplasia.

Carcinoma in situ : The theoretical concept of carcinoma in situ is that malignant

transformation has occurred but invasion is not present. It is not always possible to

recognize this morphologically. The following is recommended for the diagnosis of

carcinoma in situ: full thickness or almost full thickness architectural abnormalities in the

viable cellular layers accompanied by pronounced cytologic atypia. Atypical mitotic figures

and abnormal superficial mitoses are commonly seen in carcinoma in-situ.48

13. Binary System of Grading Dysplasia - Omar Kujan (2006)49 : Proposed a new binary

system of grading oral epithelial dysplasia for prediction of malignant transformation. They

used same morphological criteria used by the WHO classification (2005) and graded the

lesion into low- risk or high-risk based on scoring the features.

High-risk lesion (with potential susceptibility for malignant transformation) was based on

observing at least four architectural changes and five cytological changes.

Low-risk lesion (does not have the potential susceptibility for malignant transformation)

was based on observing less then four architectural changes or less than five cytological

changes.

They conducted a pilot study of 28 cases & they could predict the clinical out comes

with certainty in 84.9% of cases i.e. cases given as high risk progressed to carcinoma. The

negative value was also as high as 85% with only 5 cases of low risk category going for

malignancy. In comparison with WHO classification, where prognostic implications for

grading a case as moderate dysplasia were problematic, the binary system was more

objective by classifying them as low risk or high risk to predict the clinical outcome.

Thus they suggested that the new binary system complements the WHO 2005

classification and that it may help clinicians to make critical decisions especially regarding

cases of moderate dysplasia.

A study conducted by Omar Kujan et al (2007)49 showed low inter observer

variability in grading of oral epithelial dysplasia but good to substantial agreement was seen

when a cumulative scoring of these features was used to grade degree of dysplasia. The

efficacy of this system needs to be evaluated further.


37

Malignant transformation rates for oral epithelial dysplasia as quoted in various

studies described in the article by Bouquot JE et al50

Sl.no Reference Country

No. ofpatients

Withdysplasia

Withmalignancy

Followup Malignancy

1 Gupta et al (1980) India 10.5 Yrs 7%2 Schepman et al 1998 Netherlands 166 166 20 8.5 Yrs 8%3 Silverman et al 1984) USA 257 22 8 7.2 Yrs 36%4 Banoczy & Csiba (1976) Hungary 68 68 9 6.3 Yrs 13.2%5 Amagasa et al (1985) Japan 12 - - 10 Yrs 50%6 Vedtofte (1987) Denmark 14 - - 3.9 Yrs 35.7%7 Mincer et al (1972) USA 45 45 5 3 Yrs 11%8 Pindborg et al (1995) India 61 61 4 7 Yrs 6.6%9 Lumerman etal (1995) USA 44 44 7 1.5 Yrs 16%10 Jaber et al (2003) England UK 359 630 20 3.3 Yrs 5.6%11 Cowan et al (2001) UK Ireland 745 165 24 ? 14.5%12 Burkhardt & Maerker (1981) Germany 200 200 21 5.5 Yrs 10.5%

The above table demonstrates that dysplastic lesions are more likely to progress to

oral squamous cell carcinoma. The cancer risk in moderate to severe dysplasia appears to be

at least 2-3 times more than that in mild dysplasia or hyperplasia. This suggests that

whatever may be the inter or intra observer variability in diagnosis of oral epithelial

dysplasia, it still remains the gold standard and is an important predictor of potential

malignant transformation.50

Interestingly, in 1988 the “Bethesda classification” for cervical cytopathology,

included only two grades.45 According to this, lesions would be classified as low-grade

squamous epithelial lesions, corresponding to former cervical intraepithelial neoplasia

grade 1, and high-grade squamous epithelial lesions, corresponding to grades 2 and 3. This

system has also been mentioned in some reports for oral lesions.40

After the publication of those papers on the binary system for grading epithelial

dysplasia in oral leukoplakia( OL), a study was performed with 218 patients with OL, from

which 39 (17.9%) developed into cancer.44 The authors reported that high-risk OL was

associated with a 4.57-fold increased risk for malignant transformation, compared with low-

risk OL. The authors suggested that high-risk dysplasia would be a significant indicator for

evaluating malignant transformation risk in OL. Subsequently, the same research group

published a study in which they identified significant risk factors for malignant

transformation in a long term follow-up cohort of patients with oral epithelial dysplasia. Of

the 138 patients in this study of Maria AVDC51 with histologically confirmed oral dysplasia,


38

115 had OL and 23 had lichen planus. From these 138 lesions, 37 (26.8%) developed into

cancer and the “high-risk” degree of dysplasia was an independent risk factor for

transformation. Moreover, high-risk degree of dysplasia was associated with a 2.78-fold

increased risk of transformation compared with low-risk degree. The authors then suggest

the utilization of high-risk dysplasia as a significant indicator for evaluating malignant

transformation risk in patients with potentially malignant lesions. According to them, this

would also help guiding treatment in clinical practice. In spite of these great achievements, it

must be mentioned that malignancy also developed in some patients previously presenting

low-risk potentially malignant lesions.51

14. Grading of dysplasia Bouqout JE17 : Most pathologists grade oral epithelial dysplasia

according to a combination of cellular and tissue changes. Low-grade lesions are

considered to be less severe than high-grade lesions but this assessment is a subjective one

and it is not unusual for different pathologists to place emphasis on different aspects of

dysplastic alteration, thereby arriving at somewhat different diagnostic conclusions.

Regardless of individual emphasis, however, certain oral precancer grading criteria are in

general use:

a. the cellular atypia or dysplasia is similar to that seen in squamous cell carcinoma, b. there

is no evidence of invasion into underlying stroma (the diagnosis would then change to

carcinoma), c. the epithelium with the greatest proportion of atypical cells has the greatest

risk of being or becoming a carcinoma, d. the epithelium with the most extreme atypia of

cells has the greatest risk of being or becoming a carcinoma, e. the final grading or diagnosis

should be based on the most severely involved area of change, even if that area includes no

more than a few rete processes.

Oral epithelial dysplasia is subdivided into three prognostically significant

categories, mild, moderate and severe. Mild (grade I) dysplasia demonstrates proliferation

of atypical or immature basal cells above the parabasal region but not extending beyond the

lower third of the epithelium. Moderate (grade II) dysplasia demonstrates a similar

proliferation into the middle one-third of the epithelial cross-section. The term severe

(grade HI) dysplasia is reserved for abnormal proliferation from the basal layer into the

upper third of the epithelium. Carcinoma in situ, thought by some to be a premalignancy

and by others to be an actual pre-invasive malignancy, requires top-to-bottom change with


39

undifferentiated, primitive cells from the basal layer to the topmost layer, although most

authorities accept a thin layer of surface keratin with abrupt transition into the underlying

atypical cells. The grading of dysplastic oral epithelium must also take into consideration

the degree of cellular atypia, as described above, and those lesions with marked alteration

are elevated into a higher grade level.

It is difficult to determine the amount of basal cell hyperplasia in oral precancers

because of transverse sectioning of tissue samples and because of the natural undulation of

the inferior margin of squamous epithelium. It may be helpful to make this determination

from the top of the connective tissue papillae because of the great variability of rete process

length in acanthomatous epithelium, but there is no established standard and in many cases

the basal hyperplasia is confined to the rete processes themselves.17

f. SUBJECTIVITY / VARIABILITY IN ORAL EPITHELIAL DYSPLASIA:

The topic of subjectivity in assessing & grading oral epithelial dysplasia has often

been raised and is a cause of considerable distress among pathologists. This is largely due to

lack of well defined criteria that can be used as guidelines for grading and differences in

opinion among pathologists about what actually constitutes epithelial dysplasia. In addition

there is still a cloud over what criteria could best describe the malignant potential of oral

epithelial dysplasia. The wide intra and inter observer variability encountered in grading oral

epithelial dysplasia was reported by several authors.

At the 2nd meeting of the International Association of Oral Pathologists in 1984, an

exercise was held to study subjectivity in evaluating oral epithelial dysplasia, carcinoma in-

situ and initial squamous cell carcinoma. A poster with a color photomicrographs was

presented to be evaluated anonymously by the 72 participants. The authors diagnoses were

based on the criteria described in the report of the WHO (1978)23 international collaborating

centre for oral precancerous lesions, which ranged from mild epithelial dysplasia to early

squamous cell carcinoma. In one case 37% of the participants agreed with the original

diagnosis, whereas the best rate of agreement was 78% in another photograph (Pindborg,

1985).13 The agreement thus ranged between 1-78%. Fischer D et al (2004)52 discern

regarding this study that the data did not consider the seriousness of diagnostic

disagreements nor were the chance agreements taken into account analytically and all the

participants were not experienced oral pathologists.


40

Abbey LM et al (1995)31 evaluated inter examiner and intra examiner reliability to

objectively assess the degree of difficulty in diagnosing epithelial dysplasia. 6 oral

pathologists examined 120 oral biopsies exhibiting simple hyperkeratosis to severe

dysplasia. Exact agreement with the original sign-out diagnoses averaged 50.5%. The intra

examiner agreement was 50.8%. Agreement distinguishing dysplasia from no dysplasia

compared with original diagnosis was 80.3%. Reproducibility of the diagnosis by the same

pathologists was also a problem but was reported as 73.3%.

The results supported the view point that diagnosing epithelial dysplasia of oral

cavity is not an exact science. The authors opined that the pathologist’s diagnosis is a

necessary part of the process of treating a patient even though variation in opinion exists

among pathologists. They emphasized that the pathologist may make error in diagnosis of

epithelial dysplasia versus no dysplasia, one in 5 times when morphologic characteristics

alone are used. Therefore they mentioned that relevant clinical factors may play a role and

clinician should weigh all there factors & then arrive at a management strategy.

Karabult et al (1995)32 investigated the extent of agreement in grading epithelial

dysplasia between pathologists with the same or different educational background. 2 of the

examiners were oral and maxillofacial pathologists with dental background and the other 2

were general pathologists with medical background.

The degree of agreement ranged from 27% to 45 % (0.27-0.45) and was judged as

poor to moderate and it didn’t vary among pathologists of different backgrounds. The

authors concluded that inter examiner variability was due to individual disagreement and not

due to educational background. This variability may depend on which histologic and

cytologic characteristic are considered important for diagnosis, on the variability in the

observation of there characteristic and on the variability of grading there characters into

different categories of epithelial dysplasia, According to them, the problems arise due to

lack of sufficient knowledge of which criteria are important for predicting future

development of cancer, lack of objectivity in evaluating the established criteria, lack of

calibration of the method and arbitrary system of grading.

In a study by Abbey LM et al (1995)31, it was tested whether clinical information

could modify the diagnosis of epithelial dysplasia. But only one of the 6 pathologists came

close to reproducing the results he had achieved in the initial study. So the clinical


41

information didn’t improve the accuracy and consistency among pathologists. In fact, there

was a decrease in accuracy.

Lumermann H et al (1995)25 used their histologic criteria to grade 308 cases in their

study to define the rate of development of invasive squamous cell carcinoma in dysplastic

lesions. They compared the independent grading of 2 experienced pathologists and found

that inter examiner agreement was 54% only.

Sudbo J et al (2001)53 assessed inter examiner variability by reporting 196 lesions

which were followed up for a period of 4 to 165 months (average 103 months). All

histological sections were reviewed by 4 separate pathologists working at 3 different

institutions. The sections were graded according to WHO criteria as mild, moderate, severe

and carcinoma in situ. The calculated kappa values were in the range of 0.17-0.33,

confirming poor agreement among the scorers. To improve the agreement, they considered

further only 2 diagnostic groups as: 1. Low grade (mild dysplasia and moderate dysplasia) 2.

High grade (severe dysplasia) constituting a presumptive favorable and poor prognostic

group, respectively. Even then the agreement among pathologists didn’t improve (Kappa

values ranged 0.21-0.34). The design and statistical analysis used by previous studies were

criticized by Brothwell DJ et al (2003).46 They described a different method of statistical

analysis in assessing observer agreement and proved that there was substantial intra-and

inter observer consistency and almost perfect conformity in the grading of oral epithelial

dysplasia. 64 histologic sections were examined by 3 oral pathologists and the original sign-

out diagnoses were used as the equivalent of a 4th examiner for comparison and analysis

purposes. Dysplasia was graded according to their institutional criteria on a 5 point scale.

Inter and intra examiner agreement was assessed by using % agreement, Ks and Kw using

quadratic weighting system. The Kw values were 0.68-0.83 and 0.82-0.96 suggesting

substantial to almost perfect range of inter observer agreement and intra observer agreement

respectively. There was slight inter observer bias and this could be due to different opinion

considering the important criteria in grading dysplasia. This suggested that variation in

assessment of dysplasia might decrease if appropriate statistical method is used.

Substantial variation has been reported in the grading of OED. The design and

statistical analysis used by these studies has not been adequate to allow determination of the

extent and source of the variation. The accuracy and precision of a measurement or


42

diagnosis is normally evaluated by looking at the validity and reproducibility of examiner

observations, respectively. Validity refers to the degree to which examiner observations

represent the actual situation of interest. To assess validity, an observation is compared with

a ‘gold standard’, which is considered to best represent the truth. Observer accuracy is then

assessed by examining the degree of conformity with the gold standard.

Reproducibility refers to the degree to which observer measurement or diagnosis

remains the same on repeated independent observations of an unchanged characteristic and

considers the consistency obtained on two or more observations of the same subject. This

consistency can be assessed between different observers (inter observer) or within a single

observer (intra observer).

When studying the accuracy obtained in grading OED, the situation is slightly

different. There is no test available, which is thought to be better than the pathologist’s

observation; an accepted gold standard is not available for assessing the validity obtained

when grading OED. Therefore, reproducibility, normally used to assess precision, is used to

provide an indication of validity. When combined for this purpose, inter and intra observer

agreement levels give an estimate of the degree of bias and validity present in situations like

grading OED, where an appropriate gold standard is not available.46

A study by Fischer et al (2004)52 achieved a fair to good agreement with overall

Kappa values of 0.59 diagnosing just presence or absence of dysplasia which they believe is

higher than that obtained in other studies reported. When the histologic diagnosis was

simplified into three general categories of ‘no abnormality / hyperplasia’, ‘mild / moderate /

severe dysplasia’ & ‘carcinoma in situ / carcinoma’, the Kappa value improved and was

around 0.7. They emphasized that presence of inflammation may modify the reliability of

diagnosis as it may induce reactive atypical/dysplasia or may reduce the pathologists ability

to observe the dysplastic change leading to under reporting of lesions exhibiting

inflammation. The limitation of this study was the small sample size, differing educational

background of the pathologists, use of 2 sets of slides which could not always be identical.

They suggest combination of border line lesions into one category to improve the

diagnostic accuracy as has been suggested in the ‘Bethesda system’ for cervical

intraepithelial neoplasia.


43

In a study by Kujan O et al (2006)49 three oral pathologists & one general pathologist

examined 68 oral epithelial dysplastic lesions & scored them according to WHO 2005

criteria. The features of each of these cases were then correlated with clinical outcome to

understand which were more commonly associated with malignant transformation. There

was poor to moderate agreement in grading the individual architectural & cytological

characteristics but a good to substantial agreement was seen when cumulative scoring of

features were used to agree on the degree of dysplasia.

In a study by Warnakulasuriya S et al (2008)54 they proposed to adapt a binary

system to reduce inter and intra observer variability. They suggested that the highest

agreement scores were found for : increased number of mitotic figures, drop shape rete

ridges, increased nuclear size and abnormal variation in cell shape. On the other hand,

irregular epithelial stratification, loss of polarity of basal cells, abnormal variation in nuclear

size, atypical mitotic figures and hyperchromatism had high disagreement scores.

Considering the problems in making reliable distinctions between the different

grades the Working Group considered collapsing the four grades to two when reporting the

presence or absence of epithelial dysplasia: ‘no/questionable / mild’ as low risk, ‘moderate

or severe’ implying high risk. If mild dysplasia is used as the cutoff point when deciding

whether or not to remove the lesion surgically, near misses of grading are not particularly

relevant in this context, and the variability could be significantly reduced by having two

grades. The working group considered the two class classification and was of the view that

reducing the number of choices from 3 to 2 may increase the likelihood of agreement

between pathologists. The utility of this was recently tested and has been shown to have

merit in that better agreement was reached between those experienced in examining oral

biopsies with improvement in kappa agreements.54

Several points to increase the objectivity were put forward and authors commented

on various factors that could be responsible for the wide inter & intra observer variability

encountered in oral epithelial dysplasia.

The points considered were-

To enforce a systematic way of evaluation for the participating pathologist; they could

start examining the slides by looking at each individual feature of architecture and


44

cytological changes, and then they could build up their judgment on the degree of

dysplasia on the basis of evaluating these features.

The uncertainty on how to classify a particular lesion among pathologist could be

overcome by objective scoring for each criterion of dysplasia features.

Another source of variation might be related to the difference in the understanding of

these features in terms of recognition and impact on clinical outcomes. They felt that

even though only certain features were associated significantly with the clinical

outcomes, other features which showed no statistical significant association are also

important for the whole grading process.

Lack of standardized method for oral epithelial dysplasia; Although, the 2005 WHO

classification defined each category and pointed out how to grade the oral epithelial

dysplasia lesions, no definitive or exemplar photomicrographs that could illustrate the

individual architectural and cytological features were included .In the absence of such

information, assessment of these features depend on the educational process of grading

oral epithelial dysplasia.

It is the mission of the authoritative and regulatory bodies of the oral pathology

profession over the world to carry out more workshops, studies & meetings to improve

the educational process of grading oral epithelial dysplasia. A minimum data set similar

to that of Royal College of Pathologists (UK) for reporting head and neck cancer should

be set up.

To minimize the heterogeneity in reporting oral epithelial dysplasia; some pathologists

usually grade some oral epithelial dysplasia lesions as having features of mild dysplasia

with focal areas of moderate dysplasia. This essentially confuses the clinicians. The

advantage of minimum data set would minimize the heterogeneity of the future research

on epithelial dysplasia and helps to have homogenous data when meta-analysis or

systemic review is attempted.

Use of molecular markers: several papers exploring the prognostic and predictive values

of different molecular markers on oral epithelial dysplasia lesions are already published.

However it is not known how consistent the authors of these papers were in their

histological grading of the studied oral epithelial dysplasia lesions.


45

Reduction in number of categories in grading dysplasia: Studies on grading dysplasia

from other sites of the body demonstrated that increase in inter observer agreement

between pathologists probably is best accomplished by having fewer categories of

readily separable entities.

To conclude, grading of oral epithelial dysplasia is not an exact science and

pathologists are doing their best to reach an optimal result. This has led to the researches to

search for alternative and advanced methods for objective assessment of dysplasia and

determine its malignant potential. 54

g. VARIOUS METHODS OF EVALUATION OF ORAL EPITHELIAL DYSPLASIA:

The assessment of epithelial dysplasia by light microscopic examination of

hematoxylin and eosin stained tissue sections so far is not totally satisfactory. The grading of

dysplasia is dependent on morphological experience of the pathologist, thus it is more or less

an expression of a nominalistic concept of disease. There have been a number of other

approaches to the problem of grading dysplasia with the aim to establish a more

fundamental molecular and biochemical basis for assessing dysplastic lesions. These studies

correlate the molecular and biochemical markers with the grading of dysplasia which was

used as “gold standard” to establish a more objective method of assessing epithelial

dysplasia. Quantifiability and reproducibility could be added virtues as suggested by

Burkhardt A (1985).55

The methods are as follows : Burkhardt (1985)55, Brennan (2007)56

A. Morphological methods: Exfoliative and aspiration cytology, Quantification of

histopathology : Computer aided analysis, Stereological technique and Use of image

analysis, Histochemistry, Immunohistochemistry, In situ hybridization, Electron

microscopy.

B. Functional methods: Cellular proliferation studies, Analysis of immune status,

DNA histograms, Analysis of cell products in circulating blood, Experimental

models, Microsattellite analysis / Loss of Heterozygosity, Microarray technology .56

Exfoliative cytology was seen to be of limited use, amongst the different methods

for cytological and histological evaluation of dysplasia, as it gave high false negative results

and little information was obtained regarding the changes in the deeper layers of oral


46

mucosa.55 Aspiration cytology was seen to give better results than the smear, but it is a

micro invasive procedure.57 The micronucleus test done on the exfoliated cells as described

by Heddle JA (1973),58 was said to be a screening method for the assessment of

chromosomal damage and thus of mutagenic influences on the mucosa. An elevated

frequency of micronuclei, which were chromatin particles derived from acentric

chromosomal fragments was suggested to be indicative of the mutagenic influences.

Kramer (1970)59 said that by combining the ability of the pathologist to recognize

dysplastic features with the potential of the computer for pattern recognition, a diagnostic

hint can be obtained by processing the histological data by computer. But they opine that the

ultimate result depends on the recognition of the different dysplastic features by the

pathologist and hence prone for subjective alterations.

Franklin CD et al (1980)60 described stereological or two dimensional morphometry

technique by measuring dysplastic parameters on light or electron microscopic level. In an

experimental study on carcinogenesis, they found 3 histological features which were able to

differentiate benign and premalignant lesions, namely increase in nuclear cytoplasmic ratio,

nuclear chromatism, epithelial / mesenchymal interface. Since they found that many of the

parameters measured by these methods could be readily estimated by the human eye, the

quantification of the same was found to be unnecessary. Whether it plays any part in routine

diagnosis was not clear.55

Cahn LR et al (1962)61 noted that the chemical characteristics in the epithelium of

the premalignant lesions could be studied by application of different stains to the dysplastic

tissue sections, for example, they report that the PAS reaction indicated the presence or

absence of the glycogen. Burkhardt A (1985)55 concluded that the simple histochemistry

gave an insight into the understanding of the chemical nature in dysplastic lesions, but had

no diagnostic significance. Further he found the metabolic requirements of malignant and

premalignant cells to differ from those of the tissue of origin, by their enzyme

histochemical studies. Of the many different enzymes studied, the shift from predominantly

aerobic respiration to predominantly anaerobic glycolysis in many premalignant and

malignant cells was seen to be the fundamental metabolic change observed by Johnson et al

(1980).62 Besides the enzymes engaged in energy metabolism, Burkhardt A (1985)55 also

noted that the activities of hydrolytic enzymes may be altered in these lesions, and that an


47

increase in the production of hydrolytic enzymes by tumour cells may be related to their

invasive capacity. Although the use of enzyme histochemistry for routine purposes has been

stressed by Heyden C et al (1974),63 Burkhardt A (1985)55 opines that the variety of factors

such as sensitivity of the reactions, duration and temperature during transport, marked

regional tissue variations, greater work load in the success of immunohistochemistry,

resulting in inaccurate, variable and unreliable results lead to the apparent loss of interest in

enzyme histochemistry.

Immunohistochemistry was a very promising field of research which had

contributed much to our understanding of tissue alterations in different premalignant and

malignant lesions.55 He reviewed and stated that, since its inception by Coons A. H et al in

1941 stated that, there had been enormous and continued development in this field, in

pursuit of better methods and thus in the better understanding of premalignant and malignant

lesions. He further stated that the development of this method had been made possible, that

such tests could be applied on paraffin sections of fixed tissues and also made highly

specific monoclonal antibodies available. The aim of most of the immunohistochemical

studies had been to determine tumour associated antigenic constituents, to study the

variation in the expression of some antigens in normal, premalignant, malignant and in other

lesions, in other words to find a specific "Tumour marker". The various investigations had

been focused on epithelial surface antigens, intracellular products, constituents of the

basement membrane zone and changes in the adjacent tissue including the so called ‘stromal

reaction’.

The following markers have been used to assess dysplasia as stated by Scully C 64 :

1. Cell surface antigens: a. Carbohydrates. b. ABH blood group antigens. c.Thompson-

Friedrich antigen. d. Growth factors and receptors (EGF, EGFR). e. Integrins f.

Granulocyte colony stimulating factor receptor.

2. Intracellular components: a. Cytokeratins. b. Monoclonal antibody 17.13

3. Oncogenes : a. Tumor suppressor and apoptosis related genes – p53, Bcl-2. b. Cell

cycle regulators. c. Proliferation markers: PCNA, Ki-67, BrdU

4. Tissue Markers : a. Heat Shock Proteins. b. Matrix Metalloproteinases. c.Vascular

Endothelial growth factor

5. AgNOR

6. DNA content (DNA ploidy)


48

Silver binding nuclear organizer regions: AgNOR: Silver binding nuclear organizer

regions (AgNOR) are loops of ribosomal DNA that can transcribe for ribosomal RNA. In a

malignant cell, chromosome disarray with multiple nucleoli appears to result in an increase

in AgNOR counts suggesting poor prognosis for oral cancer. Mean count and area of

distribution of AgNOR within the nucleus was useful in diagnosing histological features

representing oral epithelial dysplasia.

In a study by Ray JG et al (2003)65, a mean Ag NOR count of 2.37 was used as a

potential cut point to distinguish between non dysplastic leukoplakia and dysplastic

leukoplakia. They compared epithelial dysplasia grading with mean AgNOR count as the

new diagnostic test due to its quantifiability and objective status and minimized

disagreements. Both the methods correlated in 65% of lesions. 17.3% of cases classified as

non-dysplastic by H & E method also had a high AgNOR count and were classified as

dysplastic by AgNOR. At the same time, 19.2% of cases classified as dysplastic by H&E

were classified as nondysplastic by the AgNOR count.

However, since mean AgNOR count of a lesion was an easily reproducible measure,

it would serve as a more objective criterion than the hematoxylin and eosin stain based

assessment for epithelial dysplasia. This proved AgNOR as a confirmatory adjunct to H&E

for diagnosing epithelial dysplasia. Additional attributes for use AgNOR are it is easy to

perform, is inexpensive & can be performed on paraffin embedded specimen. 65

Gross genomic alterations: (DNA ploidy status) (Scully C, 2003)66

DNA ploidy is a measurement of nuclear DNA content and forms a surrogate measure

of gross genetic damage. DNA ploidy status in the cells of the lesion suspected to be

premalignant had implications for the prognosis of the lesions. Subudo J et al (2001)34

conducted a study of the DNA content of 150 dysplastic oral leukoplakic lesions. They

graded the dysplasia according to WHO criteria. The genomic content was measured by

DNA image cytometry and the lesions were divided into: a) Aneuploid (high risk) lesions b)

Diploid (low risk) lesions and c) Tetraploid lesions (Intermediate risk).

They compared the genomic status with the histologic grade and the follow up clinical

outcome. It was found that no statistically significant correlation existed between histologic

grade and the DNA content of the dysplastic lesions. However, 3% of the low risk diploid


49

lesions, 60% of the intermediate risk tetraploid lesions and 84% of high risk aneuploid

lesions transformed into malignancy.

In a study of 150 patients with dysplastic oral lesions followed up for a mean of 18.6

years, it was noted that three of 105 diploid cases as against 21of 25 aneuploidy cases

developed into oral squamous cell carcinoma giving a positive predictive value of 84%. It is

significantly noted that lesions regarded as mild dysplasia histologically were aneuploid &

progressed to squamous cell carcinoma thus enhancing its efficacy in determining genetic

change ahead of histology.

The DNA ploidy status thus gave promising results as an objective marker for risk

assessment as it gave high positive and negative predictive values. The advantages include

use of automated cytometry on nuclei obtained from routinely processed samples. Thus

DNA ploidy may form a more realistic option to measure gross genomic damage.

Immunohistochemistry could be the most promising & reliable method of study at a

cellular level, and its use in grading of epithelial dysplasia could help in understanding them

better and in arriving at a universal grading system. But results with immunohistochemistry

have to be regarded with great caution until the reliability is well established and with use of

adequate standards.66

The in situ hybridization technique was thought to be of limited use in studying

premalignant lesions, as false positive results could be obtained and it had to be corroborated

by molecular hybridization to label oral carcinomas. Electron microscopy gives an insight

into the ultra-structural details of cellular differentiation in epithelial tissues and certain of

these were thought to be indicative of dysplastic lesions. Thus by knowing the defined

criteria, dysplastic lesions could be differentiated from normal tissue .As the quantification

of each feature is prone for subjective evaluation, the grading of epithelial dysplasia

becomes difficult.55 Banoczy J et al (1980)67, in their SEM studies found changes of cell

surface in dysplasia in the form of atypical or irregular pattern of microvilli, pleomorphism

and disruption of microridges. Burkhardt A (1985)55 stated that the simple electron

microscopy and SEM added to the understanding of the biology of disturbances of epithelial

differentiation in malignancybut do not contribute to routine diagnostic procedure.67

Besides these morphological methods described, a more functional type of

approach to the problem in assessing dysplastic lesions are also available. Hume WJ and


50

Potten CS (1979)68 extensively reviewed methods in the study of cellular proliferation.

Such studies included the counting of mitotic figures in conventional tissue slides (Mitotic

index) or after mitotic arrest in metaphase, and radioactive labeling with titrated thymidine.

The estimation of mitotic index was of limited use in epithelial dysplasia due to various

reasons such as - diurnal variations, difficulty in obtaining normal control values, the

increase in mitotic index not being specific of epithelial dysplasia, as it was also seen in

wound healing etc. Though radioactive labeling was described to be the most reliable

method their applications for obvious reasons was seen to be limited in humans. In vitro

methods of the same were of questionable biological significance.55, 68

The analysis of the changes in the immunostatus of an individual with premalignant

lesions did not lead to any affirmative test that could aid in diagnosing or understanding

them better. Bier J et al (1983)69 stated that the changes noted were all non-specific and

hence did not bear any prognostic implications. Kovesi G and Fekete B (1982)70 were of the

view that the assay of modified leukocyte adherence inhibition shows promising results as a

test for cancer. Tests for the analysis of the cell products such as enzymes involved in

metabolism of nucleic acids etc in serum, was done to find out if any significant changes in

these cell products helped in the prediction of malignant change. The reliability of these tests

however was not clear to be routinely used.

DNA histograms obtained by cytophotometric measurement for oral dysplastic

lesions, based on cells suspension/cytological smears following staining of nuclear DNA

have been applied. Saku T & Sato E (1983)71 put forward a dysplastic pattern of DNA

histograms suggestive of high diploid deviation of DNA ploidy pattern thought to have high

risk for malignant transformation. But this technique has high percentage of false negativity,

cannot be done on tissue section, it is time consuming, requires expensive hardware and

considerable experience therefore its value in conventional histology appears still to be

assessed.

Scully C. et al (1981)72 emphasized that analysis of cell products in circulating blood

could provide promising results. They reported an increase in concentration of β2

microglobulin, alkaline deoxyribonuclease and thymidine kinase in serum of patients with

oral dysplasia and cancer. High serum levels of deoxyribonuclease activity in dysplasia can

be a sign of poor prognosis but too little is known about the reliability and usefulness of

these tests.


51

Evson JW (1981)73 stated that in experimental studies using animal model like

hamster, mice and rats can help in assessing the progress of epithelial dysplasia which could

contribute to prevention and elimination of factors enhancing the development of

malignancy.

Microsattellite markers using DNA sequences which are relatively short tandem

repeats of nucleotide sequences, and one to four base pair units are used to detect the

allelic imbalance or loss of heterozygocity (LOH).66 Lee JS (1993); Shin DM (2001)74, 75

High frequency of LOH at 3P,9q & 17P was seen in dysplastic lesions. LOH at 9P 21 also

commonly occurs in dysplasia. Dysplasia may show chromosomal polysomy which is

associated with genomic instability reported to be increasing during tumorigenesis.

The assessment of multiple marker is complex and the techniques are not routinely

available therefore simpler and less expensive methods are usually sought. Recent advances

has made it possible of identifying molecular changes in exfoliated cells such as in saliva.76

A study done by Spafford K et al (2001)77 revealed genetic changes in exfoliated cells

identical to those found in oral squamous cell carcinoma in 80% of saliva samples.

Microarray analysis : The advent of microarray technology has allowed for

screening of the entire genome. This contrasts sharply with prior molecular biology

techniques that identified individual genes alone and thus has the potential to disclose new

pathways in the pathophysiology of dysplasia. The genes found to be unregulated in

dysplastic tissue include inflammation-related genes, cyclo-oxygenease-2(COX-2PTGS2)

and several gene receptors (PTGER 3b1, a2). Further studies are required for the validation

of role of these genes in dysplasia.56

Burkhardt A (1985)55 concluded that the future development and progress in these

advanced methods were to be expected as there were many promising aspects in these

techniques. But they suggested that at present there are still many questions related to the

diagnosis and grading of dysplastic lesions. Hence, they said that the assessment of

biological potential of oral epithelial dysplasia still relies on the “light microscopic

histological examination" and there was no alternative to this old and well proven approach.

Further they added that any advanced method apt to replace light microscopy must be able

to predict the true biological potential more accurately than the conventional histology.

Additionally, the aim of any advanced method must be to unmask those cases with

seemingly harmless histological picture but progress to malignancy.


52

Field cancerization and oral cancer

The concept and term ‘field cancerization’ was first put forward by Slaughter and

his co-workers in 1953. According to them, cancer develops multifocally as a result of

diffused and repeated carcinogenic assault and progresses at various rates within the entire

field. Slaughter stated that “Cancer does not arise as an isolated cellular phenomenon but

rather as an anaplastic tendency involving many cells at once. Subsequently the term,

“lateral cancerization” was used to indicate that the lateral spread of tumors was due to

progressive transformation of cells adjacent to a tumor, rather than the spread & destruction

of adjacent epithelium by preexisting cancer cells. The entire mucosa within a field due to

its high risk of malignant transformation has been referred to as “Condemned mucosa”. 78

This term encompassed the following issues: 1) Multiple cells form independent epithelial

tumors

2) Carcinogenic exposure affects multiple cells in a field.

Carcinogenesis is a multistep process and evidences indicate that presence of a field

lesion with genetically altered cells is a distinct biological stage in epithelial carcinogenesis

with important clinical implications.

Slaughter in 1953 defined field cancerization as “Increased risk of cancer

development in entire upper aerodigestive tract due to multiple genetic abnormalities in the

whole region after prolonged exposure to carcinogen”. Tobacco followed by alcohol has

been implied as the most common carcinogen. Boudewijin et al (2003)78 have proposed a

definition of field cancerization based on molecular findings as “the presence of one or more

areas consisting of epithelial cells that have genetic alterations. A field lesion (or shortly

field) has a monoclonal origin, and does not show invasive growth & metastatic behavior,

the hallmark criteria of cancer”. They state that a field lesion is preneoplastic and it may

have histological aberrations characteristic for dysplasia.

Gabriel DD et al (2007)79 described field cancerization as “the process where by

cells in a particular tissue or organ are transformed such that the genetically altered but

histologically normal appearing cells predate the development of neoplasia or coexist with

malignant cells, irrespective of clonality”.

Theories of field cancerization : (Monique G et al 2000)80

1. Multiple lesions develop independent of each other due to multiple genetic abnormalities

in the whole region following exposure to carcinogens. This theory is most accepted.


53

2. Other theory states that any transforming event is rare & suggests that multiple lesions

arrive due to widespread migration of transformed cells through the whole aerodigestive

tract. Two types of migration may be involved: (a) Migration of tumor cells via saliva:

e.g. micrometastasis (b) Intraepithelial migration of the progeny of the initially

transformed cells.

To understand the concept of field cancerization studies have evaluated these two

theories. Tumor associated mucosa (TAM) of smoking head and neck squamous cell

carcinoma patients show alterations not found in non smoking patients. Suggesting that

altered cells are harbored in associated margins of tumor in smoking patients, which are

absent in non-smoking patients. If the migration theory is considered, then changes should

have been observed in normal tumor associated mucosa of non smoking patients.

Additionally, it was also found that field changes like increase in nuclear area, proliferation,

p53 mutation were even observed in healthy smokers without tumor being a source for these

cells. Therefore, these results support the idea that field changes in head and neck squamous

cell carcinoma patients are carcinogen induced independent events than being a result of

migration of transformed cells.

There is a co mm on cl on al or ig in in multiple primary tumors even if the lesions

were more than 7 cm. This was based on similarity of genetic changes. Three theories have

been proposed to explain the common clonal origin of multiple primary tumors : first single

cell or small clusters of cells migrate through the submucosa, secondly they are shed in

lumen of an organ (oral cavity) at one place & regrow at another. The most recent finding in

head and neck is that a large contiguous genetically altered field exists in the epithelium

from which multiple clonally related neoplastic lesions develop which is explained based on

studies by Jang SJ et al (2001) 81; Braakhuis BJ et al (2002)82; Simon R (2001). 83 This gives

impetus to lateral clonal spread of the cancer giving rise to new primaries. Based on this

Braakhuis BJ et al (2002) suggested the use of the term ‘second field tumors’ for those

lesions that are anatomically distinct but demonstrate genetical similarities & true second

primaries are those lesions that did not share genetic similarity and develop as independent

events. 82

Recently studies have shown presence of large altered field in the epithelium in

which multiple clonally related neoplastic lesions develop. This indicates that a large


54

proportion of multiple primary tumors in the same or adjacent anatomical area have

developed within a single preneoplastic field. Based on this, a genetic progression model for

oral cancer was put forward where field change plays a central role as stated by Boudewijin

JM et al (2004).84

Patch is a clonal unit (2 mm in diameter) of mutated cells i.e. genetic change in stem

cell due to p53 mutation and has transformed this mutation to its daughter cells. This patch

appears to be first manifestation of oral cancer. Patch is cluster of cells showing p53

mutation.

Field of genetically altered cells -Tumor adjacent normal epithelium demonstrates genetic

alteration making it a field at risk. They are larger than patches, atleast 4mm in diameter but

may be over 7cm also. They show allelic loss at various chromosome arms 3p, 9p,17p &

may precede development of cancer.

Studies on Normal appearing mucosa adjacent to squamous cell carcinoma for

evaluation of field cancerization

Morphological changes : In tissue from adjacent normal margins of oral squamous cell

carcinoma (OSCC) may show epithelial hyperplasia, hyperkeratinization with

accompanying atrophy and fibrosis of the submucosa, round cell infiltration and capillary

telangectasia. Dyskaryosis with marked atypia, obvious in-situ or carcinoma is also reported

in studies. These findings are evidence of multi-centric origin and represent lateral spread of

oral squamous cell carcinoma.

Wright A et al (1985)85 studied histologically presence of dysplasia in epithelium

immediately adjacent to neoplasm in 80 cases of excised OSCC based on the concept that

most neoplasm’s are frequently preceded by precancerous lesions in an altered field. 80% of

cases showed presence of dysplasia and most common features seen were presence of basal

cell hyperplasia, nuclear hyperchromatism, loss of polarity, loss of adherence and disturbed

epithelial maturation. The occurrence of increased mitosis and abnormal mitosis was

relatively infrequent. One more important finding was that the dysplastic changes in the

adjacent epithelium were usually multi-centric.

Field change in oral cancer by quantitative assessment of cytomorphology of smears

taken from normal buccal mucosa in oral cancer & cancer free patients was reported by

Ogden GR et al(1990).86 They tried to detect changes arising in tissue peripheral to cancer


55

and effects of potential etiological factors (smoking) on occurrence of field change. A

statistically significant reduction in cytoplasmic area in normal mucosa of cancer group was

seen which suggested that decrease in cytoplasmic area preceded an increase in nuclear area

within tissue that become malignant. The authors believe that this decrease in cytoplasmic

area may not be related to smoking as half the patients did not smoke. This finding is

supported by clinical observation of patients developing second primary tumors even after

cessation of smoking. They concluded that normal oral mucosa of cancer patients is

different from that of cancer free individuals and this appears not to be dependent of alcohol

or tobacco. Cancer free smokers though showed a slight increase in nuclear area as

compared to cancer free non smokers.

Kannan S. et al (1996)87 examined fifteen cases of adjacent oral epithelium to OSCC

were under light & electron microscopy. They performed the incisional biopsy in mucosa

immediately adjacent to the cancer. 67% of the lesions exhibited hyperplasia or mild

dysplasia while 33% showed moderate to severe dysplasia.

Ultra-structural observation of all the biopsies showed subcellular alterations deep in

the lower half of the epithelium similar to those seen in frank oral carcinomas. The

important changes observed were bizarre nuclei of basal and lower spinal cells, enlarged &

multiple nuclei, presence of inter chromatin and perichromatin granules, loss of desmosomes

and marked spongiosis as well as disturbed cellular maturation evidenced by abnormal and

irregular distribution of maturation markers like keratohyaline granules and tonofilaments.

In addition, basal lamina was very thin but was intact thus heralding its breakdown in future.

The authors believe that electron microscopy could be valuable in detecting subtle changes

due to field cancerization in adjacent epithelium of OSCC.

Thompson PJ et al(2002)88 undertook study to quantify the incidence of field change

in oral mucosa by taking ‘mirror image’ biopsies in 26 patients with unilateral OSCC from

clinically normal looking mucosa at corresponding anatomic sites. 58% of patient showed

histologically abnormal tissue on microscopic examination consisting of reactive change or

cellular atypia associated with chronic irritation, frank dysplasia, carcinoma in-situ or micro

invasive squamous cell carcinoma. The authors also found increased vulnerability for

dysplastic change in lesions present in the lateral / ventral part of tongue & floor of the

mouth though these observations were not statistically significant. They suggest use of

multiple biopsies, long term follow up & use of chemoprevention.


56

h. MOLECULAR ASPECTS OF POTENTIALLY MALIGNANT DISORDERS:

Cytokeratins are intermediate filament proteins found in cytoplasm of all epithelial

cells. There are at least 20 different keratins polypeptides that are expressed in different

combinations, depending on type of epithelium and degree of differentiation. Presence of

cytokeratins 7, 8, 13, 16, 19 was observed at abnormal anatomical sites or at abnormal

intraepithelial levels in normal mucosa from head and neck OSCC patients as observed in

studies by Copper MP et al (1993)89; Ogden GR( 1993)90; Bongers V et al (1996)91; Bosch

FX et al (1989).92

Blood group antigens are cell surface carbohydrates. These show changes in

expression related to tissue type, differentiation state and cell motility capacity. Type 2 chain

ABH carbohydrate structures are distributed broadly in epithelial and endothelial cells

independent of patients ABO blood group. In normal oral epithelium, type 2 chain ABO

blood group antigens are expressed in parabasal cells. Bonger et al (1996)91 in their study

found a fourfold lower expression of type 2 chain ABH antigen in exfoliated cells from

normal mucosa from six different places distant from squamous cell carcinoma as compared

with healthy individuals. They believe that this may be a negative marker for field change

and risk indication.

Cyclins are cell cycle regulators and are functional when associated with their

concomitant Cyclin dependent kinases. Cyclin D1 regulates the G1-S transition and is

associated with CdK4 and 6. Clearly defined foci of Cyclin D1 expression in sections of

normal mucosa adjacent to squamous cell carcinoma was observed which was not observed

in the sections of normal mucosa of healthy individuals as reported by Bartkova J et al

(1995).93

King LE et al (1985)94 have shown increased expression of epidermal growth

factor receptor (EGFR) in tumor associate normal mucosa which could be due to elevated

mRNA or amplification of EGFR gene. Ligand binding to the extracellular domain of EGFR

causes receptor dimerization which activates tyronine kinase function. This leads to

autophospharylation, resulting in proliferation. EGFR expression was high in

nonsmoking/non drinking head and neck OSCC patients as well as in smoking/drinking

head and neck squamous cell carcinoma patients. It was also observed that the EGFR

expression was less elevated when the epithelium was located more distant to the tumor.


57

These results suggest paracrine effect on EGFR expression due to factors related to tumor

and not due to carcinogen (smoking).

Oijen V et al(1998)95 demonstrated increased number of proliferating cells using the

marker Ki 67 in both adjacent tumor tissue and control oral epithelium in smokers and non

smoker in the absence of histologically visible epithelial alterations and suggested that high

proliferation increases the likelihood of malignant change.

Bcl2 is an anti-apoptotic markers. Its family plays an important role in regulation of

the apoptotic pathway. In normal tissue there is a balance between cell proliferation and

apoptosis (programmed cell death). Alterations in both pathways contribute to the clonal

expression of cancer cells. There was lack of bcl2 expression adjacent normal tumor

associated mucosa in OSCC as compared to healthy control mucosa. The authors suggest

that the expression of bcl2 has to be interpreted in the context of levels of other bcl2 and Bax

family members as stated in the study by Birchal MA et al (1997).96

The expression of all gluthione S-transferase isoenzyme was significantly higher in

suprabasal and superficial layers of normal oral mucosa of OSCC patients who subsequently

developed second primary tumors as compared to its expression in normal mucosa from

OSCC patients who were free of disease after 7 years follow-up reported Bonger V et

al(1995).97 The authors believe that this increase is intriguing as elevated levels of their

detoxification enzymes actually protect against carcinogenic attacks and probably this might

represent a futile effort for combating the carcinogenic metabolites in tobacco. They believe

that a high level of this enzyme has predictive value for development of second primary

tumors.

Loss of function of tumor suppressor gene p53 might result in uncontrolled cell

division and genomic instability. Mutant p53 has higher stability than wild type p53, which

allows accumulation to the levels detectable by immunohistochemistry. Oijen V et al

(1999)98 observed a uniformly spread focal over expression of p53 occurred more frequently

in normal epithelium of smoking OSCC patients as compared to nonsmoking OSCC patients

and from healthy controls. Increase p53 cell cluster were not detected in mucosa of healthy

smokers. Tobacco in addition to other genetic and environmental factors could contribute to

abundance of p53 cell cluster in tumor associated mucosa. All these findings suggest that

p53 plays a pivotal role in early carcinogenesis and development of field cancerization. In


58

the recent patch field concept put forward by Boudewjan JM et al (2004)84 p53 positive cell

clusters are considered as the initial change (patch) in the progression of field cancerization.

Thus, to summarize the data on these molecular markers, it appears that normal

mucosa associated with carcinoma shows a range of alteration and genetic abnormalities.

Additional research is required to assess which field changes depicted by these molecular

markers have actual carcinogenetic influence & which are merely epiphenomenal.

Various clonal markers have been used to explain whether oral field cancerization

is by migration or independency theory. Patrick KH et al (2003)99 stated if multiple primary

tumors arise due to migration of malignant cells from a primary source, then all should show

identical genetic alterations. In case of independent origin, these alterations will be different.

To qualify as a clonal marker, such a genetic alteration should occur early in the

development of the primary lesion, be maintained during the progression of the lesion,

exhibit sufficient variability, be applicable in the majority of the lesions and should not

occur by chance & be present 100% of time.

Karyotype is the number and appearance of chromosome in the nucleus of

eukaryotic cell arrested in metaphase between two lesions and comparing their appearance

and search for ploidy and chromosomal breaks/rearrangements. Karyotypes of tumor have

been used as clonal markers. However, this is a rather insensitive method of detection and

requires tissue culture which in turn causes additional genetic alterations so it may not be

accurate reported by Worsham MJ et al (1995).100

Partridge M et al (2000)101 reported a case control study using microsatellite assay

with markers like 3p, 8p, 9p, 13q, Rb was done in two group of patients, one with dysplasia

who subsequently developed carcinoma on same side of mouth and another group with

similar lesions who did not develop lesions on a fixed time of follow up to identify high risk

individual for development of oral squamous cell carcinoma in a field of cancerization. They

found that dysplastic lesions harboring allelic imbalance at more polymorphic loci within

critical chromosomal regions developed tumor more frequently than matched controls with

allelic imbalance at fewer loci. They suggested that although dysplastic lesions may not be

precursor lesions per se but they developed in fields that harbored the genetic alterations

associated with tumorgenesis because microsatellite assay could identify the benign from the


59

more troublesome true precursor lesions. Process of field cancerization is very extensive

and excision of lesion might leave back mucosa which harbors aberrations that cannot be

seen. Thus normal appearing mucosa may have source of cells for development of

subsequent tumors. This risk could be detected by use of allelic imbalance for prediction of

subsequent malignancy. Based on their findings, the authors advocate complete excision of

all dysplastic lesions showing allelic imbalance of two more key microsatellite markers

regardless of degree of dysplasia to produce margins that are atleast morphologically normal

followed by dietary advice and chemoprevention. They suggest the screening of clinically

normal mucosa as a routine work up in head and neck cancer patients by histology or

molecular tests if possible to establish evidence of field cancerization and for overall risk

assessment.

Tabor BRH et al (2001)102 quantitatively measured loss of heterozygosity in head

and neck squamous cell carcinoma. They reported 10 out of 28 of patients have tumor

associated genetic alterations in biopsy taken from macroscopically normal mucosa adjacent

to the tumor. This frequency was based on analysis of 4 biopsies that were taken in each

quadrant surrounding the tumors using 15 microsatellite markers in 6 chromosomes. In

majority of cases (7 of 10), there genetically altered cells were also found in margins of

specimens removed by the surgeon. These lesions were more than 4mm in diameter and

contained more than 50% of aberrant cells.

Analysis of mitochondria DNA markers by Ha PK et al (2002)103 in normal

adjacent mucosa of dysplastic lesions and identical mitochondrial DNA mutations were

found in 3 out of 8 lesions suggesting a monoclonal origin. They emphasized that these

changes could be detected in saliva and their levels decrease with treatment indicating its

use in early detection and monitoring of oral cancer.

Chromosomal Aberrations: Genomic instability increases the risk to develop head

and neck squamous cell carcinoma. Analysis of normal and malignant lesion adjacent to

head and neck tumors for chromosome instability using in situ hybridization has suggested

that molecular aberration can be detected before histological evidence of morphologic

changes.

A whole series of chromosomal changes were observed by Ai H et al (1999)104 using

fluorescence in-situ hybridization in cells from brushes of macroscopically normal cheek


60

mucosa opposite the site of OSCC and correlated with smoking. A trend towards

aneusomies of chromosome 2, 6 and Y was observed in normal mucosa from smokers.

RNA expression analysis of oral keratinocytes can be used to detect early stages of

oral cancers was stated by Joel L et al (2008).105 RNA analysis was done using hamster

mucosa brush cytology samples measured with RTPCR. The authors suggested that brush

oral cytology may prove useful as a source of RNA for gene expression analysis during the

progression of diseases of oral epithelium such as OSCC.105

Taboor MP et al (2001)106 suggested the presence of genetic alterations in the mucosa

samples was found to be associated with histopathological changes. The majority of the

genetically altered fields were classified as dysplastic, however the cases without genetic

alterations were observed in the mild dysplasia group.

i. CORRELATION OF DYSPLASIA WITH EVENTUAL MALIGNANT TRANSFORMATION:

The development of a malignancy in stratified squamous epithelium occurs

spontaneously or as a gradual process in which multiple minor individual cellular and tissue

alterations eventually culminate in frank malignancy.14

There is a generally held view that assessment of dysplasia in pre-malignant lesions

is important because dysplastic lesions are more likely to undergo malignant change and

because it is felt that the chances of malignant transformation increase with increasing

severity of dysplasia. There is a broad measure of agreement among pathologists about the

altered epithelial features that comprise dysplasia, but there is great variation in interpreting

the severity of dysplasia and not all investigators describe the criteria they use.30

Banoczy J and Csiba A (1976)38 diagnosed a case as showing dysplasia when it had

two or more histological dysplastic features. This would tend to increase the proportion of

cases described as dysplastic by inclusion of those with minimal dysplasia that might not be

regarded as dysplastic by other pathologists. In their series of 500 cases of clinical

leukoplakia, 120 showed dysplasia. At follow-up, 13.2% of these dysplasia progressed to

malignancy. Dysplasia was graded into three categories of severity on the basis of the

numbers of histological dysplastic features evident. In their series, all of the instances of

malignant transformation arose from cases with dysplasia, as assessed by their criteria,

evident on biopsy.38

Pindborg JJ, Daftary DK & Mehta FS (1977)107 investigated carcinomas developing

from non-dysplastic lesions in a rural Indian population. They found malignant

i


61

transformation after a shorter average time follow-up in dysplastic lesions, and this change

also occurred in a younger age group.

The data from Silverman S et al (1984)108, support the concept that dysplasia is an

important predictor of risk. However, 82% of the cases of malignant transformation were in

patients diagnosed as not showing dysplasia on biopsy. They studied 257 patients in

the USA with leukoplakia, of whom 45 progressed to carcinoma in the follow-up period of

eight years. Only 22 of these patients had dysplasia in the initial biopsy. They do not state

their criteria for recording a case as dysplastic. Biopsies that other pathologists would have

regarded as containing mild dysplasia might not have been designated as such. Of the 45

cases that progressed to malignancy, 8 were derived from the 22 cases with dysplasia,

suggesting a 36% incidence of malignant transformation as contrasts with the 15.7%

incidence in the cases not showing dysplasia on biopsy.

Such a follow-up studies suggest that dysplastic lesions are more likely to proceed to

carcinoma than non-dysplastic lesions, and they also provide some evidence in support of

the concept that more severe dysplasia is associated with increased risk. However, the grade

of dysplasia is not a reliable predictor of prognosis. It also appears that carcinoma can

develop from an epithelium showing only mild dysplasia or no dysplasia. This could be a

false conclusion as it might be due to the wrong site being biopsied by bad judgment or due

to bad luck. It is also important to stress that not all cases in which dysplastic epithelium is

present inevitably progress to carcinoma.30

The microscopic features of dysplasia are now relatively well established.

Lesions with specific dysplastic grades or changes have been followed in order to

determine their natural history. This is difficult because of the following facts: 1. the

biopsy procedure itself has removed the cells upon which the diagnosis is based, 2. the

more severe dysplasia are typically removed or destroyed prior to follow-up, 3. the

grading of dysplasia is an extremely subjective pursuit and there is often poor correlation

between experienced pathologists.49

Keeping the above facts in mind, it seems to be accepted that the biological

behavior of severe epithelial dysplasia and carcinoma in situ are identical. Investigations

have found that 20.35% of severely dysplastic lesions develop carcinoma, which is

similar to the figures for carcinoma in-situ. At the opposite end of the spectrum, mild

epithelial dysplasia so seldom eventuate in carcinoma, and are so similar to reactive


62

epithelial changes, that few pathologist consider them a serious threat or recommend

complete removal of the associated white lesion.109

One of the ironies of oral pre-cancers is the apparent fact that the correlation

between the clinical appearance of leukoplakia and cancer transformation is relatively

good while any microscopic changes less than severe are not helpful in determining

prognosis. Recent investigators have even concluded that they were better able to

correlate the presence of papillomavirus with the clinical appearance of oral pre-cancers

than with the microscopic grade of dysplasia.109

There is a substantial need to improve the histologic assessment of epithelial

dysplasia, since epithelial dysplasia does not seem to be invariably associated with or even a

necessary pre-requisite for malignant development, it may be necessary to develop other

methods for predicting the malignant potential of pre-malignant lesions.30

Nevertheless, taking into account all the uncertainties and anomalies of behavior of

these lesions, it does seem justifiable in our present state of knowledge to consider in

general terms that the degree of dysplasia is linked to the degree of probability of the

development of malignancy. In any event, it is a safe rule to advise the complete removal,

wherever possible, of all lesions showing more than slight degree of dysplasia. Also, it

should again be emphasized that there are certain 'high-risk sites', especially the floor of the

mouth and ventral surface of the tongue, where greater attention should be given to even the

lesser degrees of dysplasia.23

Thus, the challenge is to predict which lesions will eventually develop into

carcinoma. Conventional clinical and histopathological aspects are not optimal for decisions

on management, which is influenced by the perceived risk of malignant development.

Resource-consuming management procedures might be spared if the risk of malignant

development could be predicted with reasonable certainty and management of risk patients

could be improved.30

Summerlin DJ (1996)110 states that cases of epithelial dysplasia, however, may

remain static or even regress, whereas examples of squamous cell carcinoma have been

shown to arise de novo. Epithelial dysplasia, though represents a pre-malignant condition

should be treated as if it will progress to carcinoma.

Progress in molecular oncology has significantly advanced our knowledge on


63

tumorgenesis; yet the practical applications of these genetic markers remain unresolved in

detecting oral dysplasia. None of the molecular markers, single or in combination appear to

be ready for use in routine clinical diagnostic practices, although many molecules have

been studied as intermediate markers. Not many studies have focused correlating the

findings at molecular level with the clinical attributes of terminal outcome. Authors suggest

that studies should aim to develop these markers for routine application with improvement

in sensitivity, predictability, and reproducibility. 16

Recent advances in the use of quantitative high throughput methods has enabled the

profiling of gene expression (microarray), protein expression (proteomics), screening of

epigenetic changes, e.g. by pyro-sequencing and single nucleotide changes (SNP) of

hundreds of pre-selected genes simultaneously. These techniques which require minute

amounts of tissue when carried out in controlled way and using careful selected samples

with proper validation will help the potentially malignant and malignant tissue samples.

Combined with routine histopathology studies, these broad based studies of gene and their

products appear to have a great deal of potential for the identification of diagnostic and

prognostic markers of pre-cancer, and data from multicenter studies using microarray

technology may be engaged in the future for a better understating of aetiopathogenesis of

pre-cancer.16


64

Ki-67 PROTEIN :

The Ki-67 protein was originally defined as the prototype monoclonal antibody Ki-

67, which was generated by immunizing mice with nuclei of the Hodgkin lymphoma cell

line L428. The name is derived from the city of origin (Kiel, Germany) and the number of

the original clone in the 96-well plate. Antigen Ki-67 also known as Ki-67 or MKi67 is

a protein that in humans is encoded by the MKi67 gene. The Ki-67 protein is a cellular

marker for proliferation. It is strictly associated with cell proliferation.

During interphase, the Ki-67 antigen can be exclusively detected within the cell

nucleus, whereas in mitosis most of the protein is relocated to the surface of

the chromosomes. Ki-67 protein is present during all active phases of the cell cycle (G1, S,

G2, and M) but is absent from resting cells (G0).111 The estimated half-life of ki-67 antigen is

60-90 minutes. It starts to be expressed in the S phase, progressively increasing through S

and G2 phases and reaching a plateau at mitosis. After cell division, the cells return to G1

with a stock of ki-67 antigen, whose level decreases rapidly during this phase. 112

Antigen Ki-67 (pKi-67) is associated with ribosomal RNA transcription. During

mitosis pKi-67 is present on all chromosomes, forming reticulate structure surrounding

metaphase chromosomes. Inactivation of antigen Ki-67 leads to inhibition of ribosomal

RNA synthesis. The expression of which is strictly associated with cell proliferation but is

absent in non-proliferating cells. This is widely used in pathology as a proliferation

marker to measure the growth fraction of cells in human tumors.113

Ki-67 antibodies have been used widely for the estimation of the growth fraction of

clinical samples of human neoplasms and of normal cells in culture. Molecular and

functional characterization of pKi-67 is necessary for a greater understanding of the role of

the antigen in the cell cycle of normal and of neoplastic cells. Western blotting has revealed

that Ki-67 reacts with two polypeptides of 345 and 395 kDa. A full length cDNA coding for

pKi-67 has been cloned by immunoscreening of human cells. Analysis using the cDNA has

identified two differentially spliced mRNAs encoding polypeptides of predicted molecular

masses of 359 kDa and 320 kDa. Both sequences contains 16 ‘Ki-67 repeat’ sequences, each

of which includes a conserved 66 bp ‘Ki-67 motif’. The deduced amino acid sequence of the

cDNA has revealed two potential nuclear localisation signals, a number of potential

sequences, implicated in proteolysis of the antigen and numerous potential sites for

phosphorylation, amidation, N-myristoylation and ATP/GTP binding. Despite recent

advances in the molecular characterisation of pKi-67, a detailed analysis of the function of


65

the antigen is yet to be determined. Significantly, incubation of IM-9 cells with synthetic

antisense deoxyoligonucleotides complementary to the deduced translation start site of pKi-

67 prevents incorporation of thymidine, suggesting that pKi-67 is necessary for cell cycle

progression.

Indirect immunofluorescence has revealed that a large number of antigens are

located exclusively in the nucleolus of interphase cells. A number of antigens which are

located in the nucleolus during interphase associate with chromosomes during mitosis.

Furthermore, a small number of antigens are found in the nucleolus of proliferating cells

only. However, pKi-67 appears unique in that it associates with nucleoli of proliferating

cells only and with chromosomes during mitosis.

The nucleolus contains rRNA gene repeats and is the site of ribosome synthesis.

Nucleoli display a unique higher-order structure. During the cell cycle, nucleoli undergo

characteristic reorganisation. During nucleologenesis, prenucleolar bodies form around

Nucleolar organising regions (NORs) which subsequently fuse to form the interphase

nucleolus. Early in mitosis, nucleoli disperse, and nucleolar antigens become redistributed.

Some nucleolar antigens remain associated with nucleolar domains, whilst others are

dispersed throughout the mitotic cytoplasm or distributed around chromosomes. Ultra

structural analysis of mature nucleoli has defined three distinct nucleolar sub-structures,

each associated with distinct nucleolar functions.

The fibrillar centres (FC) are surrounded by the dense fibrillar components (DFC).

The fibrillar components of the nucleolus are embedded within the granular components

(GC). Many nucleolar proteins are compartmentalised within nucleolar domains. For

instance, fibrillarin is localised within the DFC, nucleophosmin/B23 is localised within the

GC whilst DNA topoisomerase I and RNA polymerase I are localised within the FC.

Compartmentalisation of proteins within the nucleolus reflects compartmentalisation of

specific nucleolar functions.

Since Ki-67 is located within nucleoli of proliferating cells only, it has been

suggested that the antigen may regulate nucleolar metabolism by, for example, increasing

rates of ribosomal synthesis required by rapidly dividing cells. However, it is not entirely

clear with which nucleolar compartment Ki-67 is associated. Immunoelectron microscopy

has revealed that Ki-67 antibodies stain regions surrounding the FC, probably localised

within the DFC. Ki-67 staining is reported to be absent from the FC, GC and nucleolar

interstices.114


66

MATRIXMETALLOPROTEINASE (MMP) :

MMPs are a family of structurally related but genetically distinct enzymes that

degrade extracellular matrix and basement membrane components and regulate cell–matrix

composition. MMPs are calcium-dependent zinc-containing endopeptidases belonging to

metizincin superfamily. The interactions of cells with the ECM are critical for the normal

development and function of the organism. By regulating the integrity and composition of

the ECM structure, these enzyme systems play a pivotal role in the control of signals elicited

by matrix molecules, which regulate cell proliferation, differentiation, and cell death. The

turnover and remodeling of ECM must be highly regulated since uncontrolled proteolysis

contributes to abnormal development and to the generation of many pathological conditions

characterized by either excessive degradation or a lack of degradation of ECM

components.115

History : Woessner in 1962 first described MMP as protein enzyme in mammalian uterus

which could degrade collagen. Gross and Lapiere identified MMP, while attempting to

establish how a tadpole loses its tail during metamorphosis and discovered the first member

of this family (MMP-1). A higher molecular mass species (72 kDa) with gelatinolytic

activity in comparison with MMP- 1 was identified as MMP-2. Goldberg and colleagues

sequenced and purified MMP-2 from human rheumatoid synovial fibroblasts and isolated in

the 1970s and initially denoted as 72-kDa type IV collagenase / gelatinase A. MMP-3 was

identified as a lower molecular mass species (54 kDa) with proteoglycan and casein

degrading activity. In the late 1980s, Ed Harris and colleagues first proposed using the name

MMP for this family of enzymes. Subsequently, the International Union of Biochemistry

and Molecular Biology designated the family with the unique name MMPs and assigned

each family member with an enzyme number. By 1991, MMP-1, -2, -3, -7, -8, -9, and -10 as

well as tissue inhibitor of metalloproteinase (TIMP)-1 and -2 had been named and

characterized. Since then, the family has expanded gradually; the latest member discovered

is MMP-28, making a total of 23 human MMPs. MMPs activity is tightly regulated in

several ways, including their activation, the presence of specific endogenous inhibitors,

TIMP and by gene expression. MMP s may be secreted by tumour cells, fibroblasts,

endothelial cells and macrophages as well as mast cells and neutrophils .116


67

Functions : MMP s are involved in physiological processes such as development,

morphogenesis, angiogenesis, bone remodeling, wound repair, apoptosis. These are also

involved in pathological processes such as in periodontal diseases, cancer invasion and

metastasis. 117

TYPES OF MMP s

MMPs are classified into subgroups as collagenases, gelatinases, stromelysins, and

matrilysins on the basis of their substrate specificity and molecular structure. The

membrane-type MMPs (MT-MMPs) are subclass of MMPs that additionally contain a

transmembrane and intracellular domain, a membrane linker domain, and are membrane

associated as described Verma RP et al (2007)118 and Vihinen P et al (2002).119

1. Collagenases includes MMP-1, MMP-8 and MMP-13

• Collagenases are proteinases capable of cleaving types I, II, III, V and IX collagens and

play a decisive role in remodeling and degradation of the Extra Cellular Matrix (ECM).

MMP-1 hydrolyzes type III collagen more rapidly than type I, while MMP-8 shows a

slight preference for type I collagen. Conversely, MMP-13 prefers type II collagen and

hydrolyzes this collagen much more rapidly than MMP-1 or MMP-8.118,119

• Gelatinases includes MMP-2 and MMP-9 : MMP-2 is widely distributed in skin

fibroblasts, keratinocytes, chondrocytes, endothelial cells, monocytes, osteoblasts and in

a number of other normal and transformed cells. MMP-9 is produced by keratinocytes,

monocytes and alveolar macrophages, polymorphonuclear leukocytes and in a number of

malignant or transformed cells. In OSCC, MMP-9 is expressed in carcinoma and

inflammatory cells around carcinoma islands. MMP-2 is mainly found in carcinoma-

associated fibroblasts (CAFs). Gelatinases has the ability to break type IV collagen,

found in the basement membrane. They play an important role in angiogenesis, tumor

invasion and metastasis, and have frequently been associated with poor prognosis. They

are also responsible for the final degradation of fibrillar collagens after initial cleavage

by collagenases. Gelatinases are secreted as inactive proforms that are activated

extracellularly. 118,119

2. Stromelysins : MMP 3, MMP 10, MMP 11, MMP 12 : Degrade various components of

the ECM and activate collagenases via the proteolytic removal of a propeptide.

Stromelysin 1 (MMP - 3) and stromelysin-2 (MMP - 10) are closely related to structure


68

and substrate specificity. This subgroup also includes stromelysin-3 (MMP-11) and

metalloelastase (MMP - 12). 118,119

3. Matrilysin : MMP7 and MMP 26 : These are the smallest MMPs. They lack the C-

terminal hemopexin domain, common to other MMP family members and they have

markedly smaller molecular weights and are mainly expressed in epithelial cells in

different glandular structures. MMP-7 acts in intestinal mucosal defense by activating

antibacterial peptides, defensins. Matrilysin-2 (endometase, MMP-26) is expressed in

uterine endometrium and placenta and in cancers of lung, prostate and breast. 118,119

• Membrane type MMPs (MT- MMP) : MMP 14, MMP15, MMP 16, MMP 24, MMP

17 and MMP 25. These can activate proMMP-2 and most of them degrade ECM

components. In addition they contain a transmembrane and cytosolic domain. These are

anchored to cell membrane by a transmembrane domain in the C-terminus or attached to

cells by a C-terminal GPI anchor.

4. Additional MMPs include: MMP 19, MMP 20, MMP 21, MMP 22, MMP 23, MMP

28, MMP 29 : MMP-19 is expressed in mammary gland, placenta, lung, pancreas, ovary,

spleen, intestine, blood vessels of skin and uterine ligaments and in activated peripheral

blood mononuclear cells. MMP-20 (Enamelysin) has a restricted expression pattern in

dental tissues and is thought to play a role in tooth enamel formation. MMP-23 is

expressed in adult ovary, testis and prostate, suggesting its role in reproduction. It is also

expressed in adult heart, intestine, colon, placenta and lung. MMP-28 (Epilysin) is

expressed in testis and at lower levels in lungs, heart, colon, intestine and brain.

Expression of epilysin is detected in the basal epidermis of intact skin and in basal

keratinocytes in wounds.

REGULATION OF MMP s

Loffek S et al (2011)120 and Yan C et al (2007)121 suggested Cell-ECM interactions trigger

cellular signaling that promotes cell differentiation, migration, and mobilization which is

essential for normal cellular homeostasis. MMPs act on matrix and matrix substrates, such

as cytokines, chemokines, growth factors, cell surface receptors, and adhesion molecules.

MMPs can affect cell behavior in many ways. The cleavage products of MMPs signal in an

autocrine or paracrine manner. They cleave intercellular junctions or the BM regulating

epithelial tissue architecture and activate the action of latent and deactivate the action of


69

active signaling molecules. Balanced and regulated degradation of ECM proteins by MMPs

are involved in many physiological processes. Conversely, excess MMP activity plays a role

in several pathological processes also. Therefore, understanding of the regulation of MMP

activity, under physiological and pathological conditions, could lead to novel therapeutic

intervention.

MMP expression has both constitutive and inducible components; MMP-2 and

MMP-1 are constitutively expressed in many cell types while MMP-9 is highly inducible.

The function of MMP s can be influenced by various effectors which include growth factors,

cytokines, chemical agents, reactive oxygen species etc. The enhanced MMP gene

expression may be down-regulated by suppressive factors like transforming growth factor

beta, retinoic acids, and glucocorticoids. Significantly, fragments of ECM cleaved by MMPs

can also potentiate the expression of other MMPs. Activity of MMP is also controlled in the

pericellular environment by modulation of activation and by the action of inhibitors.

Extracellular matrix metalloproteinase inducer (EMMPRIN) is a cell surface glycoprotein

and is expressed on the surface of tumor cells which stimulates adjacent fibroblasts and

tumor cells to produce several matrix metalloproteinases.

The catalytic activity of MMPs is strongly controlled at four different levels:

1. Gene expression with transcriptional and post-transcriptional regulation;

2. Extracellular localization and tissue or cell type of MMP release, termed

compartmentalization;

3. Pro-enzyme activation by removal of the pro-domain; and

4. Inhibition by specific inhibitors, i.e. tissue inhibitors of matrix metalloproteinases

(TIMPs), and by non-specific proteinase inhibitors, e.g. -2 macroglobulin.120,121

Role of MMPs in Oral Pathologies : MMPs has a marked role in tissue destructive oral

diseases. The expression and activity of MMPs in adult tissues is normally quite low, but

increases significantly in various pathological conditions that may lead into unwanted tissue

destruction, such as inflammatory diseases, tumour growth and metastasis. The best known

examples of unwanted tissue destruction related to over expression and activity of MMP

includes periodontitis, peri-implantitis, dental caries and OSCC.122

MMPs in Oral Premalignant and Malignant Lesions : OSCC is preceded by white or

red lesions can show varying degrees of epithelial dysplasia from mild to severe. The risk of


70

progression is associated with histological grade, but it is currently impossible to predict

accurately which lesions will progress to malignancy. Degradation of the basement

membrane and invasion of the underlying connective tissue by neoplastic cells is recognized

as a fundamental step in the development of many epithelial cancers.123

Advances have been made in identifying the molecular determinants of

carcinogenesis. Alterations in specific oncogene and tumor suppressor genes have been

identified and shown to have causal roles in the initiation, maintenance and progression of

tumors. This has however, largely ignored the substantial contribution of the tumor

microenvironment to the malignant phenotype. The ‘seed and soil’ hypothesis of Paget

dates to 1889, where the molecular determinants of the ‘seed’ are much better delineated

than those of the ‘soil’ for either primary or metastatic lesions. An enzyme produced by the

normal cellular component of the tumor could be a significant player in carcinogenesis.124

Important events that determine the onset of dissemination of tumor cells are related mainly

to the invasion of surrounding tissue and the induction of angiogenesis. Local growth,

spread and metastasis are features that are shared among all malignant tumors, and require

the presence of the appropriate microenvironment of the host. For tumors to invade and

metastasize, neoplastic cells must be capable of degrading the extracellular matrix and

accessing blood vessels and lymphatics. Mounting evidence supports the view that

extracellular proteinases, such as the MMPs, mediate many of the changes in the

microenvironment during tumor progression.124

Tumour microenvironment regulation by MMP

Proteolysis of ECM: The principle components of the ECM are collagens and numerous

other proteins including laminins, entactin, and proteoglycans that make up the basement

membrane. Tumor cells overexpress proteases and/or induce expression of these enzymes in

neighboring stromal cells in order to degrade the basement membrane and invade the

surrounding tissue. Several MMPs have been implicated in the ECM degradation associated

with tumor growth and angiogenesis. This proteolytic activity is also required for a cancer

cell to invade a nearby blood vessel i.e. intravasation and then extravasate at a distant

location and invade the distant tissue in order to seed a new metastatic site.125

Modulation of Cell Adhesion, Migration and Epithelial to Mesenchymal Transition :

The loss of intercellular adhesion and the acquisition of degradative properties are


71

prerequisites for tumor cells to become fully invasive. E-cadherin-mediated cell-cell

adhesion largely contributes to the maintenance of epithelial tissue integrity. The loss of E-

cadherin expression at the cell membrane is correlated with dedifferentiation,

aggressiveness, metastasis, and poor prognosis in different cancer types. MMPs modulate

the interactions between tumor cells by cleaving E-cadherin, and between tumor cells and

ECM by processing integrins, which also enhances the invasiveness of tumor cells. The

disorganization of E-cadherin/catenin complexes and expression of MMPs are frequently

involved in the capacity of epithelial cells to acquire an invasive phenotype. 126

MMPs have also been implicated in the epithelial to mesenchymal transition (EMT), a

hallmark of cancer progression to metastasis. During EMT, tumor cells acquire migratory

characteristics and more readily invade into surrounding tissues and metastasize to

secondary sites. Activation of growth factors and cleavage of adhesion molecules are some

of the proposed mechanisms underlying MMP induced EMT. Over expression of MMP-3, a

component of the tumor microenvironment, causes EMT and induces genomic instability in

cultured mammary epithelial cells leading to all stages of neoplastic progression, malignant

transformation, and mammary carcinomas in transgenic mice. 125, 126

MMPs Affect Growth Signals : Unregulated proliferation is a common feature of cancer

cells. There are two principal ways in which the tumor achieves this condition: by acquiring

self-sufficiency in growth-promoting signals or by becoming insensitive to antigrowth

signals. MMPs may be critically involved in disrupting the balance between growth and

antigrowth signals in the tumor microenvironment, as they potently influence the

bioavailability or functionality of multiple important factors that regulate growth.

One fundamental signaling pathway with essential roles in tissue homeostasis is the TGF-β

pathway. TGF - βnormally exerts tumor suppressive effects by enforcing cytostasis. Active

TGF-βis derived from an inactive pro-form by proteolytic conversion by furin or other

proteinases, such as MMP-9, which is usually expressed by inflammatory cells. Proteolytic

activation of TGF-βby MMPs has tumor-promoting effects by selectively driving stroma-

mediated invasion and metastasis of the tumor.

Ligands for the epidermal growth factor receptor (EGFR) are potent drivers of cell

proliferation and important regulators of tissue homeostasis. Malfunction of this system by

genetic mutations of the molecules involved is frequently observed in breast cancer and


72

other malignant diseases. Activation of EGFR results in the upregulation of MMP-9, which

in turn degrades E-cadherin, a potent control element of many cellular functions including

cell-cell adhesion and differentiation. The cleavage of E-cadherin by MMPs or ADAM

proteinases has an impact on cancer cell proliferation. The association between EGFR,

MMP-9 and E- cadherin may play an important role in ovarian cancer and metastasis, as

activated EGFR and MMP-9 in these specimens co-localise with a region of reduced E-

cadherin.126

Wuertz B and F Ondrey (2010)127 from their study on role of MMP-9 in oral

keratinocytes concluded that TNF alpha and cigarette smoke may assist carcinogenesis in

vivo by inducing expression of MMP-9, potentiating invasion. They also stated that oral

leukoplakia cells are baseline cells which strongly produce functional MMP-9 that cannot be

stimulated further with TNF or CSC.

The Tumor Angiogenesis : The tumor vasculature is derived from sprouting of

local blood vessels (angiogenesis) and circulating (vasculogenic) progenitor cells derived

from the bone marrow i.e. vasculogenesis. The new vessels are often irregular and leaky due

to lack of the pericyte cover, which results in tumor cells to penetrate more easily. The

major MMPs involved in tumor angiogenesis are MMP-2, MMP -9, and MMP -14, and to a

lesser extent MMP-1 and MMP-7. Genetic MMP-9 deficiency was found to result in

retarded bone growth attributed to inhibition of angiogenesis.128

Lymphangiogenesis plays an important role in tumor biology, given that it is directly linked

with the formation of lymphatic metastases. MMPs certainly have a general impact on

lymphangiogenesis as supported by the use of broad-spectrum MMP inhibitors. However,

only a few reports directly link MMPs to the formation of new lymphatic vessels. The

modulation of VEGF bioavailability by MMPs, especially by MMP-9, may also affect

lymphangiogenesis and, in turn, promotes dissemination of metastases into the lymph.

MMPs Regulate Apoptosis : Evading programmed cell death, or apoptosis, is another

strategy that increases the cell number and the size of tumors. MMP function interferes with

the induction of apoptosis in malignant cells, which may involve the cleavage of ligands or

receptors that transduce proapoptotic signals. MMP-7 cleaves Fas ligand from the surface of

doxorubicin- treated cancer cells, lowering the impact of chemotherapy on the tumor by

abrogating apoptosis.129


73

Adipocyte Regulation Affects Tumor Progression : Unquestionably, there are

consequences for the local paracrine crosstalk between the tumor cells and adipocytes.

Adipose secretory products collectively referred to as adipokines, have been identified as

contributors to the negative consequences of adipose tissue expansion in cancer,

cardiovascular disease, and diabetes. As adipocytes are a prominent part of the tumor stroma

and they do contribute to cancer progression. Moreover, adipokines such as leptin, regulate

the expression and activation of MMPs.129

To date, MMP-11 is induced by adipose tissue that directly affects cancer progression.

MMP-11 is expressed in adipose tissue as the tumor invades the surrounding environment

and negatively regulates adipogenesis by reducing preadipocyte differentiation and reversing

mature adipocyte differentiation. Adipocyte dedifferentiation leads to the accumulation of

non malignant peritumoral fibroblast-like cells, which favor cancer cell survival and tumor

progression.129

Tumor Invasion and Metastasis: The lethal outcome of the vast majority of all cancers is

due to the dissemination of metastatic tumor cells and the outgrowth of secondary tumors at

distant sites. Metastasis not only depends on features of the cancer cells disseminating from

the primary tumor but also requires the formation of a receptive environment, a metastatic

niche, which is specifically suited for the engraftment of tumor cells at the distant organ. It

is likely that MMPs and other proteinases are crucially involved in the formation of a

metastatic niche. MMP-9 turns out to be critical for the formation of the metastatic niche,

which is most likely linked with its ability to liberate VEGF and thereby support

angiogenesis. MMP-9 releases soluble factors to recruit stem and progenitor cells from the

bone marrow, which may be of particular significance in this context.129

Richard C et al (2004)130 in his retrospective study examined changes in MMP 1,2

and 9 using polymerase chain reaction in oral dysplasia and oral squamous cell carcinoma,

found higher levels of MMP-1 and 9 mRNA and are significantly associated with oral

dysplasia’s that progress to oral cancer compared with those that did not progress. Peschos

D et al (2006)131 checked the expression of MMP-9 in benign, premalignant and malignant

laryngeal lesions. In conclusion, their study indicates that the expression of MMP-9 is up-

regulated in a stepwise fashion, with two main steps, the first one, when a dysplastic lesion

evolves and the next one, when the dysplasia progresses to invasive carcinoma. MMP-9


74

expression was related neither to survival nor to the other available clinicopathological

parameters patient age. Suvi TV et al (2013)132 reported that MMP 9 is detected in

malignant transformation of various cells and is associated with tumor metastasis and poor

prognosis. In the aggressive human tongue squamous cell carcinoma cell line, MMP-2 was

only found in its latent form, whereas MMP-9 was found in its active form. In an oral

squamous cell carcinoma cell line SCC-25, CAFs increase the expression of MMP-9, in

vitro, which is thought to occur via a fibronectin integrin v 6 pathway. MMP-9 may not be

the only, or even the most important, proteolytic enzyme in the OSCC invasion process, but

it may be important for indirect cell signaling by controlling the bioavailability and

bioactivity of molecules that target specific receptors, which regulate cell growth, migration,

inflammation, and angiogenesis.132

Gao J (2010)133 studied the expression of MMP-2/MMP-9 in normal oral mucosa,

lymph node-negative tongue cancers, lymph node-positive tongue cancers and their

metastasized tumours in cervical lymph nodes. They hardly found MMP-2/MMP-9

expression in normal epithelium. In lymph node-negative tongue cancer, 45% and 40% of

these primary tumors were positively stained for MMP-2/MMP-9. Importantly, in lymph

node-positive tongue cancer, 71% and 79% of these primary tumors were positive for MMP-

2 / MMP-9, respectively. Over expression of MMP-2/MMP-9 was present in the metastatic

lymph nodes. Their result implied the significance of MMP-2 and / or MMP-9 in predictive

value for the actual or potential presence of cervical metastases. Such activity has prognostic

value, and provides impetus for further development of biotherapies targeted at specific

inhibition of MMP activities.133

CYTOKERATIN

Cytokeratins (K) are proteins of intermediate filaments found in the intracytoplasmic

cytoskeleton and is responsible for the structural integrity of epithelial tissue. A major role

of these proteins is to act as a resilient yet adaptable arrangement that empowers epithelial

cells with the ability to sustain mechanical and non-mechanical stresses reported by

Coulumbe PA (2002).134

Jürgen S et al (2006) 135 reported that the term “CYTOKERATIN” was coined by

Franke et al in 1978. In 2006 a new systematic nomenclature for keratins was created and

now the proteins previously called "cytokeratins" are simply called keratins. This


75

nomenclature is in accordance with nomenclature of Human Genome Organization (HUGO)

for both gene and protein names. These are made up of intracellular filamentous proteins,

which include microfilaments (6-8nm), intermediate filaments (10nm) and microtubules

(25nm). Intermediate filaments (IF) form the major part and are relatively stable component

of cytoskeleton. Approximately fifty different IF genes are expressed in different cells of the

human body.135

Cytokeratins consists of 20 biochemically and antigentically different polypeptides

with molecular weights ranging from 40 kDa – 67 kDa with an isoelectric pH of 4.9 – 7.8.

These proteins have been broadly divided into two subfamilies. These are recognised by

basic proteins. The low molecular weight proteins (52-67 kDa) are numbered as 1 to 8 and

the more acidic proteins with higher molecular weight (40-56 kDa) are numbered 9 to 20.

At least one member of each family is always co expressed in any given epithelial tissue.

Structure of Cytokeratin (keratin):

These have a homologous, central helical rod domain flanked by variably sized

amino and carboxyl terminal domains. DNAs of members of subfamily cross hybridize with

one another. The central rod domain of one keratin monomer binds with another keratin

monomer to form a coiled coil dimer. One dimer associates with another to form a tetramer.

Many tetramer join to form the final intermediate filament of size around 10 nm. The keratin

forms a hetrodimers, the filament being constructed from a protein pair which consists of a

type I and a type II keratin. Intermediate filament proteins are characterized by a highly

variable central alpha-helical rich rod-shaped domain and non-helical N & C terminal

globular domains at either end. Non--helical tail domain is typical for all keratins except for

K 19 report suggested by Chu PG et al 2002.136

Types and Pairing of Cytokeratins given by Hiroshi U et al (2005)137 :

Keratins (Ks) can be classified based on distribution, amino acid sequence and charge and

based on molecular weight.

Based on molecular weight: The molecular weight of cytokeratins ranges from 40 kDa to

67 kDa and based on which they are divided into low, intermediate and high molecular

weight keratins. Generally low molecular weight pairings are seen in simple epithelia, the

higher molecular weight parings are present in stratified non-keratinizing epithelia and the

highest molecular weight keratins are found in skin and corneal epithelium.


76

Based on distribution: Soft and hard keratins. Soft keratins are epithelial cytokeratins, are

20 in number and are expressed in various types of epithelia, while hard keratins are

constituents of hard keratinizing structures such as hair and nails.137

Based on amino acid sequence: Type I and type II keratins. Type I or Acidic Keratins are

numbered 9 to 20 and Type II or basic keratin are numbered 1 to 8. It is based on the

isoelectric pH which is the point at which protein does not migrate in an electric field and

has got least migrating power.

Differences between Type I and Type II Keratins:

The major difference between type I and type II K is the molecular weight and isoelectric

pH. Cytokeratin 1 (K1) has the highest molecular weight and the highest isoelectric pH,

while K19 has the lowest molecular weight and a low isoelectric pH. Basic cytokeratins are

8 KDa larger than the acidic member. These cytokeratins are expressed in combinations

which characterize the type of epithelium i.e. each epithelial cell expresses specific pairs of

type I and type II keratins (with the exception of K19).

Thumb rules for Keratin pairing:

By Morgan PR et al (1987)138, Mei-Hua Lu et al (2000)139

1. Keratins are expressed as pairs containing one member from each subfamily

2. The two members of each pair are in the same size rank or order within their

respective family eg: Largest acidic keratin (K10) is expressed with largest basic

keratin (K1). i.e. K 1 & 10.

Each epithelial cell expresses specific pairs of type I and type II keratins (with the exception

of K19). Unpaired lowest molecular weight K19 is the type I keratin and is the smallest

keratin and is exceptional since it widely lacks the non--helical tail domain which is typical

for all other keratins.138,139

Molecular weight and chromosomal localization of Keratins: In the human

genome, the keratin genes are clustered at two different chromosomal sites, chromosome

17q21.2 (type I keratins, except K18) and chromosome 12q13.13 (type II keratins including

K18).

Keratin Expression in Adult Epithelia: The different epithelia in the human body

express cytokeratins that are characteristic of the type and degree of maturation or

differentiation within an epithelium. K5 & K14 are the major keratin expressed in the basal


77

cell layer of both keratinised and non keratinised epithelia. In the undifferentiated basal cell

layer which contain stem cells, K5 & K14 are strongly expressed and are down-regulated in

the differentiating suprabasal cell layers, where as K19 is expressed in basal cell layer only

in non keratinised epithelium. In suprabasal cell layer K1 & K10, K4 & K13 is normally

expressed in keratinized and non keratinized epithelium respectively. K 8 & K18 is

expressed embryonically and in the simple epithelia. They are the first keratins to appear in

embryogenesis and also the oldest keratins found during evolution. K6 & K16 is a marker of

fast cell turn over and is usually expressed in keratinized epithelium as given by Tencate AR

(2003) 140 and Hisham IO (2006).141

Keratin Expression Profile in Head and Neck Pathology:

Oral Epithelial Dysplasia and Oral Squamous Cell Carcinoma: During the process of

malignant transformation of oral epithelium, oral leukoplakia represents one of the first

morphologically recognizable epithelial alterations. Hence the knowledge on alteration of

cytokeratin expression is important. According to Lindberg K et al (1989).142 CK 8/18 and

suprabasal expression of K19 is correlated with the premalignant transformation in oral

epithelium.

Mild dysplasia, arising from non keratinized epithelium - K1/10 synthesis is

enhanced and K4/13 is retained. This is explained by the presence of K1/10 mRNAs,

synthesis of which is increased in dysplastic state. Moderate dysplasia from non keratinized

epithelium - K1/10 filaments completely replace the K4/13 complex. Keratins K4/13 or

K1/10 that are characteristically present in suprabasal cell layers shows reduced expression

or loss in epithelial dysplasia. Severe dysplasia - both K4/13 and K1/10 are absent. K5 and

K14 are expressed in parabasal and spinous cell layers and may reflect basilar hyperplasia in

dysplastic epithelium, where as in normal epithelia these keratins are expressed only in basal

cells. Su L et al (1996)143 found that K19 expression in epithelial dysplasia varies from case

to case but abundant nRNA appears to be consistent and suggested that K19 expression is

diminished in the tissue with higher degree of keartinization. Differing levels of K19

expression may indicate instability in this keratin gene in the course of malignant

transformation. Perhaps in WDSCC K19 was negative and in PDSCC with K14 negative

found to have enhanced K19. Hence they suggested that synthesis of K19 was to

compensate for the loss of K14 so as to maintain the intermediate filament network.143


78

The expression of K5 is a hallmark of squamous epithelium and is predominantly

seen in the basal/stem layers of stratified epithelium. Stem cells of stratified epithelium have

been repeatedly described as the major cellular targets for cancer causing mutations and

therefore might give in a long term rise to the development of SCC's. Stigbrand T et al.

showed positive correlation between increased expression K8 and K 18 and an increase in

histological grade of anaplasia. A relationship between K8 and K18 expression and location

of carcinoma cells at the tumor invasion front or in other areas of stroma interaction has

been reported. In normal fetal buccal mucosa and tongue epithelium until 27 weeks of

gestation K8 and K18 is expressed. Hence it could be explained that 8/18 expression in

tumor cells reincarnate towards embryonic expression. Study done by Thomas F et al.

showed that determination of the K 8/18 and 19 expression in OSCC can be correlated with

its prognostic potential. Furthermore in our previous study we observed that the expression

of K 8/18 was negative in all normal mucosa, and was enhanced in poorly differentiated

OSCC than well differentiated. 144

Tumors and Cysts of Odontogenic Origin: Few studies have described keratin and

vimentin in the epithelial component of odontogenic neoplasms and dental germ. The

research conducted by Crivelini MM et al (2003)145 have described the expression of K5, 7,

8, 13, 14, 17 and 19 in the human enamel organ. K19 is expressed in all odontogenic

epithelial tumors, enamel organ and as well as in dental lamina. From these observations it is

suggested that K19 may be marker for cells of odontogenic origin or ameloblast

differentiation marker.145

Tumors of salivary gland origin: Expression of K7, 8, 14 and 19 is seen in Luminal

cells of intercalated ducts of normal salivary gland and also seen in pleomorphic adenoma,

basal cell adenoma, adenoid cystic carcinoma and epithelial-myoepithelial carcinoma.146

Material & Methods

79

MATERIAL AND METHODS

After obtaining an institutional ethical clearance the present retrospective study is based

on grading of epithelial dysplasia to evaluate the inter and intra observer variability with most

commonly used classification system. The grading of epithelial dysplasia has also be done using

a proposed new binary classification system.

Pilot Study :

A pilot study was conducted to assess the inter and intra observer variability in grading

OED using a set of 25 archival cases collected from different oral pathology centers (KLE VK

IDS - Belgaum, DMDC - Wardha, GDC - Nagpur and LMDC - Nagpur). These were reviewed

independently by four oral pathologists trained under one centre. Three classification systems

were undertaken to determine the degree of dysplasia namely WHO 1978, Smith & Pindborg’s

1969 and WHO 2005. The Oral Pathologists were blinded to the clinical and histopathological

diagnosis. Findings were recorded on individual score sheets independently and a final

assessment for each case was recorded by the observer as “no dysplasia / mild / moderate /

severe dysplasia”. The results were tabulated on excel sheet for analysis of the inter and intra

observer variability. WHO 1978 classification showed poor to good agreement, maximum being

in the moderate group, whereas Smith and Pindborg’s was fair to moderate agreement, following

a moderate range. WHO 2005 demonstrated moderate agreement among all observers (Annex.

No. 1). Following this the set of same 25 cases were assessed by the binary system suggested by

Kujan Omar et al 2006, as ‘low risk and high risk, based on features described by WHO 2005 as

cytological changes and architectural features. Moderate to almost perfect agreement was

observed. Suggesting the degree of agreement was best in the binary system followed by

classification given by Smith and Pindborg.

Material & Methods

80

Estimation of Sample Size:

The sample size was calculated using the following formula :

Where: n=sample size

N= no. of subjects in the entire population.

n* = depends (r and pa-pe)

r = relative error

pa = overall agreement probability

pe = chance agreement probability

Taking into consideration, if raters will agree about 50% of times and relative error of

20%. i.e. pa-pe = 0.5 and r=20

Selection of cases (Samples) :

Based on the results obtained from the pilot study the present research protocol was

designed to further validate these grading systems and also to propose a new classification. The

validation of these classifications is also based on immunohisto chemical markers.

For the present study 100 cases of potentially malignant disorder (PMD) were retrieved

from the archives of the department which reported to KLE VK Institute of Dental Sciences

OPD from 1995 to 2012. From cases with signed out diagnosis of hyperplasia, mild, moderate,

severe dysplasia and carcinoma in-situ were selected. Samples with sufficient depth of tissue

available were included. Biopsy samples obtained from lesions on buccal mucosa were

considered. 100 slides from 100 different cases were finally included, which formulated the

study material.

In the series of selected 100 cases, the 25 cases included for pilot study were not

considered for the final assessment.

Material & Methods

81

Demographic data :

Clinical data was obtained from the biopsy requisition form for these selected 100 cases.

The demographic data like age, sex, type of habit, frequency and duration of tobacco habit was

noted. The clinical provisional diagnosis ranged as Leukoplakia, erythroplakia,

erythroleukoplakia and oral sub mucous fibrosis. Confidentiality and maintenance of all the

obtained records of all the cases included in this research protocol will be strictly maintained by

stating the waiver of consent (Annex. No. 2a).

Procedure :

Four sections of 4μeach were prepared by sectioning paraffin embedded tissue blocks.

One section was stained with H & E for histopathological evaluation and the one section each

was obtained on gel coated slide and stained immunohistochemically for MMP9, Ki-67 and with

CK-19 using HRP-IHC. The staining procedures and analysis is been stated in manual of

operation and standard operating procedures (Annex. No. 3).

Assessment Method for Epithelial Dysplasia:

For final assessment the grading classifications given by Smith and Pindborg’s grading

system (1969), WHO (2005) and the Binary grading system (2006) were used for evaluation of

H&E stained sections of 100 cases. No calibration exercise was attempted. Each slide was

labeled only with a serial order number. No clinical data or patient information was provided

with the slides to the evaluator. Four independent oral pathologists reviewed the slides. Sections

were observed under 4x; 10x; 100x magnification using Olympus BM2500 compound

microscope. The architectural characteristics of each feature was examined under the low

magnification (4x & 10x) and finished with cytological characteristics at high magnification

(40x).

A score sheet was designed for recording the individual features to aid systematic

analysis. Separate score sheets based on the selected three classifications were prepared. The

lesions were graded as ‘0 for no dysplasia ; 1 for Mild ; 2 for Moderate and 3 for Severe

dysplasia / carcinoma in-situ. For Binary grading system ‘1 for Low risk and 2 for High risk’ was

scored (Annex. No. 4).

Individual scores of all raters were obtained and in addition to the individual scores of the

raters a consensus score of 3 raters (SH, DM & PA) was obtained to compare with the

investigator’s (AK) score (Annex. No. 5). Based on the proposed classification of the present

Material & Methods

82

study the observers evaluated all the 100 cases as ‘non-risky’ and ‘risky’ for inter and intra

observer variability for all 18 dysplastic features (Annex. No. 6). For 10 important features

predicted for malignant transformation all the 100 cases were evaluated (Annex. No. 7a&b).

Classifications considered for the pilot study and the present study :

1. Histologic features for Smith & Pindborg 1969 classification of oral epithelial dysplasia.

Type of change Severity of dysplasia & scores

1. Drop shaped reteridges None-0 Slight-2 Marked-4

2. Irregular epithelial stratification None-0 Slight-2 Marked-5

3. Keratinization of cells below keratinized layer None-0 Slight-1 Marked-3

4. Basal cell hyperplasia None-0 Slight-1 Marked-4

5. Loss of intercellular adherence None-0 Slight-1 Marked-5

6. Loss of polarity None-0 Slight-2 Marked-6

7. Hyperchromatic nuclei None-0 Slight-2 Marked-5

8. Increased nucleo-cytoplasmic ratio in basal and

prickle cell layers

None-0 Slight

Increase-2

Marked

Increase-6

9. Anisocytosis and anisonucleosis None-0 Slight-2 Marked-6

10 Pleomorphic cells and nuclei None-0 Slight-2 Marked-6

11 Mitotic activity Normal-0 Slight

Increase-1

Marked

Increase-5

12 Level of mitotic activity Normal-0 Slight-3 Marked-10

13 Presence of bizarre mitoses None-0 Slight-6 Marked-10

2. WHO criteria for grading dysplasia (1978) : It graded epithelial dysplasia as: Mild,

Moderate, Severe.

Mild dysplasia: Slight nuclear abnormalities, most marked in the basal third of the epithelial

thickness and minimal in the upper layers, where the cells show maturation and stratification. A

few, but no abnormal mitosis may be present, usually accompanied by keratosis and chronic

inflammation.

Material & Methods

83

Moderate dysplasia: More marked nuclear abnormalities and nucleoli tend to be present, with

changes most marked in the basal 2/3rd of the epithelium, nuclear abnormalities may persists up

to the surface, but cell maturation and stratification are evident in the upper layers. Mitoses are

present in the parabasal and intermediate layers, but none is abnormal.





carcinoma in situ, in which the whole or almost the whole thickness of epithelium is involved.

In the present knowledge, it was not possible to say whether the presence of severe dysplasia

carried a different degree of risk of subsequent development of invasive carcinoma than the

presence of carcinoma in situ.

3. WHO criteria for grading dysplasia (2005) :

Architecture criteria Cytology criteria

1. Irregular epithelial stratification 1. Abnormal variation in nuclear size

(Anisonucleosis)

2. Loss of polarity of basal cells 2. Abnormal variation in nuclear shape

(Nuclear pleomorphism)

3. Drop-shaped rete ridges 3. Abnormal variation in cell size

(Anisocytosis)

4. Increased number of mitotic figures 4. Abnormal variation in cell shape

(Cellular pleomorphism)

5. Abnormal superficial mitoses 5. Increased nuclear cytoplasmic ratio

6. Premature keratinisation in single cells

(dyskeratosis)

6. Increased nuclear size

7. Keratin pearls within rete ridges 7. Atypical mitotic figures

8. Increased number and size of nucleoli

9. Hyperchromatism

Material & Methods

84

Grading of dysplasia:

Mild dysplasia: In general architectural disturbance limited to the lower third of the epithelium

accompanied by cytological atypia define the minimum criteria of dysplasia.

Moderate dysplasia: Architectural disturbance extending into the middle third of the

epithelium is the initial criterion for recognizing this category. However, consideration of the

degree of cytologic atypia may require upgrading i. e. those lesions that show marked cytological

alteration should be elevated to a higher grade level regardless of how extensively the atypical

cells fill the epithelium.

Severe dysplasia: Recognition of severe dysplasia starts with greater than two thirds of the


architectural disturbance extending into the middle 3rd of the epithelium with sufficient cytologic

atypia may be up graded from moderate to severe dysplasia.

Carcinoma in situ: The theoretical concept of carcinoma in situ is that malignant transformation

has occurred but invasion is not present. It is not always possible to recognize this

morphologically. The following is recommended for the diagnosis of carcinoma in situ: full

thickness or almost full thickness architectural abnormalities in the viable cellular layers

accompanied by pronounced cytologic atypia. Atypical mitotic figures and abnormal superficial

mitoses are commonly seen in carcinoma in-situ.

4. Binary Classification (2006) :



Low-risk lesion (does not have the potential susceptibility for malignant transformation) was

based on observing less than four architectural changes or less than five cytological changes. The

features to classify as High and Low Risk were based on the architectural and cytological

features suggested by WHO 2005 classification.

PROPOSED WORKING CLASSIFICATION

The proposed classification is based on Binary System (2006) and WHO (2005) criteria of

epithelial dysplasia. Additional architectural and cytological features are identified and included

in this classification. Each features of epithelial dysplasia is scored on a rating scale of 1 to 10.

The two groups identified are :

Material & Methods

85

• Non risky lesions : The total score ranged from 1 to 15

• Risky lesions : The total score ranged from 16 to 84

• ARCHITECTURAL FEATURES:

Sl. No. Particulars Scores

1. 1.1. Basal cell hyperplasia 4

2. 1.2. Drop shaped rete ridges 2

3. 1.3. Few normal mitotic figures 1

4. 1.4. Acanthosis 1

5. 1.5. Loss of polarity of basal cells 5

6. 1.6. Drop shaped rete ridges with nodules or forking 4

7. 1.7. Loss of cohesiveness with prominent intercellular bridges 5

8. 1.8. Abnormal and superficial mitoses 10

9. 1.9. Increased number of mitotic figures 5

10. 1.10. Dyskeratosis 2

11. 1.11. Loss of epithelial stratification 5

12. 1.12. Keratin pearls within rete ridges 5

• CYTOLOGICAL FEATURES:


1. 2.1. Mild to moderate degree of nuclear and cellular pleomorphism 5

2. 2.2. Marked nuclear & cellular pleomorphism &anisonucleosis 10

3. 2.3. Hyperchromasia of cells 5

4. 2.4. Increased nuclear-cytoplasmic ratio 4

5. 2.5. Atypical mitotic figures 6

6. 2.6. Increased number and size of nucleoli 4

INFLAMMATION :


1. 3.1 Mild 1

2. 3.2 Moderate 2

3. 3.3 Severe 3

Material & Methods

86

Procedure for H&E staining and Immunohistochemistry :

Four sections each of selected archival blocks were cut, one slide stained routinely with H&E

to confirm the diagnosis and the other slide stained with immnunohistochemical stain. The sections

required for immunoexpression is taken on a gel coated slides using 2%, 3-aminopropylethoxysilane

solution (APES) (Sigma Aldrich, St.Louis, MO, USA, Product code: A3648). The buffers required

for IHC are PBS as wash buffer and citrate buffer for antigen retrieval. The immunohistochemical

protocol was followed as per the standard recommendations of the company.

The IHC staining is performed using Super Sensitive one step Polymer HRP system

(QD-600, Biogenex, Bangalore) with pre-diluted primary antibodies using anti-Human antibody

for CK-19 (clone RCK 108, Biogenex, Bangalore), Anti MMP-9 (EP1-255Y, Biogenex,

Bangalore), Ki-67 (BGX-297, Biogenex, Bangalore). The storage and dilution of these

antibodies and IHC kit is done as per the manufacturers’ recommendation (Annex. No. 3).

Assessment of Immunohistochemical stained sections:

All the IHC stained slides were evaluated for immunoexpression of MMP 9, Ki

67, CK 19. The intensity of immunohistochemical staining was graded based on subjective

evaluation of color exhibited (brown color) by antigen, antibody and chromogen complex. For

Ki-67, the immunoexpression was noted for nuclear staining, CK-19 in cytoplasm and

MMP-9 for stromal reaction. Evaluation of immunoexpression was noted under x10

magnification under light microscope.

1. Analysis of immunoexpression of MMP-9

The immunoexpression of MMP-9 was based on intensity and area of expression

in the epithelium and stroma on a 4 point scale from 0 to 3. Intensity of expression was

graded as 0 when there was no evidence of immunoexpression. Intensity was considered

positive depending upon the colour intensity ranging from 1 : golden brown, 2 : light

brown, 3 : dark brown. Area of expression was based on a percentage of positivity in the

stroma and epithelium. The range was considered 1: when 1 to 25% of positive

expression was noted, 2 : 26 to 50% of positivity and 3 : more than 50% of positive

expression.

Material & Methods

87

Intensity Area in epithelium and stroma

0 = -ve 0 = -ve

1 = + , golden brown 1 = 1-25%

2 = + + , light brown 2 = 26-50%

3 = + + + , dark brown 3 = >50%

2. Analysis of immunoexpression of Ki-67

The immunoexpression of Ki-67 was mainly assessed in nuclei of epithelial cells.

This marker was evaluated in the epithelium based on the area of expression and

percentage of positive immunoexpression on a four point scale from 0 to 3. When

there was no evidence of immunoexpression it was graded as 0. In the epithelium this

proliferative marker was assessed based on the layers of the stratified squamous

epithelium involved. Whenever expression was noted only in the basal cell layer it

was graded as 1, suprabasal layer was graded as 2 and layers above this was graded as

3. Percentage of number of cells for positive immunoexpression was assessed. Area of

expression was graded as 1: when 1 to 25% positive epithelial cells were noted, 2: 26

to 50% positive cells and 3: more than 50% cells. The area with maximum number of

positive cells was considered in each section.

Area Percentage no. of cells

0 : No expression 0 = -ve

1 : Basal layer 1 = 1-25%

2 : Suprabasal layer 2 = 26-50%

3 : Above suprabasal layer 3 = >50%

3. Analysis of immunoexpression of Ck-19

The immunoexpression of Ck-19 was based on intensity and area of expression in

the epithelium on a 4 point scale from 0 to 3. Intensity of expression was graded as 0

when there was no evidence of immunoexpression. In te ns it y wa s co ns id er ed

po si ti ve depending upon the colour intensity ranging from 1 : golden brown, 2 : light

Material & Methods

88

brown, 3 : dark brown. Area of expression was based on a percentage of positivity in the

epithelium. It was considered as 1 for 1 to 25% of positive expression, 2 : 26 to 50% of

positivity and 3 : more than 50% of positive expression in the epithelium.

Intensity Area

0 = -ve 0 = -ve

1 = + golden brown 1 = 1-25%

2 = + + light brown 2 = 26-50%

3 = + + + dark brown 3 = >50%

STATISTICAL ANALYSIS:

Statistical analysis was done using :

1. Weighted Kappa and Non-Weighted Kappa tests for assessment of inter and intra observer

variability for all the classification system

Unweighted kappa takes into consideration agreement for individual cases

Weighted kappa considers the magnitude of disagreement (ordered categories) and

hence is a better measure for reliability

Kappa Agreement (Flesis and Cohen 1973)

< 0.40 Poor

0.40-0.75 Fair to Good

0.76 to 0.80 Excellent

0.81-0.99 Almost perfect

2. Fisher's Exact Test was done to evaluate the correlation of the immunohistochemical

markers with grades of epithelial dysplasia. p<0.001 was considered as statistically

significant.

Results

89

RESULTS

In the pilot study 25 cases of OED were evaluated for inter and intra observer variation

using four classifications :- a) Grading of dysplasia given by WHO (1978), b) Smith and

Pindborg Classification using photographic standards (1969) c) Classification of WHO (2005)

and d) Binary Classification by Kujan Omar et al (2006).

For the present study 100 cases of OED were subjected to evaluation for two selected

classifications: Smith and Pindborg (1969) and WHO (2005). A new classification is proposed

which included scoring system and has two categories for grading dysplasia as ‘non-risky’ and

‘risky’. IHC evaluation for Ki-67, CK-19 and MMP-9 is done to understand behavior at the

molecular level. The data obtained has been statistically analyzed.

Table No. 1 : Statistical analysis of 25 cases considered for Pilot Study

Classification/ Observers 1

st& 2

nd1

st& 3

rd1

st& 4

th2

nd& 3

rd2

nd& 4

th3

rd& 4

th

Smith &Pindborg

(1969)

k=0.307Fairkw =0.419Moderate


k=0.415Moderatekw =0.577Moderate


k=0.390Moderatekw =0.667Good


WHO(1978)

k=0.552Moderatekw=0.67Good


k=0.114Poorkw =0.316Fair




WHO(2005)

k=0.597Moderatekw =0.706Good






Binaryclassification

(2006)

k=0.920AlmostPerfect

k=0.562Moderate

k=0.448Moderate

k=0.513Moderate

k=0.522Moderate

k=0.387Fair

Table no. 1: WHO (1978) showed good to poor agreement with maximum having moderate

agreement. Smith and Pindborg (1969) had fair to moderate agreement following a moderate

range. Whereas WHO (2005) demonstrated moderate agreement amongst all observers.

Moderate to almost perfect agreement was observed in binary classification by Kujan O. The

inference drawn from the pilot study suggests that the degree of agreement was best in the

binary system followed by classification given by Smith and Pindborg. (K=kappa, kw=weighted

kappa)

Results

90

Table No. 2: Statistical analysis of 100 study cases considered for INTER OBSERVER variability

Classification Weighted kappa (kw) Unweighted kappa (k) p value

Smith & Pindborg (1969) 0.483 (Moderate) 0.275 (Fair) < 0.001

WHO (2005) 0.419 (Moderate) 0.310 (Fair) < 0.001

Proposed classification - 0.790 (Good ) < 0.001

Table no. 2 : Smith and Pindborg’s classification showed fair to moderate agreement in both

weighted and unweighted kappa statistics (p<0.001). Similarly, WHO (2005) classification

also showed fair to moderate inter observer agreement. Good agreement was evident with

proposed classification

Table No. 3 : Statistical analysis of 100 study cases for INTRA OBSERVER variability

Classification Weighted kappa Unweighted kappa p value

Smith & Pindborg (1969) 0.742 (Good) 0.612 (Good) < 0.001

WHO (2005) 0.706 (Good) 0.597 (Moderate) < 0.001

Proposed classification - 0.810 (Almost perfect) < 0.001

Table No. 3 : The intra observer agreement was found to be almost perfect in the proposed

classification as compared to the other classification.

Results

91

Table No. 4 : Scores given for atypical architectural and cytological features for theProposed Classification


Architectural Features

1. 1.1. Basal cell hyperplasia 4

2. 1.2. Drop shaped rete ridges 2

3. 1.3. Few normal mitotic figures 1

4. 1.4. Acanthosis 1

5. 1.5. Loss of polarity of basal cells 5

6. 1.6. Drop shaped rete ridges with nodules or forking 4

7. 1.7. Loss of cohesiveness with prominent intercellular bridges 5

8. 1.8. Abnormal and superficial mitoses 10

9. 1.9. Increased number of mitotic figures 5

10. 1.10. Dyskeratosis 2

11. 1.11. Loss of epithelial stratification 5

12. 1.12. Keratin pearls within rete ridges 5

Cytological Features

1. 2.1. Mild to moderate degree of nuclear and cellular pleomorphism 5

2. 2.2. Marked nuclear & cellular pleomorphism & anisonucleosis 10

3. 2.3. Hyperchromasia of cells 5

4. 2.4. Increased nuclear-cytoplasmic ratio 4

5. 2.5. Atypical mitotic figures 6

6. 2.6. Increased number and size of nucleoli 4

Inflammation

1. 3.1 Mild 1

2. 3.2 Moderate 2

3. 3.3 Severe 3

Results

92

Table No. 5 : Distribution of cases into ‘non-risky’ (total score ranged from 1-15 ) and

‘risky’ (total score ranged from 16 to 84) groups.

Category No. of Cases Percentage Score

Non risky 41 41% 2-15

Risky 59 59% 20-84

Table No. 5 : 59 cases were identified as risky dysplasia as compared to 41 cases of ‘non-

risky’ dysplasia. Suggesting that 59 cases out of 100 cases are likely to develop into

malignancy.

Table No. 6a : Calculation for risk categorization using median and inter quartile for 100cases

X SD Median 1st Quartile 3rd Quartile

Non Risk Group Score 7.7 4.1 8 4 11.5

Risk Group Score 53.5 13.01 56 45 66

Table No. 6b : Classification of risk categories for evaluation of 59 ‘risky’ cases

Category Score Number of cases

Low risk 20-44 14(23.7%)

Medium risk 45-66 39(66.1%)

High risk 67-84 6(10.2%)

Table No. 6a & 6b : Entire score for the proposed classification comprising of 1 to 84 was

evaluated for mean and standard deviation. Median and the inter quartile range was used to

define cut off points and the cases were classified into low, medium and high risk based on the

scores above. Out of the 59 ‘risky’ cases 6(10.2%) cases had highest score (67-84) were

suggestive to be at higher risk for malignant transformation.

Results

93

Table No. 7 : Risk categorization of ‘Risky’ and ‘Non-risky’ 100 cases for each feature

Features Low Risk (14) Medium Risk (39) High Risk (6) Non Risk (41)

ArchitecturalNo. ofCases % No. of

Cases % No. ofCases % No. of

Cases %

1.1 11 78.5 34 87.2 6 100.0 0 0

1.2 2 14.3 14 35.9 1 16.7 3 7.3

1.3 7 50 21 53.8 1 16.7 23 56.1

1.4 10 71.4 37 94.9 5 83.3 30 73.2

1.5 7 50 32 82.1 6 100.0 1 2.4

1.6 5 35.7 32 82.1 5 83.3 0 0

1.7 10 71.4 34 87.2 6 100.0 5 12.2

1.8* 1 7.1 4 10.3 5 83.3 0 0

1.9 6 42.8 35 89.7 6 100.0 2 4.8

1.10 1 7.1 17 43.6 6 100.0 1 2.4

1.11 4 28.6 30 76.9 6 100.0 0 0

1.12 14 100 39 100.0 6 100.0 0 0

Cytological2.1 9 64.3 39 100.0 6 100.0 0 0.0

2.2* 1 7.1 3 7.7 4 66.7 0 0.0

2.3 9 64.3 38 97.4 6 100.0 19 46.3

2.4 11 78.6 39 100.0 6 100.0 6 14.6

2.5 11 78.6 39 100.0 6 100.0 4 9.7

2.6 5 35.7 34 87.2 6 100.0 2 4.8

Table No. 7 : Cases of both non-risky and risky categorized into three risk categories as per the

scores for 18 individual features. Feature 1.8* (Abnormal and superficial mitoses) and 2.2*

(Marked nuclear & cellular pleomorphism & anisonucleosis) were significantly associated with

the risk group as compared to the non-risky group. (Fisher Exact Test p<0.0010 & p<0.003)

Results

94

Table No. 8 : Ten Important Features with higher scores considered in ‘Risky group’ aresegregated with a total score of 61.

Feature no. Particulars Scores

1.5 Loss of polarity of basal cells 5

1.7 Loss of cohesiveness with prominent intercellular bridges 5

1.8 Abnormal and superficial mitoses 10

1.9 Increased number of mitotic figures 5

1.11 Loss of epithelial stratification 5

1.12 Keratin pearls within rete ridges 5

2.1 Mild to moderate degree of nuclear and cellular pleomorphism 5

2.2 Marked nuclear & cellular pleomorphism & anisonucleosis 10

2.3 Hyperchromasia of basal cells 5

2.5 Atypical mitotic figures 6

Total = 10 Total score 61

Results

95

Table No. 9a: Calculation for risk categorization using median and inter quartile forstudy group for 10 important features

Studygroups x SD Median 1st Quartile 3rd Quartile

Non RiskyGroup Score 5.4 3.68 6 1.5 7

RiskyGroup Score 37.5 10.72 39 32 44

Table No. 9b. : Classification of risk categories for evaluation of cases for 10 features

Category Score Number of Cases

Low risk 12-31 13(22%)

Medium risk 32-44 38(64.4%)

High risk > 45 8(13.6%)

Table No, 9a and 9b : The total score of the ten individual features selected comprising of 01

to 61 was evaluated for mean and standard deviation. The median and the inter quartile range

was used to define cut off points and the cases were again classified into low, medium and high

risk based on the ten individual features. Out of the 59 ‘risky’ cases 8(13.6%) cases had

highest score (> 45) were suggestive to be at higher risk for malignant transformation.

Results

96

Table No. 10 : Risk categorization of ‘risky’ and ‘non-risky’ groups for the 10 features

Risky group Non risky group

Low Risk (13) Medium Risk (38) High Risk (8) Non Risk (41)

No. ofCases % No. of

Cases % No. ofCases % No. of

Cases %

1.5 7 53.8 30 78.9 8 100.0 1 2.4

1.7 7 53.8 35 92.1 8 100.0 5 12.2

1.8* 0 0.0 4 10.5 6 75.0 0 0

1.9 7 53.8 32 84.2 8 100.0 2 4.8

1.11 2 15.4 31 81.6 7 87.5 0 0

1.12 1 7.7 17 44.7 7 87.5 0 0

2.1 8 61.5 37 97.4 7 87.5 0 0

2.2* 0 0.0 3 7.9 5 62.5 0 0

2.3 9 69.2 36 94.7 8 100.0 19 46.3

2.5 11 84.6 37 97.4 8 100.0 4 9.7

Table No. 10 : Cases of both non-risky and risky categorized into three risk categories as per

the scores for 10 selected individual features. Feature 1.8* (Abnormal and superficial mitoses)

and 2.2* (Marked nuclear & cellular pleomorphism & anisonucleosis) were significantly more

commonly associated with the ‘risky’ group as compared to the ‘non-risky’ group (Fisher Exact

Test p<0.001).

Results

97

Table No. 11a : Area of expression of CK-19 in Epithelium of the study groups (Fisher exactTest)

Area 0 1 2 3 p value

Risky 0 15 32 12p<0.001

Non risky 20 15 6 0

Table no. 11a shows that in ‘risky’ group the area of expression of Ck-19 in 32 cases the

expression was seen in 26 to 50% of epithelial cells. In a ‘non-risky’ group 20 cases were

negative for Ck-19 expression but 6 cases had 26 to 50% positive cells. The area of CK-19

expression in the epithelium was significantly high in the risky as compared to non-risky group.

Table No. 11b : CK-19 Expression of intensity in Epithelium of the study groups (Fisherexact Test)

Table No. 11 b : The intensity of expression was light brown in 30 cases and dark brown in 14

cases of ‘risky’ group. Whereas in the ‘non-risky’ 20 cases were negative and in 21 cases the

intensity ranged from light brown to dark brown. Intensity of CK-19 expression in the

epithelium was significantly high in the ‘risky’ as compared to ‘non-risky’ group. The eight

cases with scores 45 & above had more intensity (+++) and more area (>50%) of expression of

CK-19.

Intensity 0 1 2 3 p value

Risky 0 15 30 14p<0.001

Non risky 20 19 2 0

Results

98

Table No. 12a. : Ki-67 Expression in the study groups (Fisher exact Test)

Table No. 12 a shows that 39 ‘risky’ cases had 26 to 50% positive cells for ki-67 as compared

to the non-risky group where 24 cases were negative and 11 cases had1 to 25% positive

expression. The Ki-67 expression in the epithelium was significantly higher in the ‘risky’ as

compared to ‘non-risky’ group.

Table No. 12b. : Area of Ki-67 positivity in study groups (Fisher exact Test)

Table No. 12 b shows 42 cases in the risky group had ki-67 immnoexpression upto to the supra

basal layer, whereas 6 cases had expression above the suprabasal layer. In the ‘non-risky’ 24

cases were negative but 6 cases had ki-67 immunoexpression upto the suprabasal layer. The Ki-

67 expression in the epithelium was significantly higher in the ‘risky’ group as compared to

‘non- risky’ group. The eight cases with scores 45 & above had expression upto suprabasal

layer (+++) where >50% of cells had expression of Ki-67.

Expression 0 1 2 3 p value

Risky 0 11 39 9p=0.000

Non risky 24 11 6 0

Area 0 1 2 3 p value

Risky 0 11 42 6p=0.000

Non risky 24 11 6 0

Results

99

Table No. 13a : MMP-9 Expression in Epithelium of the study groups (Fisher Exact Test)

Groups Intensity Area

0 1 2 3 0 1 2 3

Risky 2 42 13 0 2 19 33 4

Non-Risky 9 26 2 1 9 19 11 1

p value p=0.001 (S) p=0.002 (S)

Table No. 13 a shows the intensity of MMP-9 expression was golden brown in 26 ‘non-risky’

cases and 42 ‘risky’ cases. In 13 risky cases area of MMP-9 expression in epithelial cell was

seen between 26-50% whereas in non-risky 26 cases had 1 to 25% positive cells. The MMP-9

epithelium was significantly higher in the risky as compared to ‘non-risky’ group.

Table No. 13b : MMP-9 Expression in stroma of the study groups (Fisher exact Test)

Groups Intensity Area

0 1 2 3 0 1 2 3

Risky 1 20 26 11 1 17 37 2

Non-Risky 2 15 18 4 2 19 15 2

p value p=0.551 (NS) p=0.700 (NS)

Table No. 13 b : shows the intensity of MMP-9 expression was light brown in 18 ‘non-risky’

cases and 26 ‘risky’ cases. In 37 risky cases area of MMP 9 expression in stromal cell was seen

between 26-50% whereas in non-risky 19 cases had 1 to 25% positive cells. No association was

observed in the stromal expression of MMP-9 in the ‘risky’ as compared to ‘non-risky’ group.

H&E Photographs

100

H&E Stained Sections of Dysplasia

Figure 1. Section showing ‘Non-risky’ dysplasia (a - 10x, b - 40x, H&E stain)

Figure 2. Section showing ‘risky’ dysplasia (a - 10x, b & c - 40x, H&E stain)

a b

a cb

MMP-9 Photographs

101

MMP-9 Immunoexpression in Dysplasia

Figure 3. ‘Non-risky’ dysplasia with minimal MMP-9 Expression (a- 10x, b- 40x, IHC stain)

b

Figure 4. ‘Risky’ dysplasia with enhanced MMP-9 Expression (a & b - 10x, c- 40x, IHC stain)

a c

a b

Ki-67 Photographs

102

Ki-67 Immunoexpression in Dysplasia

a b

b

Figure 6. ‘Risky’ dysplasia. Enhanced Ki-67 expression in suprabasal layer.(a- 10x, b & c- 40x, IHC stain)

c

a

Figure 5. ‘Non-risky’ dysplasia. Restricted Ki-67 expression in basal cell layer.(a- 10x, b- 40x, IHC stain)

CK-19 Photographs

103

CK-19 Immunoexpression in Dysplasia

Figure 7. ‘Non-risky’ dysplasia. Restricted CK-19 expression in basal cell layer.(a -10x, b- 40x IHC stain)

b

Figure 8. ‘Risky’ dysplasia. Enhanced CK-19 expression in basal and suprabasal layers.(a & b -10x, c- 40x, IHC stain)

c

a b

a

Epithelial Atypia Photographs

104

Fig. F : Loss ofCohesiveness (40x)

Fig. G : Loss ofstratification (40x)

Fig. A : Dysplastic OralEpithelium (10x)

Fig. B : Prominent IntercellularBridges (10x)

Fig. E : Increased Nucleoli (40x)

Fig. C : Loss of BasalCell Polarity (40x)

Fig. D : BasilarHyperplasia &

Hyperchromasia(40x)

Fig. I : Superficial &Abnormal Mitosis (40x)

Fig. H : Marked Pleomorphism (40x)

a

CBA

D E F

G H I

Figure 9(A-I) : Features of Epithelial Atypia (H&E Stain)

Discussion

105

DISCUSSION

Oral squamous cell carcinoma (OSCC) is one of the most commonly encountered

malignancies, which not only involves mortality but also disfigurement and dysfunction

associated with treatment. For intra oral cancers with cervical lymph node metastasis, the

five year survival rate is less than 50%.2 To improve the mortality and morbidity adequate

early diagnosis is urgently needed. It is not known if all oral cancers are preceded by

precursor lesions but a majority of OSCC do present as visible changes of the oral mucosa.

These precursor lesions are defined as an altered epithelium with an increased likelihood for

progression to OSCC. The altered epithelium shows a variety of cytological and architectural

changes that have traditionally been brought under the common denominator: Dysplasia.

In the oral cavity the most common lesions recognized as potentially malignant are

leukoplakia and erythroplakia. In the Indian subcontinent oral submucous fibrosis has been

reported to have malignant potential. It is known that 50% of OSCC arise from apparently

clinically normal mucosa. The prognostic significance of an individual lesion is difficult to

determine. None of the currently available molecular markers have proved to neither be

prognostically significant nor have yet been evaluated in large prospective studies. The gold

standard for the assessment of oral potentially malignant lesions is microscopic evaluation of

standard H&E stained sections for the presence of architectural and cytological changes,

generally referred as epithelial dysplasia.147

Assessment of dysplasia depends upon the microscopic diagnosis which is based on

grading the changes considering combination of architectural and cytological features.147

Grading of dysplasia is subjective and lacks intra and inter observer reproducibility due to

insufficiency of validated morphological criteria and the biological nature of dysplasia.54

Hence, due to the absence of a consensus, several systems are currently employed.

During the last decade many classifications of potentially malignant lesions

synonymously referred as epithelial dysplastic lesions, premalignant conditions, squamous

intra epithelial neoplasia etc. have been proposed. The majority of the classifications in the

current literature have followed criteria similar to those in common use for the grading of

epithelial lesions of the uterine cervix.148 In cervical epithelium there is clear distinction

between normal and abnormal layers of the epithelium and consequently the degree of

dysplasia can be assessed by determining the horizontal level of this border in the epithelium,

Discussion

106

resulting in substantial intra and inter observer consistency. Such a sharp distinction between

normal and abnormal layers in the epithelium of the upper aero-digestive tract is less obvious

and consequently the definitions for degree of dysplasia are much more susceptible for

discussion, thus resulting in substantial intra and inter observer variability.149

A pilot study was conducted to assess the inter and intra observer variability in

grading OED using a set of 25 archival cases with signed out diagnosis as hyperplasia in 2

cases, mild dysplasia in 3 cases, moderate in 7 cases, severe dysplasia in 8 cases and

carcinoma in situ in 4 cases. Four oral pathologists graded them independently using three

classification systems. The observers were blinded for clinical and histopathological

diagnosis. The classification systems followed were (1) WHO (1978) (ii) Smith and

Pindborg’s (1969) system (iii) Architectural & Cytological criteria given by WHO in 2005.

The results of the pilot study were evaluated statistically for assessment of inter and

intra observer variability:

1. Smith and Pindborg’s System: The inter and intra observer K values ranged from

0.241(fair) to 0.443 (moderate) whereas kw values ranged from 0.417(moderate) to

0.667(good) (Table No. 1). Agreement was better between the raters as each category is

assigned a score and the sum of the scores for the 13 categories provided an epithelial

dysplasia index for each slide. It produces a dysplasia assessment on an ordinal scale

which facilitates statistical analysis. The disadvantage was that the analysis was time

consuming as each feature had to be compared with the monograph pictures.

2. WHO Criteria (1978) : has summarized the features of epithelial dysplasia under two

heads namely (a) Epithelial organization & (b) Cytological features. The inter and intra

observer K values ranged from 0.114(poor) to 0.552 (moderate) whereas kw values

ranged from 0.316(fair) to 0.67(good) suggesting wide variation in observation ranging

from poor to good agreement, maximum observations being in the moderate group (Table

No. 1).When only a few features are present, more emphasis is placed on them but which

independent feature has to be given weightage to grade dysplasia will vary from case to

case. This lead to disparity in observations which was not only between histopathologist

(inter observer) but also when the same investigator viewed the case on a different

occasion (intra observer).

Discussion

107

3. WHO Criteria (2005) : WHO consensus group modified the features published in 1978

and broadly categorized as changes in the architecture (strata) of the epithelium and the

cytological features which manifest as cellular atypia and divided into grades of mild,

moderate and severe. The more numerous the features, the more severe is the grade of

dysplasia. However, dysplasia represents a spectrum of change rather than discrete

identifiable stage. In the pilot study the inter and intra observer K values ranged from

0.212(fair) to 0.597(moderate) whereas kw values ranged from 0.404(moderate) to

0.706(good) (Table No. 1). This system lacks guide lines in interpretation of the

presence, degree and significance of each feature and hence contributed to inconsistent

assessment.

Kujan O et al (2007)167 demonstrated a variation in the inter observer agreement on

grading OED using two different grading systems; the 2005 WHO classification and new

binary system. A further analysis of the inter-observer agreement on the architectural and

cytological features that were used in the grading process was undertaken to help to

understand where the inter-observer variation stems from. Their data showed wide inter-

observer variability in recognizing the individual features of OED. The observers have

shown poor to moderate agreement on grading these individual features, but a good to

substantial agreement was seen when a cumulative scoring was used to agree on the

degree of dysplasia either by using the new binary system of ‘‘low-risk’’ or ‘‘high-risk’’

or by using the 2005 WHO classification. This study states that the observers had

compensated the discrepancies in recognizing these features and this could be related to

the fact that most pathologists do initially screen the slides at low magnification and

make a provisional diagnosis. They then use a higher magnification to ascertain the final

diagnosis by looking for specific details in the slide.

4. Binary grading system (2006) : Thus the set of same 25 cases were once again

evaluated using the binary system of grading as ‘high / low risk’ as given by Kujan O et

al.49 The K values were calculated which ranged from k=0.387(fair) to k=0.920 (almost

perfect) (Table No. 1). Fair to almost perfect agreement among the observers was noted.

A statistically significant overall correlation was found between histopathological

grading of dysplasia using this low / high risk system for all observers (p<0.001). But

difficulties in assessment observed were the same as encountered with WHO criteria –

Discussion

108

2005. The disagreement was due to interpretation of individual features of dysplasia. By

reducing features to definable components increased the objectivity and insight but

evaluation and justification of specific features could not be achieved as the cut-off point

for a ‘high risk’ lesion was based on observing at least four architectural changes and five

cytological changes. The cut-off point for a ‘low risk’ lesion considered less than four

architectural changes or less than five cytological changes as suggested by the authors.

Based on the observations from pilot study and after considering an extensive

search of literature, the criteria under architectural features and morphological features

defined by WHO in 2005 were redefined to form the basis for the new proposal to grade

epithelial dysplasia for the present study. We have made an attempt to score the

individual features of dysplasia. The criteria for assigning the scores was based on the

severity and occurrence as seen in OSCC or in tissues adjacent to OSCC, features

considered for severe dysplasia by other authors and those suggested by Smith and

Pindborg. This was suggested with the objective of reducing the inter and intra observer

variability and for making assessment objectively reproducible. (Table No. 4)

The two basic requirement of any grading system should be : reproducibility

and prognostic value. Reproducibility refers to the degree to which observer

measurement or diagnosis remains the same on repeated independent observations of an

unchanged characteristic. This consistency can be assessed between different observers

(inter observer) or within a single observer (intra observer). Prognostic value is the

natural evolution of a disease without treatment and predictive value is the natural

evolution of the disease in response to treatment.150 For studying accuracy in grading

dysplasia of upper aero-digestive tract, there is no test available, which is thought to be

better than the pathologist’s observation when an accepted gold standard is not available

for assessing the validity obtained for grading these lesions. Therefore, reproducibility is

needed to provide indication of validity. The inter and intra observer agreement levels

give an estimate of the degree of bias and validity, where an appropriate gold standard is

not available.46

Hence, we tried to assess the inter and intra observer agreement levels to give an

estimate of the degree of bias and validity for assessing epithelial dysplasia. To make our

grading system reproducible two important aspects were identified (i) to obtain highest

Discussion

109

agreement for individual features of dysplasia / atypia and (ii) to assign scoring system

for individual features to reduce subjectivity and make assessment easy.

From the analysis of pilot study and from data obtained from the studies of Smith and

Pindborg, Kujan O, RJ Oliver, WM Tilakarane, the highest agreement scores were seen for

(1) atypical and abnormal mitosis (ii) Marked cellular and nuclear pleomorphism (iii)

irregular epithelial stratification (iv) loss of cohesiveness (v) Keratin pearls within the

epithelium (vi) loss of polarity of basal cells. The interpretation of the features required to be

present for a diagnosis varies from observer to observer as does their actual recognition and

the assessment of their degree of importance. There is a relatively high degree of

inconsistency in evaluation due to the overwhelming dependence upon subjective factors.

Therefore it was decided that an attempt should be made to place the diagnosis of epithelial

dysplasia on an objective level. A total of 18 features were identified to score for grading

dysplasia in the present study which was scored on a numerical scale of 1 to 10.

Rationale for assigning scores to the observations made in the present study:

Each individual characteristic dysplastic feature to be evaluated is listed along with

the scores as shown in Table No. 4. Many of the clinically diagnosed potentially malignant

disorders histopathologically show excess surface keratin, a prominent basal cell layer ,

hyperplasia of spinous cell layer (acanthosis) and a few normal mitotic figures in the

proliferative layers. These changes may be common to a number of mucosal lesions without

any cancer transformation potential. Warnakulasuriya S. et al54 suggested that the magnitude

of surface keratinization (keratosis or hyperkeratosis) is of no importance in the assessment

of dysplasia. As these features are of common occurrence and less significant for malignant

transformation, they were assigned lower scores of 1 and 2. Cells appear primitive and show

nuclear hyperplasia (enlarged nuclei), hyperchromatism of nucleus of basal cells, enlarged,

often eosinophilic nuclei (prominent nucleoli) and increased nuclear to cytoplasmic ratio.

These features are not exclusive to carcinogenesis as they are seen in reactive epithelium or

epithelium influenced by a variety of systemic alterations. Similarly not all dyskeratosis is

related to dysplasia, for example individual cell keratinization can be seen in Witkop disease

and in smokeless tobacco keratosis. Dyskeratosis is rare in carcinoma in situ, perhaps

because the cells are too primitive to produce keratin in quantities capable of being seen with

Discussion

110

light microscopy. It is also rare in moderate dysplasia and never seen in mild dysplasia.42

These cellular changes in dysplasia were assigned lower scores in the present assessment.

Elongated rete-processes lined by well polarized basal cells with normal stratification

and minimal or no cellular atypia were given lower score. These changes are seen in a variety

of inflammatory hyperplastic conditions like infections, frictional keratosis, psoriasis and

pseudoepitheliomatous hyperplasia, hence considered for lower scores in the present study.

On the other side of the spectrum some architectural and cytological features were

considered for higher score. The lower most region of the epithelial rete ridges with bulbous

enlargement results due to excessive basal cell proliferation. Especially if secondary

projections or nodules and forking are seen to arise from the basal layer that branch at

different angles into the subepithelial lamina propria or irregular ‘drop-shaped’ rete ridges

running in different directions. Budding of the rete-processes is a sinister feature. Small

clusters of basaloid cells may bulge or project from the basal layer into the stroma. These

features are considered for higher score. There is no physiological explanation for secondary

nodules and forking and hence were considered as ominous enough to alter the

histopathologic grading of a lesion to a higher level.

Dysplastic epithelium may also be atrophic which lacks retepegs and may be

ulcerated, thereby mimicking a traumatic or inflammatory lesion with thin, regenerating

epithelium creeping-in from the margins. Regenerating epithelium has granulation tissue

beneath it to distinguish it from dysplasia. Speight PM et al42 suggest that, in atrophic

dysplastic epithelium the basal cell hyperplasia appears to extend to the surface and they

recommend such lesions as severe dysplasia. This feature was considered for higher scores.

Loss of cellular cohesiveness not accompanied by severe inflammatory changes,

referred to as increased intercellular space or prominent intercellular bridges or loss of

intercellular adherence: is a feature of invasive carcinoma and is rarely seen otherwise. Intra

epithelial deposits of extra cellular matrix molecules can be regarded as intra epithelial

stroma. This result in the loss of intercellular adherence.47 Whereas Bouquet JE suggests that

loss of cohesiveness must be distinguished from intercellular edema, intra epithelial

inflammatory cells or degenerating cells with pyknotic nuclei and vacuolated cytoplasm,

which may also cause loss of cellular cohesiveness. These mild expansions of intercellular

spaces in few areas or along the length of epithelium are marked by the presence of

Discussion

111

inflammation in the underlying connective tissue.50 The loss of cohesiveness of cells in

pathologies of buccal mucosa could be due to the fact that there are fewer desmosomes in the

keratinocytes of the non-keratinized stratified squamous epithelium (SSE) as compared to

those of cornified SSE. The smaller number of desmosomes in non-cornified epithelia

reflects the lesser amount of keratin filament bundling (tonofilament formation) observed in

those cell types as noted by Garrant PR.151 Hence when otherwise loss of cohesiveness is

present it should be considered as a definite feature of dysplasia and thus assigned higher

score in our study.

Loss of polarity due to irregular orientation of basal cells has been scored high.

Polarity is lost when the normal alignment of the basal cells with their long axis

perpendicular to the basement membrane becomes disturbed.35 Due to hyperchromasia the

cells have basaloid appearance which indicates active proliferation. The oral mucosal

epithelium has germinal center in the parabasal area, from which cells with proliferative

potential are differentiated into prickle cells as well as basal cells. “The loss of polarity of

basal cells” (WHO classification) can be interpreted as loss of basal cell alignment. This loss

of alignment results from focal proliferation of basaloid (parabasal) cells. Therefore it is

important to pay careful attention to this feature to consider for accurately diagnosing

dysplasia.47 The loss of basal cell alignment is more clearly observed in dysplasia with a two

phase appearance. This is generated by a focal proliferation of basaloid cells in the lower part

of rete processes which replace not only basal cells but also lower prickle cells. Such

proliferating cell foci make a straight interface with parakeratinized cells in the upper layer,

which shows distinct contrast between the upper and lower layer of the epithelium. This

sudden transition is caused by loss of prickle cells due to their apoptosis. This feature is

referred to as the ‘two phase appearance’. There is a stepwise development of two phase

dysplasia, from small buddings of basaloid cells, their fusion, to fully developed ones. The

working committee for new histopathological criteria for borderline malignancy of oral

mucosa47 suggested that such lesions are diagnosed as true epithelial dysplasia.

Loss of epithelial stratification which is a more comprehensive loss of the normal

layered effect throughout the entire thickness and length of the epithelium is considered a

high risk feature and has been given higher score. Loss of differentiation of epithelial cells in

oral mucosa is the most fundamental morphological change of dysplasia. Bouquot JE et al

Discussion

112

(2006)50 describe this as an alarming morphological alteration of dysplastic epithelium which

is due to an apparent inability to properly differentiate and mature. Cells high in the

epithelium have the same immature appearance as those in the basal layers. Dysplastic

epithelium may be atrophic or acanthotic and some authors believe that atrophic forms have a

higher risk of malignant transformation.

Mitosis : The proliferative compartment contains two pools of dividing cells : Stem

cells and Transit amplifying cells. Stem cells are located in the basal cell layer and are slow

cycling cells; divide one to five times while migrating laterally and upward towards the

epithelial surface, producing a clone of differentiating cells.151 These different types of

proliferating cells cannot be distinguished by histologic methods. Regardless of whether the

cells are of the stem or amplifying type, cell division is a cyclic activity. The only phase

distinguished histologically is mitosis, which can be further subdivided into the recognizable

stages of prophase, metaphase, anaphase and telophase.

The mitotic activity can be affected by a number of factors such as time of the day,

stress and inflammation. Though slight inflammatory cell infiltrate stimulates mitosis, severe

inflammation causes a marked reduction in proliferative activity.152 In healing tissues,

inflammatory or reactive proliferations, metaphase of the dividing cell in the cell cycle is

most commonly seen and is considered as normal mitotic activity. Abnormal or atypical

mitosis appear as bizarre shaped mitotic cells and are easily identified in routine H&E stained

sections. In potentially malignant disorders and cancers different phases of mitosis are seen.

In the present study identification of abnormal and atypical mitotic cells, their presence in the

different layers of epithelium and their numbers were considered for highest score and hence

considered as high risk feature.

Cellular pleomorphism implies differences or bizarre forms in the cell size and cell

shape inappropriate for a given layer in the epithelium. Altered polarity sometimes with the

appearance of spindle shaped cells, and increased size unaccompanied by cell division are

components of pleomorphism. Bizarre mitoses and extreme of anisonucleosis are quite rare

in oral epithelium and hence should be considered as component of dysplasia.28

Abnormal mitosis, disturbed polarity, hyperchromatism, enlarged nucleoli,

hyperchromatism and intra epithelial keratinization were considered as features of

Discussion

113

significance for malignant transformation by Tilakaratne WM et al153 and the same is

considered for the present study.

Increased number and size of nucleoli : The nucleus of a fixed cell appears to have a

complex net like framework along with granules of varying size irregularly arranged. These

granules represent chromatin, a name given by Flemming because they stain brilliantly with

basic coal tar dye. The vast majority of nucleus has within them small, round, sharply

demarcated bodies called nucleoli. These are of two types, “true nucleoli” or plasmosomes

which take cytoplastic stains and “false nucleoli” or karyosomes which are reserve store

house of basic chromatin. They are known to contribute their substance to the chromosomes

at the time of cell division. The plasmosomes are related to metabolic activity in certain cells,

especially in the production of secretary products.154

Koss LG155 suggested that in normal cell, nucleoli, show cells with active RNA and

protein production. Size and configuration of nucleoli depends on two factors (i) Level of

production of RNA in nucleolus (ii) Cytoplasmic demand of the RNA. Therefore, large

nucleoli and increase of numbers reflect uninhibited formation of RNA in the presence of

limited cytoplasmic demand. These regions are called nucleolar organizer regions (NOR).156

While still within the nucleolus, these rRNA molecules are assembled together with

many protein molecules into two subunits of future ribosomes. The subunits leave the

nucleus through pores in the nuclear membrane and enter the cytoplasm. Here the two

ribosome subunits are brought together to form a functional ribosome at an early stage in

translation and separated again when the ribosome has fully translated a molecule of

messenger RNA (mRNA). Synthesis of ribosomes happens in the nucleolus. Ribosomes are

actually sites of protein synthesis.157

Recent studies showed that the nucleolus has emerged as a highly complex and

multifunctional regulatory compartment. The nucleolus is an extremely sensitive structure to

monitor ; it responds to cellular apoptotic stresses such as radiation, or exposure to cytotoxic

agents. Nucleolar proteins such as nucleophosmin, nucleostemin, have important roles in

these non-ribosomal functions, including regulation of cell growth and death, stress

responses, mitosis and the cell cycle.158,159

In the present study, increase in number and size of nucleoli, a feature also suggested

by WHO 2005 as a feature of dysplastic has been given higher score. Abnormalities in the

Discussion

114

nucleolar morphology of cancer cells attracted the attention of tumor pathologists as early as

the late 19th century, when Pianese G et al160 reported that hypertrophic and irregularly

shaped nucleoli were characteristic of malignant cells. The morphological and functional

changes of the nucleolus, widely observed in cancer tissues, are the consequence of the

increased metabolic necessities that characterize proliferating cells. 160

Study by Rajput DV et al161 suggested that NOR’s are the sites of rRNA synthesis

that reflect protein synthesis. AgNOR’s that stain for NOR associated proteins is known to

increase with increase in cell ploidy, increased transcriptional activity and in state of active

cell proliferation. Malati N. et al (2008) 162 from their study concluded that AgNOR quantity

is strictly proportional to the proliferative of AgNORs per nucleus suggests a marker of

proliferative activity. Quantitative assessment of size, shape and pattern of distribution acts

as a marker of premalignant / malignant change. Studies have shown that Nucleoli are

usually inconspicuous in normal cells. Nucleoli are prominent, more in number, enlarged and

irregular with occasional sharp pointed projections in malignant cells. This is due to

increased protein synthesis. It is reported that number and irregularity of nucleoli increased

with increasing grades of carcinoma.162

Keratisation of cells below the keratinized layer and Keratin pearls within reteridges :

Keratin formation is a feature of normal tissue maturation. As the basal cells of oral

epithelium move from one layer to the other and as the cells increase in age with progressive

maturation, they result in formation of keratin in the oral epithelium. This pattern of

maturation differs in different type of epithelium. Keratin is the protein formed by ribosomes,

secreted through the golgi apparatus and consist of aggregation of tonofilaments, involucrin

and filaggrin. As the epithelial cells contain these keratin filaments, they are referred to as

keratinocytes.

In dysplastic epithelium the irregular maturation and proliferation of cells lead to

abnormal formation of keratin and lower layers of epithelium. These are recognized as

individual cell keratinization and as keratin pearls. The presence of this is similar to the

keratin pearls formed in squamous cell carcinoma. Hence this feature in grading dysplasia is

scored high.

The changes seen in lamina propria should also be considered in grading dysplasia.

Lamina propria plays an important role in determining the phenotype of the overlying

Discussion

115

epithelium and directs the pattern of epithelial morphogenesis. Specificity is the result of

epithelial mesenchymal interaction. Though regional specificity exits, which may be consider

as being the result of extrinsic inductive stimuli from underlying lamina propria or an

intrinsic property of the basal layer of the epithelium.163 Inflammation in the lamina propria

was considered for grading dysplasia as it influences epithelial proliferation, keratinization

and cytological changes like cellular edema and altered nuclear cytoplasmic ratio in the

present study.

Appraisal of comparison of Epithelial Dysplasia classifications:

Grading is hampered by the arbitrary division into distinct categories of a continually

progressing process without any naturally and sharply defined borders. It is in fact, an

attempt to impose discrete categories on what is in effect a continuous grey scale and

therefore any grading scheme is by definition artificial.29 The ultimate diagnosis depends on

the emphasis put upon each of the characteristics in grading by the observer. Several studies

have shown inter and intra observer variability in the assessment of presence or absence of

individual characteristic features and thus the grading of dysplasia using different

classification system.

In the present study 100 cases of PMD with epithelial dysplasia were assessed using

Smith and Pindborg’s classification given in 1969, WHO in 2005 and the proposed new

classification for assessing the inter and intra observer variability. Kappa agreement was used

to evaluate the degree of variation in grading dysplasia for each cytological and architectural

feature. Both unweighted kappa and weighted kappa analysis were considered.

Unweighted or simple kappa (k) is a measurement of inter-rater agreement which

provides quantitative estimate of reliability and represents agreement greater than that

expected by chance alone. As stated by Cohen the kappa values fall between ‘-1’ : complete

disagreement, ‘0’ : agreement accepted by chance and ‘1’ : perfect agreement. Thus kappa

gives credit only for complete agreement.

Whereas weighted kappa (kw) considers the magnitude of disagreement by giving

credit to complete and partial agreement and hence a better measurement for reliability.

Weighted kappa results in a better inter and intra observer agreement values.164, 165 In the

present study the interpretation of kappa values suggested by Flesis and Cohen (1973)166

Discussion

116

were considered, where the values < 0.40 was rated as poor, 0.40-0.75 as fair to good, 0.76-

0.80 as excellent and above 0.81 as almost perfect.

The overall inter observer variability in our study for Smith and Pindborg and WHO

2005 classification showed moderate to fair agreement kw=0.48 & 0.149 and k=0.275 &

0.310 respectively (Table No. 2). Whereas the proposed classification had 90% agreement

(p=0.001) k=0.79. The intra observer variation for the 100 cases of dysplasia showed similar

observation when compared between the different classifications. Grading was reproducible

with good agreement for Smith and Pindborg and almost perfect agreement (91%, k=0.81)

for the proposed classification (Table No. 3).

Great variability in inter and intra observer agreement has been reported in several

studies. They ranged from poor to substantial agreement by use of different statistical

methods. Pindborg et al reported a wide range of agreement (1-78%) for 9 photomicrographs

evaluated by 72 pathologists. Abbey LM et al (1995)3l examined 120 oral biopsies of a range

of pathologies from hyperkeratosis to severe dysplasia. The agreement ranged from 35.8% to

55.8%. The kappa values improved to 0.70 – 0.88 for inter examiner comparisons when the

grading was expanded to “within one histological grade”.

Sudbo J et al (2001)53 s ho we d po or inter observer agreement using 3 point ordinal

scale (k values: 0.17 – 0.33). A slight improvement was observed when 2 categories of

“favorable” (Mild and moderate dysplasia) and “poor” (severe dysplasia) were used. (K

value ranged 0.21 – 0.34). Brothwell DJ et al (2003)46 showed a substantial agreement using

a 5 point ordinal scale with a group weighed Kappa (Kw) of 0.74 (94%, Cl=0.64 – 0.85).

However similar to other studies, in their study the group unweighted Kappa showed a fair to

moderate agreement (ks) of 0.37 with 95% Cl = 0.32 – 0.42.

Fischer D et al (2004)52 reported an overall kappa weighted value of 0.59 (95%

cl=0.45 – 0.72) An improvement in the kappa agreement was observed when the histological

diagnosis was simplified into three general categories of “no abnormality / hyperplasia”,

“mild / moderate / severe dysplasia” and “Carcinoma in situ” (Kw=0.70 (95% Cl : 0.56-

0.84). Subjective judgment and individual histopathological experience of the participating

pathologists was the core for the histopathological assessment.

Kujan O et al (2006)49 observed in their study that for a substantial group weighted

agreement was reported when four observers graded oral biopsy specimens into categories

Discussion

117

varying from hyperplasia (no dysplasia) to carcinoma in-situ (Kw=0.63, (95% CI: 0.42-

0.78)). A fair to moderate agreement was found using unweighted kappa test (ks=0.22,

(95CI:0.11 : 0.35). These variations were due to the fact that the unweighted kappa

considered all disagreements between raters. Whereas weighted kappa yields a higher

reliability when disagreements are small compared to, when they are large. An improvement

in the pairwise and overall agreement was reported in their study when a binary grading

system was used. They observed a better agreement between those experienced in examining

oral mucosal biopsies as compared with the general pathologist. Bias could also be due to

established departmental practice in oral pathology as stated by Kujan O et al.

All previously published reports on the inter observer agreement on grading OED

lesions showed considerable variations. Warnakulasuriya S et al26 suggested that this might

be due to the lack of objectivity in the evaluation of established criteria, arbitrary division of

the grading, lack of calibration of criteria and grading, and lack of sufficient knowledge

about which criteria are important for the prediction of malignant potential.

To overcome some of these limitations, we in our proposed classification have given

score to each of the 18 features identified for dysplasia and are predicted to have potential for

malignant transformation. The sum total of the scores ranged from 1 to 84. When the scores

ranged from 1 to 15 for a PMD it was categorized as ‘non-risky’ and when the scores ranged

between 16 to 84 the lesion was graded as ‘risky’. Accordingly 59% of the cases were

grouped under ‘risky’ (Table No. 5) suggesting that these are likely to develop into

malignancy. Whereas literature suggests, 2% to 17% chance of malignant transformation in

PMD. The scores assigned to individual dysplastic characters were in agreement with that

given by Smith and Pindborg for evaluating the dysplastic features more objectively. The

total score in our study was probably over estimated for the group categorized as ‘risky’. The

features considered under the group of ‘risky’ were further evaluated for risk categorization.

Entire score of the proposed classification comprising of 1 to 84 was evaluated for mean and

standard deviation (Table no. 6a, 6b). Median and inter quartile range was used to define cut

off point for scoring dysplasia and the cases were regrouped into low risk (score between 22 -

44) which consisted of 23.7% (14) cases, medium risk with scores between 45-66 were

66.1% (39) cases and in the high risk group we found 10.2% (6) cases with a scored between

67-84. This suggested that out of 59 cases initially grouped as ‘risky’, 6 cases could be

Discussion

118

considered for having potential to undergo malignant transformation as they were on the

higher side of the cut-off values.

To evaluate the 18 individual dysplastic features the ‘non-risk’ and ‘risky’ cases were

categorized for each feature of dysplasia (Table no. 7). Feature assigned number 1.8 i.e.

abnormal and superficial mitosis and 2.2 i.e. marked nuclear and cellular pleomorphism and

anisonucleosis were significantly associated with ‘risky’ group as compared to ‘non-risky’

where the Fisher’s Exact Test value were p<0.001, p<0.003 respectively. Fisher’s Exact test

of statistical significance is useful for categorical data that result from classifying objects into

different ways: it is used to examine the significance of association (contingency) between

two kinds of classification of samples.

To further evaluate the significance of individual features, those dysplastic features

that had score of 5 and above were segregated. A total of 10 such important features

predicted for malignant transformation were identified (Table No. 8). These are loss of

polarity, loss of cohesiveness with prominent intercellular bridges, abnormal and superficial

mitoses, increased number of mitotic figures, loss of epithelial stratification, keratin pearls

within rete ridges, mild to moderate degree of nuclear and cellular pleomorphism, marked

nuclear & cellular pleomorphism & anisonucleosis, hyperchromasia of basal cells and

atypical mitotic figures. The total score of these 10 features was calculated as 61.

All the 100 cases were once again categorized into the three risk categories as per the

scores of ten above mentioned selected features. Calculations using median and interquartile

range to define the cut-off point based on which of the ‘risky’ cases were regrouped as low,

medium and high risk (Table No. 9a & 9b). 13.6% (8) cases were identified with the total

score of more than 45, suggesting these cases which had score more than cut-off value could

be at a higher risk of undergoing malignant transformation.

Cases of both ‘non-risky’ (41 cases) and ‘risky’ (59 cases) were categorized into risk

categories as per the scores of 10 above mentioned selected individual dysplastic features.

Abnormal and superficial mitosis (feature no.1.8) and marked nuclear & cellular

pleomorphism & and anisonucleosis (feature no. 2.2) were significantly and more commonly

associated with ‘risky’ group as compared to the ‘non-risky’ group (Fischer’s Exact Test

p<0.001) (Table No. 10). These observations suggest that these two features observed in the

biopsy of any potentially malignant disorder would be important criteria for consideration for

Discussion

119

malignant transformation. Another important feature observed in the ‘risky’ category was

loss of cohesiveness which histologically presents as prominent intercellular bridges. Our

results are in agreement with the data obtained from the studies of Smith and Pindborg,

Kujan Omar, Oliver RJ, Tilakaranate WM.

Variation could result due to the difference in the understanding of these features in

terms of recognition and impact on clinical outcomes. Though some features are associated

significantly with the clinical outcomes, the other features which showed no statistical

significant association are as suggested by Patridge M. et al (1997)167 considered to be

important for the whole grading process.

One of the main objectives of the study by Kujan O et al (2006)49 was to assess the

prognostic value of the binary system using qualitative morphological characteristics. They

found the new binary grading system to be a very good predictor for the malignant changes

in oral epithelial dysplasia with reasonable values of sensitivity and specificity (85% and

80%, respectively). They predicted the clinical outcomes of dysplastic lesions with certainty

(28/33;84.9%), while the negative predictive value of system was 85%.

They also suggested that cases with either hyperplasia or mild dysplasia could be

considered as ‘watch and wait’, whereas the cases with severe dysplasia or CIS in situ need

intervention. The cases with moderate dysplasia belong to the category of ‘watch and wait’.

Their data showed that moderate dysplasia cases had significant potential for malignant

changes (14/30 ; 46.7%). 14 out of 16 (87.5%) cases, in their study with moderate dysplasia

had been assigned as high risk and progressed to OSCC. Their results further demonstrated

that patients whose OED lesions had been assigned as “high risk” tended to have

significantly greater transformation rates than those with lesions assigned as “low risk”

(p=0.004, by log-rank test). In contrast to other studies, the results of the Kujan O et al

(2006) suggest that the new binary grading system has the potential to help the clinicians to

make more appropriate treatment decisions when dealing with OED.

Appraisal of IHC in Epithelial Dysplasia:

Epithelial alterations are the first morphological changes seen during the progress of

malignant transformation of the oral epithelium. These potential malignant disorders are

important lesions for clinical preventive investigations. For better understanding not only are

the clinicopathological evaluation essential but also the molecular-biological changes during

Discussion

120

transformation process should be known. A variety of cell biological markers mainly

involved in cell proliferation and apoptosis are studied and some association is described in

the literature.

Identical genetic changes in dysplastic oral lesions and oral cancers in some patients

suggest that the progeny of the cells in dysplastic lesions may further modify into a

malignant lesion. A clone of oral keratinocytes may proceed through stages of transformation

from normal, to dysplastic, to malignant. The major event in oral malignancy is excessive

proliferation of oral keratinocytes, which break through the basal membrane and spread in

the stroma as suggested by Upasani OS et al (2004).168

Though progress in molecular oncology has significantly advanced our knowledge on

tumorogenesis; yet the practical applications of these genetic markers remain unresolved in

detecting oral dysplasia. Although many molecules have been studied as intermediate

markers, not many studies have focused correlating the findings with the clinical attributes of

terminal outcome. Combined with routine histopathology studies, the broad based studies of

molecular markers could have a great potential for the determination of prognosis of PMD.

The aim of the present study was to correlate the molecular changes in the epithelium

and basement membrane zone of lamina propria of PMD. An attempt is made in the present

study to correlate the histopathological features with IHC markers for CK-19, Ki-67 and

MMP-9 so as to substantiate the observation of dysplasia observed in H&E stained tissue

sections.

Cytokeratins (CK) are epithelial specific intermediate filament proteins. They are 20

different types and exhibit distinct patterns of expression in specific epithelial tissues. The

CK polypeptides have common epitopes against which antibodies are developed and these

tend to cross react. A large number of monoclonal and polyclonal antibodies have been

developed that are being used for diagnosis using IHC. CK are connected to the nuclear

envelope on one side, while they interact with plasma membrane proteins. These in turn

interact with ECM proteins. Thus CK are involved in transduction of signals and transport of

nutrients from inside to outside of cells and vice a versa. CK-19 is expressed in simple

epithelium and basal cells of non-keratinizing squamous epithelium reported by Takeda TK

et al (2006).169

Discussion

121

Basal cells situated along the base of the epithelial ridge have a flattened cell surface

in contact with connective tissue. These non-serrated cells contain only a few cytoplasmic

organelles and appear to be least differentiated cells. These represent the stem cells and

transit amplifying cells. They are characterized by a high nucleus to cytoplasmic ratio,

expression of K-19, a relative lack of keratin filament bundle and high levels of β1

integrins.151

In the present study for all 100 cases the intensity of CK-19 immunoexpression in the

epithelial cells and in the different cell layers were observed (Table no. 11a & 11b). The

immunoexpression of Ck-19 was found to be statistically significantly higher (p<0.001) in

the ‘risky’ cases as compared to ‘non-risky’. In the ‘risky’ group all the cases showed CK-19

expression whereas 20 ‘non risky’ cases had no CK-19 expression. In the ‘risky’ group, 44

cases out of 59 had more than 26% of cells with immunoexpression and the intensity of

expression ranged from light brown to dark brown suggesting increase in the number of

proliferating cells. 21 ‘non-risky’ cases had immune expression of CK-19 with maximum

having less than 26% of immunopositive cells with intensity of expression being light brown

suggesting the presence of proliferating cells in this group. The intensity of expression of

CK-19 was significantly higher in the ‘risky’ group as compared to the ‘non-risky’ (p <

0.001). We also found that 6 out of the 8 cases (75%) which were categorized as high risk

with scores more than 45 in the ‘risky’ group had high expression (+++, >50%) of CK-19.

These findings suggest disturbance in the distribution of proliferating cells and the stem cells

during proliferation in the epithelium.

Lindberg K.(1989)142 in their study suggested CK-19 as a marker of premalignancy

when the supra basal cells of dysplastic epithelium were immunopositive. This is in

accordance with our observation. Takeda TK et al (2006)169 reported CK-19 and p63 over

expression in their series of cases and suggested that CK-19 positive cells are transient

applying cells located in the parabasal layers. Loss of asymmetric cell division leads to

increase in number of stem cells in oral dysplasia. This disturbance of stem cells and

asymmetry in cell division can be recognized as dispolarity of basal cells, an essential

histological feature of dysplasia. Thus an increase in suprabasal cells could be an indicator of

higher grade of dysplasia. They also suggested that the decrease in CK-19 expression could

be due to alteration of stem cell function and that these cells could be replaced by

Discussion

122

proliferating cells in the basal layer of dysplastic epithelium. Hence this feature could be a

useful index in grading.

The expression of terminal differentiation products differ from keratinized to non-

keratinized SSE in which K4 and K13 keratins are expressed in the suprabasal cells, whereas

the suprabasal cells of cornified SSE express high molecular weight keratins (K1, K2, K10 &

K11). This normal pattern of keratin expression is altered in hyperplastic and cancerous

lesions of the oral mucosa. The appearance of K8, K18 and K-19, typical of simple highly

proliferative epithelial cells, becomes prominent in oral leukoplakia and stratified squamous

cell carcinoma was reported by Lan Su et al (1996).143

Ki-67 is a nuclear antigen present in the proliferating cells. It is present in all phases

of cell cycle, where its expression is low in G1 phase and increase to peak levels in the G2

and mitotic phase. In our study ki-67 expression and the area involved in the epithelial cell

layer showed a significantly marked increase of immunoexpression in the ‘risky’ group as

compared to the ‘non-risky’ (p<0.001). In the ‘risky’ group 48 cases had immunoexpression

in more than 26% of the epithelial cells and in 6 cases these cells were extending above the

suprabasal layer. Whereas in the ‘non-risky’ group 17 cases were immune positive for Ki-67,

maximum of which had less than 25% of immune positive cells and were limited to the basal

layer. The 8 cases with scores of more than 45 (high risk in the risky group) in the proposed

classification had more intensely stained and more than 50% immunopositive cells in the

upper layers of dysplastic epithelium (Table 12a & 12b). As Ki-67 is located within nucleoli

of proliferating cells only, it may be suggested that the antigen may regulate nucleolar

metabolism by increasing rates of ribosomal synthesis required by rapidly dividing cells

which can be histologically interpreted as basilar hyperplasia. Our findings are in accordance

to observations of Takeda TK et al169, they found increase in ki-67 expression in the basal

and parabasal cells of dysplasia group as compared to normal oral epithelium. They did not

notice any significant difference between the mild and moderate dysplasia cases. Takashi

Saku47 suggested that differentiation of keratinocytes happens in two directions : 1) towards

the surface – keratinization and 2) towards basal cells. Thus basal cells are neither the

germinal cells nor are they a source for squamous epithelial regeneration, but are terminally

differentiated cells. The cells with proliferating potential are located in the parabasal cell

layer which may be regarded germinal center of oral squamous epithelium.

Discussion

123

When epithelial cell nuclei are Ki-67 positive they are regarded as cells in

proliferation. In normal epithelium these are located in parabasal cells. In ‘risky’ dysplasia

group in our study, ki-67 positive cells were seen up to the middle third of epithelium. With

the presence of dyskeratosis or sudden keratinization in the lower layer of the epithelium the

ki-67 positive cells could indicate that the keratinocyte differentiation is not well regulated

that could lead to lack of basal cell alignment and loss of stratification.

Oral cancer may result from mutation of ras oncogene and a simultaneous

deactivating mutation of p53, a tumors suppresser gene in oral keratinocyte. Oral cancer may

result from mutation in other oncogenes and in EGFr / ras / kinase cascade / c-myc signaling

pathway. Authors correlated this with the stages of carcinogenesis from normal mucosa, to

squamous hyperplasia (9p), to dysplasia (3p, 17p), to carcinoma in situ (11q, 13q, 14q), to

invasive carcinoma (6p, 8q, 4q).2

In dysplasia the oral keratinocytes proliferation is limited within the epithelial

compartment. Eventually the proliferating keratinocytes may break through the epithelial

basement membrane and expand into the underlying connective tissue resulting in carcinoma.

Frank invasion into the stroma by cancer cells are obvious histologically. However in some

cases the invasiveness is not clearly determined. This so called micro invasion cannot be

ruled out in carcinoma in-situ cases. Takashi Saku47 suggests that invasion of cancer cells is

accompanied by a neoplastic stromal change around the cell nests. On H&E stained sections,

the newly formed neoplastic stroma is more diffusely eosinophilic than the lose connective

tissue of the lamina propria but more basophilically myxoid than surrounding fibrous

granulation tissue. Carcinoma in-situ does not induce such diffuse neoplastic stroma. To

evaluate the basement membrane and stromal changes MMP-9 was considered in the present

study.

MMPs are a family of structurally related but genetically distinct enzymes that

degrade extracellular matrix and basement membrane components and regulate cell–matrix

composition. MMPs may be secreted by tumour cells, fibroblasts, endothelial cells and

macrophages as well as mast cells and neutrophils.116

In our study MMP-9 immunoexpression was studied in the epithelium and connective

tissue in all 100 cases of PMD for grading dysplasia. The intensity and area of expression in

the epithelium was found to be statistically significant when compared between the ‘risky’

Discussion

124

and ‘non-risky’ cases (Fisher Exact Test, Intensity: p<0.001, Area: p<0.002). (Table No. 13a

& 13b). The intensity of MMP-9 expression in the epithelium was golden brown in majority

of the ‘risky’ and ‘non-risky’ cases. Where in more than 26% of epithelial cells had enhanced

MMP-9 immunoexpression in 37 cases of ‘risky’ dysplasia. The immunoexpression in

stromal cells did not show any significant correlation / association in the intensity and area

between the two study groups (Fisher Exact Test, Intensity: p<0.551 NS, Area: p<0.700 NS).

The expression of MMP-9 in ‘risky’ group was in accordance with the study on

OSCC where it was noted that MMP-9 is expressed in carcinoma and inflammatory cells

around carcinoma islands.118,119 For tumors to invade and metastasize, neoplastic cells must

be capable of degrading the extracellular matrix and accessing blood vessels and lymphatics.

Mounting evidence supports the view that extracellular proteinases, such as the MMPs,

mediate many of the changes in the microenvironment during tumor progression.124

Richard et al130 in his retrospective study examined changes in MMP 1, 2 and 9 using

polymerase chain reaction in oral dysplasia and oral squamous cell carcinoma and found

higher levels of MMP-1 and 9 mRNA that was significantly associated with dysplasia that

progressed to oral cancer as compared with those that did not. Peschos et al131 checked

expression of MMP-9 in benign, premalignant and malignant laryngeal lesions. They

suggested a two-step model of up-regulation of MMP-9 expression, first when a dysplastic

lesion evolves and the next when the dysplasia progresses to invasive carcinoma. MMP-9

expression was related neither to survival nor to the other available clinicopathological

parameters.

In the present study the stromal reaction of MMP-9 did not show any significant

difference between the cases. Though the stromal reaction in the high risk category cases

from the ‘risky’ dysplasia group were more pronounced as compared to the other cases.

Recent studies have shown that MMPs contribute to tumor progression, through its effect on

cell proliferation, survival and angiogenesis.170

In addition to carcinoma cells, cancers consist of tumor-associated stromal cells,

which include fibroblasts, endothelial cells, leukocytes, macrophages, nerve cells, and

adipocytes. The cancer cells crosstalk with stromal components during cancer progression

and they are mediated by transmembrane receptors, which are expressed on cancer cells and

stromal cells.132 Tumor-associated cells promote angiogenesis, inflammation, invasion, and

Discussion

125

ECM modeling through cell-cell contact and the production of growth factors, hormones,

cytokines, and proteinases such as MMPs.170

MMP 9 is detected in malignant transformation of various cells and is associated with

tumor metastasis and poor prognosis. In the aggressive human tongue squamous cell

carcinoma cell line, MMP-2 was only found in its latent form, whereas MMP-9 was found in

its active form. In an oral squamous cell carcinoma cell line SCC-25, increase expression of

MMP-9, in vitro, this was thought to occur via fibronectin integrin v 6 pathway. MMP-9

may not be the only factor for OSCC invasion, but it may be important for indirect cell

signaling by controlling the bioavailability and bioactivity of molecules that target specific

receptors, which regulate cell growth, migration, inflammation, and angiogenesis.132

Gao J (2010)133 studied the expression of MMP-2/MMP-9 in normal oral mucosa,

lymph node-negative tongue cancers, lymph node-positive tongue cancers and their

metastasized tumors in cervical lymph nodes. They hardly found MMP-2/MMP-9 expression

in normal epithelium. In lymph node-negative tongue cancer, 45% and 40% of these primary

tumors were positively stained for MMP-2/MMP-9. Importantly, in lymph node-positive

tongue cancer, 71% and 79% of these primary tumors were positive for MMP-2/MMP-9,

respectively. Over expression of MMP-2/MMP-9 was present in the metastatic lymph nodes.

Their result implied the significance of MMP-2 and / or MMP-9 in predictive value for the

actual or potential presence of cervical metastases. Such activity has prognostic value, and

provides impetus for further development of biotherapies targeted at specific inhibition of

MMP activities.



detecting oral dysplasia. None of the molecular markers single or in combination appear to

be ready for use in routine clinical diagnostic practices, although many molecules have been

studied as intermediate markers.16

Furthermore, as suggested by Gayani P. et al (2009)16 recent advances in the use of

quantitative methods has enabled the profiling of gene expression (microarray), protein

expression (proteomics), screening of epigenetic changes, e.g. by pyro-sequencing and single

nucleotide changes (SNP) of hundreds of pre-selected genes simultaneously. These

techniques which require minute amount of tissue when carried out in controlled way and

Discussion

126

using carefully selected samples with proper validation will help the potentially malignant

and malignant tissue samples. Combined with routine histopathology studies, these broad

based studies of gene and their products appear to have a great deal of potential for the

identification of diagnostic and prognostic markers of pre-cancer, and data from multicenter

studies using microarray technology may be engaged in the future for a better understating of

aetiopathogenesis of pre-cancer.

CONCLUSION

Conclusions

127

CONCLUSIONS

Oral epithelial dysplasia : a histopathologic marker of premalignancy, is a diagnostic

term used to describe the histopathological changes seen in progressive chronic and

premalignant disorders of the oral mucosa.. It may also be seen in verrucous or papillary

leukoplakia or in the margins of chronic mucosal ulcers. It is also consistently seen in the

apparently normal mucosa adjacent to the squamous cell carcinoma. Clinically it may present

as leukoplakia, erythroplakia or leukoerythroplakia 25

A histological dysplasia classification system should ideally meet two basic

requirements. First it should be easily applicable in daily routine practice with low inter and

intra observer variability. Secondly, it should allow a clear separation and give clear

suggestions to the clinician to distinguish between patients who need treatment to prevent

progression and those whom no treatment is needed and should be kept on routine

observation.

We propose a new binary classification system consisting of two categories where the

individual architectural and cytological features are given a score. This new classification has

been validated and has shown good to almost perfect agreement for both inter and intra

observer variation. This classification has been designed with the following objectives : a) to

reduce the inter and intra observer variability, b) to make it objectively reproducible,

c) transmit maximum information possible from interpretation of routine H&E stained

sections, d) give a clear diagnosis for clinicians to make critical treatment decision.

From the observations of the present study, ten important morphological and

architectural features are identified to be considered for categorizing potentially malignant

disorder histologically as ‘risky’. Abnormal superficial mitosis and marked cellular and

nuclear pleomorphism were two important features that could predict malignant

transformation are found to be statistically significant and that should be identified and

considered for risk categorization and malignant transformation.

The IHC markers using Ki-67 for observing the irregular epithelial cell proliferation,

Ck-19 for loss of polarity and basal cell hyperplasia in the suprabasal layers and MMP-9 for

stromal changes, micro-invasion and changes in the basement membrane and lamina propria

have substantiated the H&E observation.

Conclusions

128

Our proposed system of grading dysplasia is designed for histopathological

evaluation using H&E stained sections. It was designed to extract a larger set of data with

detailed observations of individual features of dysplasia based on architectural and

cytological changes. Our aim was to generate objective data to correlate with defined patterns

of dysplastic features.

Grading dysplasia is not an exact science and pathologists are doing their best to

reach an optimal result. In this study we have tried to weigh individual features by defining

their importance and scoring them for OED grading process obtaining a common consensus

on the features. However, data reveals that the proposed classification system can be used to

improve or develop simpler and easy diagnostic methods.

The outcome of ‘risky’ dysplasia for malignant transformation could not be assessed

in the study due to lack of follow up data. Within the limitations of the analysis we have

shown that many individual features of dysplasia can be assessed reproducibly as they were

categorized under high risk with a definite score. Reducing features to definable components

increases objectivity and insight, though it makes it difficult for comparison with the

published data.

Future studies should compare lesional responses related to patient population and

their geographic areas of origin. As well as specific carcinogenic factors, diagnostic

usefulness of aids like Toluidine blue application, oral brush biopsy techniques and

spectroscopy and the natural history of the potentially malignant disorders should be

correlated.

Relevance of the observations and future scope of this study

• The proposed classification is objective and is easy to understand. As the inter and

intra observer agreement is statistically significant and hence is reproducible.

• The proposed classification has been assigned with scores for specific features that

has helped to minimizes the inter and intra observer variability.

• Reporting and grading of oral epithelial dysplasia in potentially malignant disorders

has been made easy and valid with the use of routine H&E stain section from any

histopathology diagnostic service, center or laboratory.

Conclusions

129

• Larger sample size studies, at various oral pathology centers, in different

geographical locations could help in developing this proposed classification

universally acceptable.

• Our classification of ‘risky’ and ‘non-risky’ will be helpful for clinical application in

imparting proper treatment, patient education, ‘wait & watch’ policy and follow-up.

Hence, the need is for:

• Application of the proposed binary classification as ‘risky’ and ‘non-risky’ for

routine use.

• Multicentric focused group assessment using the proposed binary classification.

• Longitudinal follow-up study for evaluating the malignant potential of ‘risky’ and

‘non-risky’ cases of potentially malignant disorders.

Bibliography

130

BIBLIOGRAPHY

1. Corten RS, Kumar V, Robbs SL. Robins Basic Pathology, Neoplasia, 7th edition

Elsevier Publisher 2003; 165-210.

2. Sugerman PB, Savage NW. Current concepts in oral Cancer. Australian Dental Journal

1999; 44(3): 147-156.

3. Rani RM, Bonu S, Jha P. Tobacco use in India: prevalence and predictors of smoking

and chewing in a national cross sectional household survey. An international peer

reviewed journal for health professionals. J of Control 2003 ; 12.

4. Mehrotra R, Pandya S, Chaudhary A, Kumar M, Singh M. Prevalence of oral pre

malignant and malignant lesions at a tertiary level hospital in Allahabad, India. Asian

Pacific Journal of Cancer Prevalence 2008(9) ;

5. Byakodi R, Byakodi S, Hiremath S, Byakodi J, Adaki S, Marathe K, Mahind P. Oral

cancer in India : an epidemiology and clinical review. J. Community Health 2012;

37(2): 316-9.

6. Sankaranarayana R. Oral Cancer in India: An epidermal. Oral Pathol, Oral Med, Oral

Surg, 1990; 69(3): 325 – 530.

7. Bouquet JE, Speight PM, Farthing PM. Epithelial dysplasia of oral mucosa diagnostic

problems and prognostic features. Current Diagnostic Pathology 2006; 12: 11-21.

8. Stedman's Medical Dictionary. Copyright © 2006 Lippincott Williams & Wilkins.

9. Napier SS, Speight PM. National history of potentially malignant disorders of the oral

lesions and conditions: an overview of the literature. J. Oral Pathol. Med 2008; 37; 1-10.

10. Gupta PC, Mehta FS, Daftary DK. Incidence rates of oral cancer and natural history of

oral precancerous lesions in a 10 year follow up study of Indian villagers. Community

Dent Oral Epidemiol 1980; 8: 287-333.

11. Pindborg JJ, Reibel J, Roed PB, Mehta FS. Tobacco induced changes in oral

leukoplakic epithelium. Cancer 1980; 45: 1330-6.

12. Warnakulasuriya S, Johnson NW, van derWaal I. Nomenclature and classification of

potentially malignant disorders of the oral mucosa. J Oral Pathol Med 2007; 36: 575-80.

13. Pindborg JJ, Reibel J. Subjectivity in evaluating oral epithelial dysplasia, carcinoma in

situ and initial carcinoma. J Oral Pathology 1985; 14: 695-708.

Bibliography

131

14. Reibel J. Prognosis of Oral Premalignant lesions, Significance of clinical,

histopathological and molecular biological characteristics. Crit Rev Oral Biol Med

2003; 14(1): 47-62.

15. Holmstrup P, Vedtofe P, Reibel J. Long term treatment outcome of oral premalignant

lesions. Oral Oncology 2006; 42: 461-74.

16. Gayani P, Tilakaratne WM, Tavassoli M, Warnakulasuriya S. Molecular markers in

oral epithelial dysplasia: review. J Oral Pathol Med 2009; 38: 737-752.

17. Bouquot JE. The Pathology and progression of Oral Premalignancy. Proceedings,

Epithelial dysplasia symposium, 5th International Congress on Oral Cancer, Royal

College of Physician, London, UK September 1997. Invited review.

18. Coelho KR. Challenges of the oral cancer burden in India. J. Cancer Epidemiology

2012.

19. Saddichha S, Khass CR. Prevalence of tobacco use among young adult males in India: A

community based epidemiological study. Am J Drug Alcohol Abual 2010; 36(1) : 73-7.

20. Makwana NR, Shah V, Gadav S. A study on prevalence of smoking and tobacco

chewing among adolescents in rural areas of Jamnagar District, Gujarat State. Journal of

Medical Science Research 2007; (1).

21. Hazarey VK, Erlewad DM, Mundhe KA, Ughade SN. Oral submucous fibrosis: Study of

1000 cases from Central India. Journal of Oral Pathology and Medicine 2007; 36 (1):

12–17.

22. Ivan D, James L. Anderson’s Pathology, Vol 1 .10th edition, 1996pp 61, Mosby – Year

Book, London.

23. Kramer IRH, Lucas RB, Pindborg JJ, Sobbin LH. Definition of leukoplakia and related

lesions; An aid to studies on oral precancerous lesions. WHO Collaborating Centre for

Oral Precancerous lesions. Oral Surg. Oral Med Oral Pathol 1978; 46(4); 518-539.

24. Sheffield Edward. Pathology in Dentistry. 1st Edition Oxford University Press 1966.

281.

25. Lumerman H, Freedman P, Kerpel S. Oral Epithelial dysplasia and the development of

invasive squamous cell carcinoma. Oral Surg Oral Med Oral Pathol Oral Radiol Endod

1995; 79: 321-9.

Bibliography

132

26. Warnakulasuriya S. Histologic grading of oral epithelial dysplasia: revisited. J Pathol.

2001; 194: 294-297.

27. Soames JV, Southam JC. Oral Pathology. Oxford Medical Publication, New York 1998:

138-166.

28. Odell EW, Morgan PR. Oral Squamous carcinoma and premalignancy in Biopsy

Pathology of the Oral Tissues. First Edition, Chapman & Hall, London, Chapter 8,

1998; 181-241.

29. Bosman FT. Dysplasia classification: Pathology in disgrace? J Pathol. 2001; 194: 143-

144.

30. Johnson NW. Risk markers for oral diseases. Vol 2. ‘Oral Cancer’ – Detection of

patients and lesions at risk. Cambridge University Press 1991.

31. Abbey LM, Gunsolle JC, Page DG. Intra examiner and inter examiner reliability in

diagnosis of oral epithelial dysplasia. Oral Surg. Oral Med. Oral Pathol. 1995; 80: 188-

191.

32. Karabulut A, Reibel J, Thekildoen MH. Observer variability in the histologic assessment

of oral premalignant lesions. J Oral Pathol. Med. 1995; 24: 198-200.

33. WHO collaborating Centre for Oral Precancerous Lesions. Definition of leukoplakia

and related lesions: An aid to studies on oral precancer. Oral Surg. 1978; 46(10).

34. Subudo J, Byrne M, Johanesen A, Reith A. Comparison of histologic grading and large

scale genomic status (DNA) as prognostic tools in oral dysplasia. Pathol 2001; 194:303-

310.

35. Smith C, Pindborg JJ. Histological grading of oral epithelial atypia by the use of

photographic standards. Copenhagen: Hambergers Bogtrykkeri WHO reference centre

for Precancerous Conditions 1969.

36. Fishman SL, Ulmansky M, Sela J, Bab I, Gazit D. Correlative clinicopathological

evaluation of oral premalignancy. J Oral Pathology 1982; 11: 283-289.

37. Katz HC, Shear M, Altini M. A critical evaluation of epithelial dysplasia in oral

mucosal lesions using the Smith and Pindborg method of standardization. J Oral

Pathology 1985; 14; 476-482.

Bibliography

133

38. Bancozy J, Csiba A. Occurrences of epithelial dysplasia in oral leukoplakia-Analysis

and follow up study of 120 cases. Oral Surg Oral Med Oral Pathol 1976; 42(6): 766-

774.

39. Burkhardt A, Maerker R. A colour atlas of oral cancer. The diagnosis and classification

of leukoplakia, precancer condition and carcinomas Mosby 1st Edition, London. 1981:

34-35.

40. Girod SC, Pfeiffer P, Ries J, Pape HD. Proliferative activity and loss of function of

tumor suppressor genes as biomarkers in diagnosis and prognosis of benign and

preneoplastic oral lesions and squamous cell carcinoma. British Journal of Oral and

Maxillofacial Surgery 1998; 36:252-260.

41. Neville BW, Damm DD, Allen CM, Bouquot JE. Oral and maxillofacial pathology. W

B Saunders Company, London, 2nd edition, Chapter No. 10, 2002; 342-345.

42. Speight PM, Farthing PM, Bouquot JE. The pathology of oral cancer and precancer.

Curr. Diag. Path 1996; 3: 165-7.

43. Kuffer J, Lambardi S. Reconsideration oral risk lesions. Oral Diseases 2002; 38:302-

307.

44. Moles MAG. Commenon Kuffer & Lambardi classification. Oral Oncology 2002; 38:

561.

45. Gnepp DR. Diagnostic surgical pathology of the head and neck. Saunders Company,

Philadelphia. 1st Edition, chapter 2; 2001: 12.13.

46. Brothwell DJ, Lewis DW, Bradley G, Leong I, Jordan RCK. Observer agreement in the

grading of oral epithelial dysplasia. Community Dent Oral Epidemiol.2003; 31:300-

305.

47. Takashi S. The working Committee for New Histopathological criteria for Borderline

Malignancy of Oral Mucosa, The Japanese Society for Oral Pathology (2004).

Proceedings of symposium of JSOP Tokyo.

48. Barnes L, Eveson J, Reichart P, Sidransky D. World Health Organization’s.

Classification of tumors. Pathology and genetics of tumors of the head and neck. Lyon:

IARC Press; 2005.

Bibliography

134

49. Kujan O, Khattab A, Richard O, Stephen R, Thakker N, Sloan P. Evaluation of a new

binary system of grading oral epithelial dysplasia for prediction of malignant

transformation. Oral Oncology 2006; 42:987-993.

50. Bouquot JE, Speight PM, Farthing PM. Epithelial dysplasia of oral mucosa diagnostic


51. Maria A, Carmo VD, Patrícia CC. Binary System of Grading Epithelial Dysplasia in

Oral Leukoplakias. Carcinogenesis. http// dx.doi.org/10.5772/54466

52. Fischer D, Joel E, Morton TH, Schwartz S. Inter observer reliability in the

histopathologic diagnosis of oral pre malignant and malignant lesions. J Oral Pathol

Med 2004; 33:65-70.

53. Sudbo J, Kildal W, Risberg B, Koppang HS. DNA content as a prognostic marker on

patients with oral leukoplakia. New Engl J Med 2001; 344:1270-8.

54. Warnakulasuriya S, Reibel J, Bouquot J, Dabelsteen E. Oral epithelial dysplasia

classification systems : predictive value, utility, weakness and scope for improvement.

J Oral Pathol Med 2008; 37: 127-133.

55. Burkhardt A. Advanced methods in the evaluation of premalignant lesions and

carcinomas of the oral mucosa. J Oral Pathology1985; 14:751-778.

56. Brennan M, Cesar M, Lockhart P, Wary D, I Al-Hashimi, Tony A et al. Management of

oral epithelial dysplasia : a review. Oral Surg Oral Med Oral Pathol Oral Radiol Endod

2007; 103 (suppl 1):s19.el-s19.el2.

57. Sigurdson A, Willen R. Collection of cytological specimens with a cell aspirator with

reference to the wooden spatula method: Comparative, quantitative and qualitative

study. Int J Oral Surg 1979; 8: 43.

58. Heddle JA. A rapid in-vivo test for chromosomal damage. Mutation Res 1973; 18:187.

59. Kramer IRH, Lucas RB, El-Labban, Lister. A computer aided study on the tissue

changes in oral keratosis and Lichen Planus and an analysis of case groupings by

subjective and objective criteria. British Journal of Cancer 1970; 24(3) : 407-423.

60. Franklin CD, Gohari K, Smith CJ, White FH. Quantitative evaluation of normal

hyperplastic and premalignant epithelium by stereological methods. Oral malignancy

243 University of IOWA Press, 1980.

Bibliography

135

61. Cahn LR, Eisenbud L, Blake MN. Histochemical analysis of white lesions of the mouth:

Analysis of glycogen content. Oral Surg 1962; 15:458.

62. Johnson NW, Evans AW, Morgan PR, Butcher RG. Biochemical and histochemical

changes in oral premalignancy. University of Iowa Press 1980:312.

63. Heyden C. Histochemical investigation of malignant cells. Histochemistry 1974;

39:327.

64. Scully C, Burkhardt A. Tissue markers of potentially malignant human oral epithelial

lesions. J Oral Pathol Med 1993; 22:246-56.

65. Ray JG, Chattopadhyay A, Caplan J. Usefulness of AgNOR counts in diagnosing

epithelial dysplasia. J Oral Pathol Med 2003; 32:71-6.

66. Scully C. Progress in determining the malignant potential of oral lesions. J Oral Pathol

Med 2003; 32:251-6.

67. Bancozy J, Lapis K, Albrecht M. Scanning electron microscopic study of oral

leukoplakia. J Oral Pathol 1980; 9: 145.

68. Hume WJ, Potten CS. Advances in epithelial kinetics: an oral view. J Oral Pathol 1979;

8: 3.

69. Bier J, Nicklisch U, Platz H. The doubtful relevance of non-specific immunoreactivity

in patients with squamous cell carcinoma of the head and neck region. Cancer 1983;

52:1165–1171.

70. Kovesi G, Fekete B. Potential diagnostic value of a new modality of leukocyte

adherence inhibition assay in oral cancer. J Maxillofacial Surg 1982; 10:28.

71. Sako T, Sato E. Prediction of malignant change in oral precancerous lesions by DNA

cytofluorometry. J of Oral Pathol 1983; 12:90.

72. Scully C. Serum beta 2 microglobulin in oral malignancy and premalignancy. J of Oral

Pathol 1981b; 10:354.

73. Eveson JW. Animal models of intra-oral chemical carcinogenesis: a review. J Oral

Pathol 1981; 10:129.

74. Lee JS, Kim SY, Hong WK. Detection of chromosome polysomy in oral leukoplakia, a

premalignant lesion. J Natl Cancer Inst 1993; 85:1951-4.

Bibliography

136

75. Shin DM, Charuruks N, Lippman SM, Lee JJ, Ro JY, Hong WK, Hittelman WN.

p53 protein accumulation and genomic instability in head and neck multistep

tumurogenesis. Cancer Epidemiol Biomarkers Prev. 2001; 10(6):603-9.

76. Sidransky D. Nucleic acid-based methods for the detection of cancer. Science 1997;

278: 1054-9.

77. Spafford K, Recee AB. Detection of head and neck squamous cell carcinoma among

exfoliated oral mucosal cells by microsatellite analysis. Clin Cancer Res.2001;7;607-12.

78. Boudewijin JM, Braakhuis BJ, Tabor MP, Kummar JA, Leemans R, Barkenhoff RH. et

al. A Genetic Explanation of Slaughter’s Concept of Field Cancerization : Evidence and

Clinical Implications. Cancer research 2003 (April 15) ;63:1727-1730.

79. Gabriel DD, Jakupciak JP, Machin M, Parr R. Clinical 172 implications and utility of

field cancerization. Cancer Cell International 2007;7:2.

80. Monique G, Oijen V, Slootweg PJ. Oral Field Cancerization: Carcinogen-induced

Independent Events or Micrometastic deposits. Cancer Epidemiology, Biomarkers &

Prevention. 2000; 9:249-256.

81. Jang SJ, Chiba I, Hirai A, Hong WK, Mao L. Multiple oral squamous epithelial lesions:

are they genetically related? Oncogene 2001; 20: 2235-2242.

82. Braakhuis BJM, Tabor MP, Leemans CR. Second primary tumors and field

cancerization in oral and or pharyngeal cancer: molecular techniques provide new

insights and definitions. Head Neck 2002; 24:198-206.

83. Simon R, Eltze E, Schafer KL, Burger H, Semjonow A, Hertle L, Dockhorn D, Terpe

HJ, Bocker. Cytogenic analysis of multifocal bladder cancer supports a monoclonal

origin and intraepithelial spread of tumor cells. Cancer Res 2001; 61:355.

84. Boudewijin JM, Brakhuis C, Leemana R, Brakenhoff RH. A genetic progression model

of oral cancer: current evidence and clinical implications. J Oral Pathol Med 2004;

33:317-22.

85. Wright A, Shear M. Epithelial dysplasia immediately adjacent to oral squamous cell

carcinomas. Journal of Oral Pathology1985; 14:559-564.

86. Ogden GR, Cowpe JG, Green MW. Evidence of field change in oral cancer. British

Journal of oral and maxillofacial surgery.1990; 28: 390-392.

Bibliography

137

87. Kannan S, Kartha P, Balaram G, Jagadeesh C, Pillai R, Pillai R, et al. Ultrastructural

analysis of the adjacent epithelium of oral squamous cell carcinoma. British Journal of

Oral Maxillofacial Surgery 1996; 34:51-157.

88. Thomson PJ. Field change and oral cancer: new evidence for widespread

carcinogenesis. Int J Oral Maxillofac Surg 2002; 31:262-266.

89. Copper MP, Braakhuis BJM, de Vries N, Dongen V, Nauta JJP, Snow GB. A panel of

biomarkers of carcinogenesis of the upper aerodigestive tract as a potential immediate

endpoints in chemoprevention trials. Cancer 1993; 71:825-830.

90. Ogden GR, Lane EB, Hopwood DV, Chisholm DM. Evidence of field change in oral

cancer based on cytokeratin expression. British Journal of Oral and Maxillofacial

Surgery 1993; 67:1324-1330.

91. Bongers V, Snow GB, de Vries N, Braakhuis BJM. Potential early markers of

carcinogenesis in the mucosa of the head and neck using exfoliative cytology. J Pathol

1996; 178:284-289.

92. Bosch FX, Ouhayoun JP, Bader BL, Collin C, Grund C, Lee I, Franke WW. Extensive

changes in cytokeratin expression pattern in pathologically affected human gingiva.

Virchows Arch B Cell Pathol 1989;58:59-77.

93. Bartkova J, Lukas J, Muller H, Strauss M, Gusterson B, Bartek J. Abnormal patterns of

D-type cyclin expression and G1 regulation in human head and neck cancer. Cancer Res

1995;55: 949-956.

94. King LE. What does the epidermal growth factor do and how does it do it? J Invest

Dermatol 1985;84:165-169.

95. Oijen V, Rijksen G, Ten Broek FW, Slootweg PJ. Increased number of proliferating

cells in oral epithelium from smokers and ex-smokers. Oral Oncol 1998 ; 16 : 3158-

3168.

96. Birchall MA, Schock E, Harmon BV, Gobe G. Apoptosis, mitoses, PCNA and bcl-2 in

normal, leukoplakic and malignant epithelia of the human oral cavity; A prospective

Study. Oral Oncol 1997 ; 33 : 419-425.

97. Bongers V, Snow GB, de Vries N, Cattan AR, Hall AG, van der Waal I, Braakhuis BJ.

Second primary head and neck squamous cell carcinoma predicated by the s-transferase

Bibliography

138

expression in healthy tissue in the direct vicinity of the first tumor. Lab Invest 1995

;73:503-510.

98. Oijen V, de Craats V, Slootweg JG. p53 overexpression in oral mucosa in relation to

smoking. J. Pathol 1999 ;187: 469–74.

99. Patrick KH, Califano AJ. The Molecular Biology of Mucosal Field Cancerization of the

Head and Neck. Crit Rev Oral Biol Med. 2003; 14(5) : 363-369.

100. Worsham MJ, Wolman SR, Carey TE, Zarbo RJ, Benninger MS, Van Dyke DL.

Common clonal origin of synchronous primary head and 179 neck squamous cell

carcinomas: analysis by tumor karyotypes and fluorescence in situ hybridization.

Human Pathol 1995 ;26: 251–261.

101. Partridge M, Stelios P, Elaine P, Emilion GG, Roger P, John DL. A case control study

confirms that microsatellite assay can identify patients at risk of developing oral

squamous cell carcinoma within a field of cancerization. Cancer Research 2000 ; 60 :

3893-3898.

102. Tabor BRH, Houten V. Persistence of genetically altered fields in head and neck cancer

patients; biological and clinical implications. Clin Cancer Res 2001;7:1523-32.

103. Ha PK, Tong BC, Westra WH, Sanchez CM, Parrella P, Zahurak M, Sidransky D,

Califano JA. Mitochondrial C-tract alteration in premalignant lesions of the head and

neck : a marker for progression and clonal proliferation. Clin Cancer Res 2002 ; 8 :

2260-2265.

104. Ai H, Barrera J, Pan Z, Meyers AD, Varella GM. Identification of individuals at high

risk for head and neck carcinogenesis using chromosome aneuploidy detected by

fluorescence in situ hybridization. Mutat. Res. 1999 ; 439 : 223–232.

105. Schwatz JL, Panda S, Beam C, Bach LE, Adami GR. RNA from brush oral cytology to

measure squamous cell carcinoma gene expression. J Oral Pathol Med. 2008 37(2) :

70-77.

106. Tabor MP, Brakenhoff RH, Viola MM, Houten V, Kummer JA, Mireille HJ, Peter JF et

al. Persistence of genetically altered fields in head and neck cancer patients : Biological

and clinical implications. Clin. Cancer Res. 2001 ; 7 : 1523-1532.

107. Pindborg JJ, Daftary DK, Mehta FS. A follow up study of sixty one oral dysplastic

precancerous lesions in Indian villages. Oral Surg 1977 ; 43(3) : 383-390.

Bibliography

139

108. Silverman S, Gorsky M, Lozada F. Oral Leukoplakia and malignant transformation.

Cancer 1984 ; 53 : 563-568.

109. Oral Precancer : Dysplasia, Molecular Biology, Microbiology. The maxillofacial Center

for Diagnostics and Research. http: Precancer dysplasia and markers.html

110. Summerlin DJ. Precancerous and cancerous lesions of the oral Cavity. Dermatol Clin.

1996 ; 14(2) : 205-23.

111. Wikipedia, the free encyclopedia - Antigen identified by monoclonal antibody Ki-67

112. Humayun S, Prasad VR. Expression of p53 protein and ki-67 antigen in oral

premalignant lesions and oral squamous cell carcinomas: An immunohistochemical

study. National Journal of Maxillofacial Surgery 2011 ; 2(1) : 38-46.

113. Takafumi U. Prognostic Significance of Ki-67 Reactivity in Soft Tissue Sarcomas.

Cancer1989; 63:1607-1611.

114. Kill IR. Localisation of the Ki-67 antigen within the nucleolus evidence for a fibrillarin

deficient region of the dense fibrillar component. Journal of Cell Science 1996 ; 109 :

1253-1263.

115. Massova I, Kotra LP, Fridman R and Mobashery S. Matrix metalloproteinases:

structures, evolution and diversification. FASEB J. 1998 ;12 : 1075–1095.

116. EA Baker, Leaper DJ, Hayter JP, Dickenson AJ. The matrix metalloproteinase system

in oral squamous cell carcinoma. British Journal of Oral and Maxillofacial Surgery

2006 ; 44 : 482-486.

117. Iyer RP, Patterson NL, Fields GB, Lindsey ML. The history of matrix

metalloproteinases: milestones, myths and misperceptions. Am J Physiol Heart Circ

Physiol 2012 ; 303: H919–H930.

118. Verma RP, Corwin H. Matrix metalloproteinases (MMPs): Chemical–biological

functions and (Q)SARs. Bioorganic & Medicinal Chemistry 2007;15 :2223–2268.

119. Vihinen P, Kahari VM. Matrix metalloproteinases in cancer: prognostic markers and

therapeutic targets. Int. J. Cancer 2002 ; 99 : 157–166.

120. Loffek S, Schilling O, Franzke CW. Biological role of matrix metalloproteinases: a

critical balance. Eur Respir J. 2011; 38: 191–208.

121. Yan C, Boyd DD. Regulation of matrix Metalloproteinase gene Expression. J. Cell

Physiol 2007 ;211: 19–26.

Bibliography

140

122. Sorsa T, Tjaderhane L, Salo T. Matrix metalloproteinases (MMPs) in oral diseases. Oral

Diseases 2004 ;10 : 311-218.

123. Alessandro F, Marco S, Masini E, Milena P, Gallo O, Sardi I. Expression of matrix

metalloproteinase-1, matrix metalloproteinase-2, and matrix metalloproteinase-9 in

carcinoma of the head and neck correlation with p53 status, inducible nitric oxide

synthase activity, and angiogenesis. J of Cancer Research 2002 ; 95 : 1902-1910.

124. Lisa J. Mc Cawley, Matrisian LM. Tumor progression: Defining the soil round the

tumor seed. Current Biology 2001; 11(1):25–27.

125. Roopali R, Jiang J, Marsha. MMP as Novel Biomarkers and Potential Therapeutic

Targets in Human Cancer. Journal of Clinical Oncology 2009;27(30) : 5287-97.

126. Kai K, Plaks V, Werb Z. Matrix Metalloproteinases: Regulators of the Tumor

Microenvironment. Cell 2010 ; April 2 : 52-67.

127. Wuertz B, Ondrey F. MMP-9 Analysis in Carcinogen-Treated Oral Keratinocytes

Otolaryngology-Head and Neck Surgery, 2010 August ; Vol 143, No 2S2,

(http://oto.sagepub.com/content/143/2_suppl/P190.3T).

128. Thiennu HV. MMP-9/Gelatinase B is a key regulator of growth plate angiogenesis and

apoptosis of hypertrophic chondrocytes. Cell 1998 May 1; 93(3): 411–422.

129. Philippe V, Claude L. Chemokine and cytokine processing by matrix metalloproteinases

and its effect on leukocyte migration and inflammation. J of Leukocyte Biology 2007 ;

82 : 1375-81.

130. Richard C, Maricris M, Caroline H, Shiboski. Overexpression of MMP-1 and 9 mRNA

is associated with the progression of oral dysplasia to cancer. Clin Cancer Res. 2004 ;

10 : 6460-65.

131. Peschos D. Expression of matrix metalloproteinase-9 (gelatinase B) in benign,

premalignant and malignant laryngeal lesions. Histol Histopathol 2006; 21: 603-608.

132. Suvi TV, Tuula S, Timo S, Nyberg P. Fluctuating Roles of Matrix Metalloproteinase-9

in Oral Squamous Cell Carcinoma. Scientific World Journal Volume 2013.

133. Gao J. Immunoexpression of matrix metalloproteinase-2 and matrix metalloproteinase-9

in the metastasis of squamous cell carcinoma of the human tongue. Australian Dental

Journal 2010 ; 55 : 385-89.

Bibliography

141

134. Coulombe PA, Omary MB. 'Hard' and 'soft' principles defining the structure, function

and regulation of keratin intermediate filaments. Current Opinion in Cell Biology 2002 ;

14 : 110-22.

135. Jürgen S, Paul EB, Pierre AC, Lutz L, Birgitte L, Thomas MM et al. New consensus

nomenclature for mammalian keratins. Journal of Cell Biology 2006; 174: 169-74.

136. Chu PG, Weiss LM. Keratin expression in human tissues and neoplasms.

Histopathology 2002; 40:403-39.

137. Hiroshi U, Ja-Mun C, Eka F, Miho I, Miyuki S, Kazuya S, Makoto S et al. Soft and

hard keratin expression in Epstein-Barr virus-associated gastric carcinoma. Anticancer

Research 2005; 25: 3183-90.

138. Morgan PR, Shirlaw PJ, Johnson NW, Leigh IM, Lane EB. Potential applications of

anti-keratin antibodies in oral diagnosis. J Oral Pathol 1987;16:212-22.

139. Mei HL, Yang PC, Chang LT, Chung FC. Temporal and spatial sequence expression of

K19 in cultured human keratinocyte. Proc. Natl. Sci. Counc. 2000; 24: 169-77.

140. Tencate AR. Oral Histology: development, structure and function by C. A. Squier and

M. W. Hill. C. V. Mosby Company. 6thedition 2003 ;54-56.

141. Hisham IO, Kora AS. Immunohistochemical detection of cytokeratin 14 in developing

enamel organ of rat molars. Egyptian Dental Journal. 2006; 52: 1399-1407.

142. Lindberg K, Rheinwald J. Suprabasal 40kd (K 19) expression as an immunohistologic

marker of premalignancy in oral epithelium. Am J Pathol. 1989; 134: 89-98.

143. Lan Su, Morgan P, Birgitte L. Keratin 14 and 19 expression in normal, dysplastic and

malignant oral epithelia. A study using in situ hybridization and immunohistochemistry.

J Oral Pathol Med. 1996 ; 25: 293-301.

144. Kale A, Mane D, Babji D, Gupta K. Establishment of field change by expression of

cytokeratins 8/18, 19 and MMP-9 in an apparently normal oral mucosa adjacent to

squamous cell carcinoma: An immunohistochemical study. JOMFP. 2012; 16: 10-15

145. Crivelini MM, de Araujo VC, de Araujo NS. Cytokeratins in epithelia of odontogenic

neoplasms. Oral Diseases 2003; 9: 1–6.

146. De Araujo, Cavalcanti V, De Sousa, Orsini S, Carvalho, Rodarte Y et al. Application of

immunohistochemistry to the diagnosis of salivary gland tumors. Applied

immunohistochemistry and molecular morphology. 2000; 8: 195-202.

Bibliography

142

147. Speight PM. Update an oral epithelial dysplasia and progression to cancer. Head and

Neck Pathol 2007 ; 1 : 61-66.

148. Richart RM. Natural history of cervical intra epithelial neoplasia. Clin Obstet Gynecol

1967 ; 10 : 748-84.

149. Fleskeus S, Slootweg P. Grading systems in head and neck dysplasia : their prognostic

value, weakness and utility. Head and Neck Oncology 2009 : 11.

150. Houston WJ. The analysis of errors in orthodontic measurements. Am J Orthod1983 ;

83 : 382-390.

151. Garrant PR. Oral mucosa in Oral cells and tissues Quintessence Publishing Co. Inc.1st

Edition 2003, Chapter 3 & 4 ; 81 - 121.

152. Tencate AR. Oral Mucosa : Oral Histology development, structure and function by C.

A. Squier , M. W. Hill. C. V. Mosby Company 3rd Edition Chapter 18; 1989: 345-347.

153. Tilakaratne WM, Sherriff M, Morgan PR, Odell EW. Grading oral epithelial dysplasia:

analysis of individual Features. J Oral Pathol Med 2011 ;40: 533–540.

154. Smith PE, Wilfred M. Bailey’s textbook of Histology, The William and Wilkins

Company Baltimore, 11th Ed. 1944 Chapter I. The Cell Page No. 11-12.

155. Koss LG, Melamed RM. Diagnostic cytology and its histological bases. Volume 1, 5th

edition, Lippincott.

156. Hernandez VD. Nucleolus: from structure to dynamics. Histochem. Cell Biol 2006;

125(1-2) : 127-137.

157. http://wiki.answers.com/Q/What_does_the_nucleolus_do?

158. Boulon S, Westman BJ, Hutten S, Boisvert F, Lamond AI. The nucleolus under

stress. Mol Cell 2010; 40: 216–227.

159. Pederson T, Tsai R. In search of non ribosomal nucleolar protein function and

regulation. J Cell Biol 2009; 184: 771–776.

160. Pianese G. Beitragzur Histologie and Aetiologie der Carcinoma. Histologische und

experimentelle Untersuchungen Beitr Pathol Anat Allgem Pathol. 1896; 142: 1–

193.http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2475768/

161. Rajput DV, Tupkari JV. Early detection of oral cancer: PAP and AgNOR staining in

brush biopsies. Journal of Oral and Maxillofacial Pathology 2010 ;14 Issue 2 Jul -

Dec:52- 58.

Bibliography

143

162. Elangovan T, Mani NJ, Malathi N. Argyophilic nucleolar organizer regions in

inflammatory, premalignant and malignant oral lesions. A quantitative and qualitative

assessment. Indian J Dent Res 2008 ;19(2).

163. Kujan O, Oliver RJ, Khattab A, Roberts SA, Thakker N, Sloan P. Evaluation of a new

binary system of grading oral epithelial dysplasia for prediction of malignant

transformation. Oral Oncology 2005;12 : 14.

164. Cohen J. Weighted kappa: nominal scale agreement with provision for scaled

disagreement or partial credit. Psychol bull. 1968;70 : 213-30.

165. Fleiss LJ, Cohen J. The equivalence of weighted kappa and the intra class correlation

coefficient as measures of reliability 1973 ;33 : 613-619.

166. Kujan O, Khattab A, Oliver RJ, Roberts SA, Thakker N, Sloan P. Why oral

histopathology suffers inter observer variability on grading oral epithelial dysplasia : An

attempt to understand the sources of variation. Oral Oncology 2007 ; 43 : 224-231.

167. Patridge M, Emilion G, Pateromichelakis S, Phillips E, Langdon J. Field cancerization

of oral cavity : comparison of the spectrum of molecular alterations in cases presenting

with both dysplastic and malignant lesions. Eur J Cancer B Oral Oncology 1997 ; 33

(b) : 332-337.

168. Upasani OS, Vaidya MM, Bhisey AN. Database on monoclonal antibodies to

cytokeratins. Oral oncology 2004 ; 40 : 236-256.

169. Takeda TK, Sugihera YH, Hirano M., Tanuma JB, Semba I. Immunohistological

evaluation of ki-67, p63, CK-19 and p63 expression in oral epithelial dysplasia. J Oral

Pathol Med 2006; 35 : 369-75.

170. McCawley LJ, Matrisian LM. Tumor progression: Defining the soil round the tumor

seed. Current Biology 2001; 11(1):25–27.

CONTENT

Sl. No. Contents Page No.

1 Background of the proposed research 1 – 3

2 Aim and Objectives 3

3 Materials and Methods 4 – 12

4 Results 12 – 15

5 Discussion 15 – 27

6 Conclusion 27 – 29

7 Bibliography 29 – 32

1

BACKGROUND OF THE PROPOSED RESEARCH

The significance of oral cancer involves its mortality and the disfigurement and

dysfunction associated with treatment. The rate of curability of cancer depends on the stage and

site. Small cancers for example on anterior tongue, floor of the mouth and buccal mucosa have a

local control rate upto 90%. Moderately advanced lesions without evidence of spread to cervical

lymph nodes have control rates upto 80-90%, whereas the intra oral carcinomas with cervical

lymph node metastasis, the survival rate is less than 50%.1

The mortality of oral cancer has increased seven times in the last 50 years in-spite of

millions worth research going on globally. India has one of the highest rates in oral cancer in the

world partly attributed to high prevalence of tobacco chewing.

In parallel to the increase in Oral Cancer, border-line malignant lesions which range from

epithelial dysplasia to intra epithelial carcinoma have also increased in numbers. These lesions are

significant part of routine oral pathology service. This is due to increased awareness, improved

consciousness towards health, availability of health care centers at primary, secondary and tertiary

levels, which results in early detection of oral borderline cases.

Over 90% of all oral cancers are squamous cell carcinomas (SCC) arising from the lining

mucosa. SCC may arise de novo in clinically normal mucosa, though a majority is preceded by

visible changes of the mucosa. There are several histologically distinct lesions of oral mucosa

which have malignant potential. Among these are leukoplakia, erythroplakia, proliferative

verrucous leukoplakia, oral submucous fibrosis etc. These vary in their malignant potential and

genetic background. The lesions may be found in association with and / or preceding SCC.

The term ‘Precancer’ has assumed newer terminologies as “Potentially malignant lesions,

oral precursor lesions, premalignant conditions etc”. These lesions present either as white macule

with excess surface keratin or as red, non-blanching macules with minimal surface keratin.

Histologically these precancerous lesions present with epithelial dysplasia, which is described by

Stedman's Medical Dictionary as a disorder of differentiation of epithelial cells which may

regress, remain stable, or progress to invasive carcinoma. The features of epithelial dysplasia

include: Increased nuclear to cytoplasmic ratio, acanthosis, atypical mitosis, pleomorphism,

nuclear hyperchromatism, enlarged darkly stained nucleoli, loss of cellular adhesion and

abnormal keratinization. These features vary with individual tissue specimens and assessment

2

varies with pathologists viewing the same tissue specimen. Prognostic significance of a lesion is

difficult to determine because of the above factors.2,3

Not all precancerous lesions and conditions transform to cancer. Morphological

alterations amongst some may have an increased potential for malignant transformation. These

are also indicators of risk of likely future malignancies elsewhere in clinically normal appearing

oral mucosa and not only site specific predictors.4

There is always a challenge for oral pathologists to assess the degree of dysplasia in

potentially premalignant oral lesions with accuracy, so that the burden of malignant

transformation can be predicted and treatment modalities can ensure reduced morbidity.5


mainly due to lack of objective and validated grading features. Many studies have shown wide

variability in the diagnosis and grading of oral epithelial dysplasia (OED) with results

demonstrating only poor to moderate agreement.6

The customary grading system of oral epithelial dysplasia into mild, moderate and severe

does not allow accurate prediction of which cases may eventually transform into malignancy.

Many studies have remarked that precancers with epithelial dysplasia have shown to develop

into cancer more readily than lesions without.7 Nevertheless all precancer or epithelial dysplasia

do not develop into cancer, whereas, some have even shown to regress with time.8

The limitations inherent in the histologic grading, suggest the necessity to improve

histologic assessment and the need to identify more reliable makers. In the last decade approach

to understanding of molecular carcinogensis has revealed newer directions. Many molecular

markers functioning as elements of cell cycle control have been recognized. These have been

tested individually and in combinations to develop diagnostic markers for identification of

dysplasia and their ability to predict malignant potential.9

Most of the authors have concluded that though the progress in molecular oncology has

significantly advanced the knowledge on tumorigenesis, yet the practical applications of these

genetic makers remain unresolved in detecting oral dysplasia. Not many studies have focused

correlating the findings at molecular level with the clinical attributes of terminal outcome.

The present study was undertaken to assess the different grading systems to determine the

inter observer and intra observer variation and to study the individual features defining atypia

and general disturbance in the epithelium of oral epithelial lesions; dysplasia. An attempt is also

3

made to correlate the histopathological features with Immunohistochemistry markers for

cytokeratin (CK-19) basement membrane enhancement protein (Ki-67) and extracellular matrix

change (MMP-9) so as to substantiate the observation of atypia observed in H&E stained tissue

sections. The study will also try to validate a new binary classification which will objectively

define each feature and classify OED as ‘risky’ and ‘non-risky’ using the Hematoxylin and Eosin

stains as gold standard.

AIM AND OBJECTIVES

Primary Objective (Aim) :

To propose a new classification for dysplasia to categories dysplasia into ‘non risky’ and

‘risky’ groups

Secondary Objective :

1. To histopathologically evaluate 100 cases of potentially malignant disorders (PMD) for

degree of dysplasia based on the classification system given by (a) Smith and Pindborg’s

classification (1969) as mild, moderate and severe, using photographic standardization

and scores (b) WHO classification (2005) as mild, moderate & severe dysplasia and (c)

Binary classification as low risk and high risk cases given by Kujan Omar (2006).

2. To assess the inter and intra observer variability and agreement for the above mentioned

3 classification systems.

3. To histopathologically evaluate the degree of dysplasia for the same 100 cases of

potential malignant disorders (PMD) using the proposed classification.

4. To assess the inter and intra observer variability and agreement for the proposed

classification.

5. To identify individual histopathological features within the proposed classification that

would aid in categorization of risk behavior.

6. To correlate the proposed classification for degree of dysplasia with

immunohistochemical markers for : CK-19 for epithelial differentiation, Ki-67 as

proliferative marker and MMP-9 as marker for epithelial - connective tissue interaction.

7. To correlate the observations of IHC with histopathological features analyzed with the

new proposed classification.


After obtaining an institutional ethical clearance the present retrospective study is based


commonly used classification


Pilot Study :


OED using a set of 25 archival cases

IDS - Belgaum, DMDC - Wardha, GDC


were undertaken to determine the degree of dysplasia namely WHO 1978, Smith & Pindborg’s



assessment for each case was recorded by the observer as “no dysplasia / mild / moderate /

severe dysplasia”. The results were tabulated on excel sheet

observer variability. WHO 1978

in the moderate group, whereas

a moderate range. WHO 2005

No. 1). Following this the set of same 25 cases were assessed


cytological changes and architectural features.

observed. Suggesting the degree of

classification given by Smith and Pindborg

Estimation of Sample Size:


4


obtaining an institutional ethical clearance the present retrospective study is based


commonly used classification system. The grading of epithelial dysplasia has also be done using



OED using a set of 25 archival cases collected from different oral pathology centers (KLE VK

Wardha, GDC - Nagpur and LMDC - Nagpur). These were reviewed


ermine the degree of dysplasia namely WHO 1978, Smith & Pindborg’s



for each case was recorded by the observer as “no dysplasia / mild / moderate /

The results were tabulated on excel sheet for analysis of

WHO 1978 classification showed poor to good agreement, m

, whereas Smith and Pindborg’s was fair to moderate agreement, following

a moderate range. WHO 2005 demonstrated moderate agreement among all observers

Following this the set of same 25 cases were assessed by the binary system suggested by


cytological changes and architectural features. Moderate to almost perfect agreement

Suggesting the degree of agreement was best in the binary system followed by

classification given by Smith and Pindborg.


obtaining an institutional ethical clearance the present retrospective study is based


system. The grading of epithelial dysplasia has also be done using


collected from different oral pathology centers (KLE VK

Nagpur). These were reviewed


ermine the degree of dysplasia namely WHO 1978, Smith & Pindborg’s



for each case was recorded by the observer as “no dysplasia / mild / moderate /

for analysis of the inter and intra

ood agreement, maximum being

air to moderate agreement, following

emonstrated moderate agreement among all observers (Annex.

by the binary system suggested by


Moderate to almost perfect agreement was

agreement was best in the binary system followed by

Where: n=sample size

N= no. of subjects in the entire population.

n* = depends (r

r = relative error

pa = overall agreement probability

pe = chance agreement probability


20%. i.e. pa-pe = 0.5 and r=20




validation of these classifications is also based on immunohisto chemical markers.


from the archives of the department which reported to KLE VK Institute of Dental Sciences

OPD from 1995 to 2012. From cases

severe dysplasia and carcinoma in

available were included. Biopsy samples obtained from lesions on buccal mucosa were

considered. 100 slides from 100 different cases

study material.

In the series of selected 100 cases, the


Demographic data :

Clinical data was obtained from the biopsy requisition form f

The demographic data like age, sex, type of habit, frequency and duration of tobacco habit w

5

n=sample size

of subjects in the entire population.

depends (r and pa-pe)

r = relative error

overall agreement probability

chance agreement probability


pe = 0.5 and r=20




tion of these classifications is also based on immunohisto chemical markers.


department which reported to KLE VK Institute of Dental Sciences

From cases with signed out diagnosis of hyperplasia, mild, moderate,

severe dysplasia and carcinoma in-situ were selected. Samples with sufficient depth of tissue

lable were included. Biopsy samples obtained from lesions on buccal mucosa were

from 100 different cases were finally included,

In the series of selected 100 cases, the 25 cases included for pilot


Clinical data was obtained from the biopsy requisition form f or these selected 100 cases.

The demographic data like age, sex, type of habit, frequency and duration of tobacco habit w




tion of these classifications is also based on immunohisto chemical markers.


department which reported to KLE VK Institute of Dental Sciences

with signed out diagnosis of hyperplasia, mild, moderate,

Samples with sufficient depth of tissue

lable were included. Biopsy samples obtained from lesions on buccal mucosa were

finally included, which formulated the

25 cases included for pilot study were not

r these selected 100 cases.

The demographic data like age, sex, type of habit, frequency and duration of tobacco habit w as

6

noted. The clinical provisional diagnosis ranged as Leukoplakia, erythroplakia,

erythroleukoplakia and oral sub mucous fibrosis. Confidentiality and maintenance of all the

obtained records of all the cases included in this research protocol will be strictly maintained by

stating the waiver of consent.

Procedure :

Four sections of 4μeach were prepared by sectioning paraffin embedded tissue blocks.

One section was stained with H & E for histopathological evaluation and the one section each

was obtained on gel coated slide and stained immunohistochemically for MMP9, Ki-67 and with

CK-19 using HRP-IHC. The staining procedures and analysis is been stated in manual of

operation and standard operating procedures.

Assessment Method for Epithelial Dysplasia:

For final assessment the grading classifications given by Smith and Pindborg’s grading

system (1969), WHO (2005) and the Binary grading system (2006) were used for evaluation of

H&E stained sections of 100 cases. No calibration exercise was attempted. Each slide was

labeled only with a serial order number. No clinical data or patient information was provided

with the slides to the evaluator. Four independent oral pathologists reviewed the slides. Sections

were observed under 4x; 10x; 100x magnification using Olympus BM2500 compound

microscope. The architectural characteristics of each feature was examined under the low

magnification (4x & 10x) and finished with cytological characteristics at high magnification

(40x).

A score sheet was designed for recording the individual features to aid systematic

analysis. Separate score sheets based on the selected three classifications were prepared. The

lesions were graded as ‘0 for no dysplasia ; 1 for Mild ; 2 for Moderate and 3 for Severe

dysplasia / carcinoma in-situ. For Binary grading system ‘1 for Low risk and 2 for High risk’ was

scored.

Individual scores of all raters were obtained and in addition to the individual scores of the

raters a consensus score of 3 raters (SH, DM & PA) was obtained to compare with the

investigator’s (AK) score. Based on the proposed classification of the present study the observers

evaluated all the 100 cases as ‘non-risky’ and ‘risky’ for inter and intra observer variability for

all 18 dysplastic features. For 10 important features predicted for malignant transformation all

the 100 cases were evaluated.

7

Classifications considered for the pilot study and the present study :1. Histologic features for Smith & Pindborg 1969 classification of oral epithelial dysplasia.

Type of change Severity of dysplasia & scores1. Drop shaped reteridges None-0 Slight-2 Marked-42. Irregular epithelial stratification None-0 Slight-2 Marked-53. Keratinization of cells below keratinized layer None-0 Slight-1 Marked-34. Basal cell hyperplasia None-0 Slight-1 Marked-45. Loss of intercellular adherence None-0 Slight-1 Marked-56. Loss of polarity None-0 Slight-2 Marked-67. Hyperchromatic nuclei None-0 Slight-2 Marked-58. Increased nucleo-cytoplasmic ratio in basal and

prickle cell layersNone-0 Slight

Increase-2MarkedIncrease-6

9. Anisocytosis and anisonucleosis None-0 Slight-2 Marked-610 Pleomorphic cells and nuclei None-0 Slight-2 Marked-611 Mitotic activity Normal-0 Slight

Increase-1MarkedIncrease-5

12 Level of mitotic activity Normal-0 Slight-3 Marked-1013 Presence of bizarre mitoses None-0 Slight-6 Marked-10

2. WHO criteria for grading dysplasia (1978) : It graded epithelial dysplasia as: Mild,

Moderate, Severe.

Mild dysplasia: Slight nuclear abnormalities, most marked in the basal third of the epithelial

thickness and minimal in the upper layers, where the cells show maturation and stratification. A

few, but no abnormal mitosis may be present, usually accompanied by keratosis and chronic

inflammation.

Moderate dysplasia: More marked nuclear abnormalities and nucleoli tend to be present, with

changes most marked in the basal 2/3rd of the epithelium, nuclear abnormalities may persists up

to the surface, but cell maturation and stratification are evident in the upper layers. Mitoses are

present in the parabasal and intermediate layers, but none is abnormal.





carcinoma in situ, in which the whole or almost the whole thickness of epithelium is involved.

In the present knowledge, it was not possible to say whether the presence of severe dysplasia

carried a different degree of risk of subsequent development of invasive carcinoma than the

presence of carcinoma in situ.

8

3. WHO criteria for grading dysplasia (2005) :

Architecture criteria Cytology criteria1. Irregular epithelial stratification 1. Abnormal variation in nuclear size

(Anisonucleosis)2. Loss of polarity of basal cells 2. Abnormal variation in nuclear shape

(Nuclear pleomorphism)3. Drop-shaped rete ridges 3. Abnormal variation in cell size

(Anisocytosis)4. Increased number of mitotic figures 4. Abnormal variation in cell shape

(Cellular pleomorphism)5. Abnormal superficial mitoses 5. Increased nuclear cytoplasmic ratio6. Premature keratinisation in single cells(dyskeratosis)

6. Increased nuclear size

7. Keratin pearls within rete ridges 7. Atypical mitotic figures8. Increased number and size of nucleoli9. Hyperchromatism

Grading of dysplasia:

Mild dysplasia: In general architectural disturbance limited to the lower third of the epithelium

accompanied by cytological atypia define the minimum criteria of dysplasia.

Moderate dysplasia: Architectural disturbance extending into the middle third of the

epithelium is the initial criterion for recognizing this category. However, consideration of the

degree of cytologic atypia may require upgrading i. e. those lesions that show marked cytological

alteration should be elevated to a higher grade level regardless of how extensively the atypical

cells fill the epithelium.

Severe dysplasia: Recognition of severe dysplasia starts with greater than two thirds of the


architectural disturbance extending into the middle 3rd of the epithelium with sufficient cytologic

atypia may be up graded from moderate to severe dysplasia.

Carcinoma in situ: The theoretical concept of carcinoma in situ is that malignant transformation

has occurred but invasion is not present. It is not always possible to recognize this

morphologically. The following is recommended for the diagnosis of carcinoma in situ: full

thickness or almost full thickness architectural abnormalities in the viable cellular layers

accompanied by pronounced cytologic atypia. Atypical mitotic figures and abnormal superficial

mitoses are commonly seen in carcinoma in-situ.

9

4. Binary Classification (2006) :



Low-risk lesion (does not have the potential susceptibility for malignant transformation) was

based on observing less than four architectural changes or less than five cytological changes. The

features to classify as High and Low Risk were based on the architectural and cytological

features suggested by WHO 2005 classification.

PROPOSED WORKING CLASSIFICATION

The proposed classification is based on Binary System (2006) and WHO (2005) criteria of

epithelial dysplasia. Additional architectural and cytological features are identified and included

in this classification. Each features of epithelial dysplasia is scored on a rating scale of 1 to 10.

The two groups identified are :

• Non risky lesions : The total score ranged from 1 to 15

• Risky lesions : The total score ranged from 16 to 84

• ARCHITECTURAL FEATURES:

Sl. No. Particulars Scores1. 1.1. Basal cell hyperplasia 42. 1.2. Drop shaped rete ridges 23. 1.3. Few normal mitotic figures 14. 1.4. Acanthosis 15. 1.5. Loss of polarity of basal cells 56. 1.6. Drop shaped rete ridges with nodules or forking 47. 1.7. Loss of cohesiveness with prominent intercellular bridges 58. 1.8. Abnormal and superficial mitoses 109. 1.9. Increased number of mitotic figures 510. 1.10. Dyskeratosis 211. 1.11. Loss of epithelial stratification 512. 1.12. Keratin pearls within rete ridges 5

• CYTOLOGICAL FEATURES:

Sl. No. Particulars Scores1. 2.1. Mild to moderate degree of nuclear and cellular pleomorphism 52. 2.2. Marked nuclear & cellular pleomorphism &anisonucleosis 103. 2.3. Hyperchromasia of cells 54. 2.4. Increased nuclear-cytoplasmic ratio 45. 2.5. Atypical mitotic figures 66. 2.6. Increased number and size of nucleoli 4

10

INFLAMMATION :

Sl. No. Particulars Scores1. 3.1 Mild 12. 3.2 Moderate 23. 3.3 Severe 3

Procedure for H&E staining and Immunohistochemistry :

Four sections each of selected archival blocks were cut, one slide stained routinely with H&E

to confirm the diagnosis and the other slide stained with immnunohistochemical stain. The sections

required for immunoexpression is taken on a gel coated slides using 2%, 3-aminopropylethoxysilane

solution (APES) (Sigma Aldrich, St.Louis, MO, USA, Product code: A3648). The buffers required

for IHC are PBS as wash buffer and citrate buffer for antigen retrieval. The immunohistochemical

protocol was followed as per the standard recommendations of the company.

The IHC staining is performed using Super Sensitive one step Polymer HRP system

(QD-600, Biogenex, Bangalore) with pre-diluted primary antibodies using anti-Human antibody

for CK-19 (clone RCK 108, Biogenex, Bangalore), Anti MMP-9 (EP1-255Y, Biogenex,

Bangalore), Ki-67 (BGX-297, Biogenex, Bangalore). The storage and dilution of these

antibodies and IHC kit is done as per the manufacturers’ recommendation (Annex. No. 3).

Assessment of Immunohistochemical stained sections:

All the IHC stained slides were evaluated for immunoexpression of MMP 9, Ki 67, CK

19. The intensity of immunohistochemical staining was graded based on subjective evaluation ofcolor exhibited (brown color) by antigen, antibody and chromogen complex. For Ki-67, the

immunoexpression was noted for nuclear staining, CK-19 in cytoplasm and MMP-9 for stromalreaction. Evaluation of immunoexpression was noted under x10 magnification under light

microscope.

1. Analysis of immunoexpression of MMP-9

The immunoexpression of MMP-9 was based on intensity and area of expression in the

epithelium and stroma on a 4 point scale from 0 to 3. Intensity of expression was graded as 0

when there was no evidence of immunoexpression. Intensity was considered positive depending

upon the colour intensity ranging from 1 : golden brown, 2 : light brown, 3 : dark brown. Area

of expression was based on a percentage of positivity in the stroma and epithelium. The range

was considered 1: when 1 to 25% of positive expression was noted, 2 : 26 to 50% of positivity

and 3 : more than 50% of positive expression.

11

Intensity Area in epithelium and stroma0 = -ve 0 = -ve1 = + , golden brown 1 = 1-25%2 = + + , light brown 2 = 26-50%3 = + + + , dark brown 3 = >50%

2. Analysis of immunoexpression of Ki-67

The immunoexpression of Ki-67 was mainly assessed in nuclei of epithelial cells. This

marker was evaluated in the epithelium based on the area of expression and percentage ofpositive immunoexpression on a four point scale from 0 to 3. When there was no evidence of

immunoexpression it was graded as 0. In the epithelium this proliferative marker wasassessed based on the layers of the stratified squamous epithelium involved. Whenever

expression was noted only in the basal cell layer it was graded as 1, suprabasal layer wasgraded as 2 and layers above this was graded as 3. Percentage of number of cells for positive

immunoexpression was assessed. Area of expression was graded as 1: when 1 to 25%positive epithelial cells were noted, 2: 26 to 50% positive cells and 3: more than 50% cells.

The area with maximum number of positive cells was considered in each section.

Area Percentage no. of cells0 : No expression 0 = -ve1 : Basal layer 1 = 1-25%2 : Suprabasal layer 2 = 26-50%3 : Above suprabasal layer 3 = >50%

3. Analysis of immunoexpression of Ck-19

The immunoexpression of Ck-19 was based on intensity and area of expression in the

epithelium on a 4 point scale from 0 to 3. Intensity of expression was graded as 0 when there was

no evidence of immunoexpression. In te ns it y wa s co ns id er ed po si ti ve depending upon the

colour intensity ranging from 1 : golden brown, 2 : light brown, 3 : dark brown. Area of

expression was based on a percentage of positivity in the epithelium. It was considered as 1 for 1

to 25% of positive expression, 2 : 26 to 50% of positivity and 3 : more than 50% of positive

expression in the epithelium.

Intensity Area0 = -ve 0 = -ve1 = + golden brown 1 = 1-25%2 = + + light brown 2 = 26-50%3 = + + + dark brown 3 = >50%

12

STATISTICAL ANALYSIS:

Statistical analysis was done using :1. Weighted Kappa and Non-Weighted Kappa tests for assessment of inter and intra observer

variability for all the classification system

Unweighted kappa takes into consideration agreement for individual cases

Weighted kappa considers the magnitude of disagreement (ordered categories) and

hence is a better measure for reliability

Kappa Agreement (Flesis and Cohen 1973)

< 0.40 Poor

0.40-0.75 Fair to Good

0.76 to 0.80 Excellent

0.81-0.99 Almost perfect

2. Fisher's Exact Test was done to evaluate the correlation of the immunohistochemical

markers with grades of epithelial dysplasia. p<0.001 was considered as statistically

significant.

RESULTS

In the pilot study 25 cases of OED were evaluated for inter and intra observer variation

using four classifications :- a) Grading of dysplasia given by WHO (1978), b) Smith and

Pindborg Classification using photographic standards (1969) c) Classification of WHO (2005)

and d) Binary Classification by Kujan Omar et al (2006).

For the present study 100 cases of OED were subjected to evaluation for two selected

classifications: Smith and Pindborg (1969) and WHO (2005). A new classification is proposed

which included scoring system and has two categories for grading dysplasia as ‘non-risky’ and

‘risky’. IHC evaluation for Ki-67, CK-19 and MMP-9 is done to understand behavior at the

molecular level. The data obtained has been statistically analyzed.

Statistical analysis of 25 cases considered for Pilot Study : WHO (1978) showed good to poor

agreement with maximum having moderate agreement. Smith and Pindborg (1969) had fair to

moderate agreement following a moderate range. Whereas WHO (2005) demonstrated moderate

agreement amongst all observers. Moderate to almost perfect agreement was observed in binary

classification by Kujan O. The inference drawn from the pilot study suggests that the degree of

13

agreement was best in the binary system followed by classification given by Smith and Pindborg.

(K=kappa, kw=weighted kappa)

Statistical analysis of 100 study cases considered for INTER OBSERVER variability : Smith and

Pindborg’s classification showed fair to moderate agreement in both weighted and unweighted

kappa statistics (p<0.001). Similarly, WHO (2005) classification also showed fair to moderate

inter observer agreement. Good agreement was evident with proposed classification.

Statistical analysis of 100 study cases for INTRA OBSERVER variability : The intra observer

agreement was found to be almost perfect in the proposed classification as compared to the other

classification.

Distribution of cases into ‘non-risky’ (total score ranged from 1-15 ) and ‘risky’ (total

score ranged from 16 to 84) groups : 59 cases were identified as risky dysplasia as compared

to 41 cases of ‘non-risky’ dysplasia. Suggesting that 59 cases out of 100 cases are likely to

develop into malignancy.

Calculation for risk categorization using median and inter quartile for 100 cases.

Classification of risk categories for evaluation of 59 ‘risky’ cases : Entire score for the

proposed classification comprising of 1 to 84 was evaluated for mean and standard deviation.

Median and the inter quartile range was used to define cut off points and the cases were classified

into low, medium and high risk based on the scores above. Out of the 59 ‘risky’ cases 6(10.2%)

cases had highest score (67-84) were suggestive to be at higher risk for malignant transformation.

Risk categorization of ‘Risky’ and ‘Non-risky’ 100 cases for each feature : Cases of both

non-risky and risky categorized into three risk categories as per the scores for 18 individual

features. Feature 1.8* (Abnormal and superficial mitoses) and 2.2* (Marked nuclear & cellular

pleomorphism & anisonucleosis) were significantly associated with the risk group as compared

to the non-risky group. (Fisher Exact Test p<0.0010 & p<0.003).

Ten Important Features with higher scores considered in ‘Risky group’ are segregated

with a total score of 61 : Loss of polarity of basal cells, Loss of cohesiveness with prominent

intercellular bridges, Abnormal and superficial mitoses, Increased number of mitotic figures,

Loss of epithelial stratification, Keratin pearls within rete ridges, Mild to moderate degree of

nuclear and cellular pleomorphism, Marked nuclear & cellular pleomorphism & anisonucleosis,

Hyperchromasia of basal cells, Atypical mitotic figures.

14

Calculation for risk categorization using median and inter quartile for study group for 10

important features. Classification of risk categories for evaluation of cases for 10 features :

The total score of the ten individual features selected comprising of 01 to 61 was evaluated for

mean and standard deviation. The median and the inter quartile range was used to define cut off

points and the cases were again classified into low, medium and high risk based on the ten

individual features. Out of the 59 ‘risky’ cases 8(13.6%) cases had highest score (> 45) were

suggestive to be at higher risk for malignant transformation.

Risk categorization of ‘risky’ and ‘non-risky’ groups for the 10 features : Cases of both non-

risky and risky categorized into three risk categories as per the scores for 10 selected individual

features. Feature 1.8* (Abnormal and superficial mitoses) and 2.2* (Marked nuclear & cellular

pleomorphism & anisonucleosis) were significantly more commonly associated with the ‘risky’

group as compared to the ‘non-risky’ group (Fisher Exact Test p<0.001).

Area of expression of CK-19 in Epithelium of the study groups (Fisher exact Test) : ‘Risky’

group the area of expression of Ck-19 in 32 cases the expression was seen in 26 to 50% of

epithelial cells. In a ‘non-risky’ group 20 cases were negative for Ck-19 expression but 6 cases

had 26 to 50% positive cells. The area of CK-19 expression in the epithelium was significantly

high in the risky as compared to non-risky group.

CK-19 Expression of intensity in Epithelium of the study groups (Fisher exact Test) : The

intensity of expression was light brown in 30 cases and dark brown in 14 cases of ‘risky’ group.

Whereas in the ‘non-risky’ 20 cases were negative and in 21 cases the intensity ranged from light

brown to dark brown. Intensity of CK-19 expression in the epithelium was significantly high in

the ‘risky’ as compared to ‘non-risky’ group. The eight cases with scores 45 & above had more

intensity (+++) and more area (>50%) of expression of CK-19.

Ki-67 Expression in the study groups (Fisher Exact Test) : 39 ‘risky’ cases had 26 to 50%

positive cells for ki-67 as compared to the non-risky group where 24 cases were negative and 11

cases had1 to 25% positive expression. The Ki-67 expression in the epithelium was significantly

higher in the ‘risky’ as compared to ‘non-risky’ group.

Area of Ki-67 positivity in study groups (Fisher Exact Test) : 42 cases in the risky group had

ki-67 immnoexpression upto to the supra basal layer, whereas 6 cases had expression above the

suprabasal layer. In the ‘non-risky’ 24 cases were negative but 6 cases had ki-67

immunoexpression upto the suprabasal layer. The Ki-67 expression in the epithelium was

15

significantly higher in the ‘risky’ group as compared to ‘non- risky’ group. The eight cases with

scores 45 & above had expression upto suprabasal layer (+++) where >50% of cells had

expression of Ki-67.

MMP-9 Expression in Epithelium of the study groups (Fisher Exact Test) : The intensity of

MMP-9 expression was golden brown in 26 ‘non-risky’ cases and 42 ‘risky’ cases. In 13 risky

cases area of MMP-9 expression in epithelial cell was seen between 26-50% whereas in non-

risky 26 cases had 1 to 25% positive cells. The MMP-9 epithelium was significantly higher in

the risky as compared to ‘non-risky’ group.

MMP-9 Expression in stroma of the study groups (Fisher Exact Test) : The intensity of

MMP-9 expression was light brown in 18 ‘non-risky’ cases and 26 ‘risky’ cases. In 37 risky

cases area of MMP 9 expression in stromal cell was seen between 26-50% whereas in non-risky

19 cases had 1 to 25% positive cells. No association was observed in the stromal expression of

MMP-9 in the ‘risky’ as compared to ‘non-risky’ group.

DISCUSSION

Assessment of dysplasia depends upon the microscopic diagnosis which is based on

grading the changes considering combination of architectural and cytological features.10 Grading

of dysplasia is subjective and lacks intra and inter observer reproducibility due to insufficiency

of validated morphological criteria and the biological nature of dysplasia.11 Hence, due to the

absence of a consensus, several systems are currently employed.

During the last decade many classifications of potentially malignant lesions

synonymously referred as epithelial dysplastic lesions, premalignant conditions, squamous intra

epithelial neoplasia etc. have been proposed. The majority of the classifications in the current

literature have followed criteria similar to those in common use for the grading of epithelial

lesions of the uterine cervix.12 In cervical epithelium there is clear distinction between normal

and abnormal layers of the epithelium and consequently the degree of dysplasia can be assessed

by determining the horizontal level of this border in the epithelium, resulting in substantial intra

and inter observer consistency. Such a sharp distinction between normal and abnormal layers in

the epithelium of the upper aero-digestive tract is less obvious and consequently the definitions

for degree of dysplasia are much more susceptible for discussion, thus resulting in substantial

intra and inter observer variability.13

16

Based on the observations from pilot study and after considering an extensive search of

literature, the criteria under architectural features and morphological features defined by WHO in

2005 were redefined to form the basis for the new proposal to grade epithelial dysplasia for the

present study. We have made an attempt to score the individual features of dysplasia. The criteria

for assigning the scores was based on the severity and occurrence as seen in OSCC or in tissues

adjacent to OSCC, features considered for severe dysplasia by other authors and those suggested

by Smith and Pindborg. This was suggested with the objective of reducing the inter and intra

observer variability and for making assessment objectively reproducible.

The two basic requirement of any grading system should be : reproducibility and

prognostic value. Reproducibility refers to the degree to which observer measurement or

diagnosis remains the same on repeated independent observations of an unchanged characteristic.

This consistency can be assessed between different observers (inter observer) or within a single

observer (intra observer). Prognostic value is the natural evolution of a disease without treatment

and predictive value is the natural evolution of the disease in response to treatment.14 For

studying accuracy in grading dysplasia of upper aero-digestive tract, there is no test available,

which is thought to be better than the pathologist’s observation when an accepted gold standard

is not available for assessing the validity obtained for grading these lesions. Therefore,

reproducibility is needed to provide indication of validity. The inter and intra observer agreement

levels give an estimate of the degree of bias and validity, where an appropriate gold standard is

not available.15

Appraisal of comparison of Epithelial Dysplasia classifications:

Grading is hampered by the arbitrary division into distinct categories of a continually

progressing process without any naturally and sharply defined borders. It is in fact, an attempt to

impose discrete categories on what is in effect a continuous grey scale and therefore any grading

scheme is by definition artificial.16 The ultimate diagnosis depends on the emphasis put upon

each of the characteristics in grading by the observer. Several studies have shown inter and intra

observer variability in the assessment of presence or absence of individual characteristic features

and thus the grading of dysplasia using different classification system.

In the present study 100 cases of PMD with epithelial dysplasia were assessed using

Smith and Pindborg’s classification given in 1969, WHO in 2005 and the proposed new

classification for assessing the inter and intra observer variability. Kappa agreement was used to

17

evaluate the degree of variation in grading dysplasia for each cytological and architectural

feature. Both unweighted kappa and weighted kappa analysis were considered.

Unweighted or simple kappa (k) is a measurement of inter-rater agreement which

provides quantitative estimate of reliability and represents agreement greater than that expected

by chance alone. As stated by Cohen the kappa values fall between ‘-1’ : complete disagreement,

‘0’ : agreement accepted by chance and ‘1’ : perfect agreement. Thus kappa gives credit only for

complete agreement.

Whereas weighted kappa (kw) considers the magnitude of disagreement by giving credit

to complete and partial agreement and hence a better measurement for reliability. Weighted

kappa results in a better inter and intra observer agreement values.17, 18 In the present study the

interpretation of kappa values suggested by Flesis and Cohen (1973)19 were considered, where

the values < 0.40 was rated as poor, 0.40-0.75 as fair to good, 0.76-0.80 as excellent and above

0.81 as almost perfect.

The overall inter observer variability in our study for Smith and Pindborg and WHO 2005

classification showed moderate to fair agreement kw=0.48 & 0.149 and k=0.275 & 0.310

respectively. Whereas the proposed classification had 90% agreement (p=0.001) k=0.79. The

intra observer variation for the 100 cases of dysplasia showed similar observation when

compared between the different classifications. Grading was reproducible with good agreement

for Smith and Pindborg and almost perfect agreement (91%, k=0.81) for the proposed

classification.

Great variability in inter and intra observer agreement has been reported in several

studies. They ranged from poor to substantial agreement by use of different statistical methods.

Pindborg et al reported a wide range of agreement (1-78%) for 9 photomicrographs evaluated by

72 pathologists. Abbey LM et al (1995)20 examined 120 oral biopsies of a range of pathologies

from hyperkeratosis to severe dysplasia. The agreement ranged from 35.8% to 55.8%. The kappa

values improved to 0.70 – 0.88 for inter examiner comparisons when the grading was expanded

to “within one histological grade”.

Sudbo J et al (2001)21 sh ow ed po or inter observer agreement using 3 point ordinal

scale (k values: 0.17 – 0.33). A slight improvement was observed when 2 categories of

“favorable” (Mild and moderate dysplasia) and “poor” (severe dysplasia) were used. (K value

ranged 0.21 – 0.34). Brothwell DJ et al (2003)15 showed a substantial agreement using a 5 point

18

ordinal scale with a group weighed Kappa (Kw) of 0.74 (94%, Cl=0.64 – 0.85). However similar

to other studies, in their study the group unweighted Kappa showed a fair to moderate agreement

(ks) of 0.37 with 95% Cl = 0.32 – 0.42.

Fischer D et al (2004)22 reported an overall kappa weighted value of 0.59 (95% cl=0.45 –

0.72) An improvement in the kappa agreement was observed when the histological diagnosis was

simplified into three general categories of “no abnormality / hyperplasia”, “mild / moderate /

severe dysplasia” and “Carcinoma in situ” (Kw=0.70 (95% Cl : 0.56-0.84). Subjective judgment

and individual histopathological experience of the participating pathologists was the core for the

histopathological assessment.

Kujan O et al (2006)23 observed in their study that for a substantial group weighted

agreement was reported when four observers graded oral biopsy specimens into categories

varying from hyperplasia (no dysplasia) to carcinoma in-situ (Kw=0.63, (95% CI: 0.42- 0.78)). A

fair to moderate agreement was found using unweighted kappa test (ks=0.22, (95CI:0.11 : 0.35).

These variations were due to the fact that the unweighted kappa considered all disagreements

between raters. Whereas weighted kappa yields a higher reliability when disagreements are small

compared to, when they are large. An improvement in the pairwise and overall agreement was

reported in their study when a binary grading system was used. They observed a better

agreement between those experienced in examining oral mucosal biopsies as compared with the

general pathologist. Bias could also be due to established departmental practice in oral pathology

as stated by Kujan O et al.

All previously published reports on the inter observer agreement on grading OED lesions

showed considerable variations. Warnakulasuriya S et al24 suggested that this might be due to the

lack of objectivity in the evaluation of established criteria, arbitrary division of the grading, lack

of calibration of criteria and grading, and lack of sufficient knowledge about which criteria are

important for the prediction of malignant potential.

To overcome some of these limitations, we in our proposed classification have given

score to each of the 18 features identified for dysplasia and are predicted to have potential for

malignant transformation. The sum total of the scores ranged from 1 to 84. When the scores

ranged from 1 to 15 for a PMD it was categorized as ‘non-risky’ and when the scores ranged

between 16 to 84 the lesion was graded as ‘risky’. Accordingly 59% of the cases were grouped

under ‘risky’ suggesting that these are likely to develop into malignancy. Whereas literature

19

suggests, 2% to 17% chance of malignant transformation in PMD. The scores assigned to

individual dysplastic characters were in agreement with that given by Smith and Pindborg for

evaluating the dysplastic features more objectively. The total score in our study was probably

over estimated for the group categorized as ‘risky’. The features considered under the group of

‘risky’ were further evaluated for risk categorization. Entire score of the proposed classification

comprising of 1 to 84 was evaluated for mean and standard deviation. Median and inter quartile

range was used to define cut off point for scoring dysplasia and the cases were regrouped into

low risk (score between 22 - 44) which consisted of 23.7% (14) cases, medium risk with scores

between 45-66 were 66.1% (39) cases and in the high risk group we found 10.2% (6) cases with

a scored between 67-84. This suggested that out of 59 cases initially grouped as ‘risky’, 6 cases

could be considered for having potential to undergo malignant transformation as they were on

the higher side of the cut-off values.

To evaluate the 18 individual dysplastic features the ‘non-risk’ and ‘risky’ cases were

categorized for each feature of dysplasia. Feature assigned number 1.8 i.e. abnormal and

superficial mitosis and 2.2 i.e. marked nuclear and cellular pleomorphism and anisonucleosis

were significantly associated with ‘risky’ group as compared to ‘non-risky’ where the Fisher’s

Exact Test value were p<0.001, p<0.003 respectively. Fisher’s Exact test of statistical

significance is useful for categorical data that result from classifying objects into different ways:

it is used to examine the significance of association (contingency) between two kinds of

classification of samples.

To further evaluate the significance of individual features, those dysplastic features that

had score of 5 and above were segregated. A total of 10 such important features predicted for

malignant transformation were identified. These are loss of polarity, loss of cohesiveness with

prominent intercellular bridges, abnormal and superficial mitoses, increased number of mitotic

figures, loss of epithelial stratification, keratin pearls within rete ridges, mild to moderate degree

of nuclear and cellular pleomorphism, marked nuclear & cellular pleomorphism &

anisonucleosis, hyperchromasia of basal cells and atypical mitotic figures. The total score of

these 10 features was calculated as 61.

All the 100 cases were once again categorized into the three risk categories as per the

scores of ten above mentioned selected features. Calculations using median and interquartile

range to define the cut-off point based on which of the ‘risky’ cases were regrouped as low,

20

medium and high risk. 13.6% (8) cases were identified with the total score of more than 45,

suggesting these cases which had score more than cut-off value could be at a higher risk of

undergoing malignant transformation.

Cases of both ‘non-risky’ (41 cases) and ‘risky’ (59 cases) were categorized into risk

categories as per the scores of 10 above mentioned selected individual dysplastic features.

Abnormal and superficial mitosis (feature no.1.8) and marked nuclear & cellular pleomorphism

& and anisonucleosis (feature no. 2.2) were significantly and more commonly associated with

‘risky’ group as compared to the ‘non-risky’ group (Fischer’s Exact Test p<0.001). These

observations suggest that these two features observed in the biopsy of any potentially malignant

disorder would be important criteria for consideration for malignant transformation. Another

important feature observed in the ‘risky’ category was loss of cohesiveness which histologically

presents as prominent intercellular bridges. Our results are in agreement with the data obtained

from the studies of Smith and Pindborg, Kujan Omar, Oliver RJ, Tilakaranate WM.

Variation could result due to the difference in the understanding of these features in terms

of recognition and impact on clinical outcomes. Though some features are associated

significantly with the clinical outcomes, the other features which showed no statistical significant

association are as suggested by Patridge M. et al (1997)25 considered to be important for the

whole grading process.

One of the main objectives of the study by Kujan O et al (2006)23 was to assess the

prognostic value of the binary system using qualitative morphological characteristics. They

found the new binary grading system to be a very good predictor for the malignant changes in

oral epithelial dysplasia with reasonable values of sensitivity and specificity (85% and 80%,

respectively). They predicted the clinical outcomes of dysplastic lesions with certainty

(28/33;84.9%), while the negative predictive value of system was 85%.

They also suggested that cases with either hyperplasia or mild dysplasia could be

considered as ‘watch and wait’, whereas the cases with severe dysplasia or CIS in situ need

intervention. The cases with moderate dysplasia belong to the category of ‘watch and wait’.

Their data showed that moderate dysplasia cases had significant potential for malignant changes

(14/30 ; 46.7%). 14 out of 16 (87.5%) cases, in their study with moderate dysplasia had been

assigned as high risk and progressed to OSCC. Their results further demonstrated that patients

whose OED lesions had been assigned as “high risk” tended to have significantly greater

21

transformation rates than those with lesions assigned as “low risk” (p=0.004, by log-rank test). In

contrast to other studies, the results of the Kujan O et al (2006) suggest that the new binary

grading system has the potential to help the clinicians to make more appropriate treatment

decisions when dealing with OED.

Appraisal of IHC in Epithelial Dysplasia:

Epithelial alterations are the first morphological changes seen during the progress of

malignant transformation of the oral epithelium. These potential malignant disorders are

important lesions for clinical preventive investigations. For better understanding not only are the

clinicopathological evaluation essential but also the molecular-biological changes during

transformation process should be known. A variety of cell biological markers mainly involved in

cell proliferation and apoptosis are studied and some association is described in the literature.

Identical genetic changes in dysplastic oral lesions and oral cancers in some patients

suggest that the progeny of the cells in dysplastic lesions may further modify into a malignant

lesion. A clone of oral keratinocytes may proceed through stages of transformation from normal,

to dysplastic, to malignant. The major event in oral malignancy is excessive proliferation of oral

keratinocytes, which break through the basal membrane and spread in the stroma as suggested by

Upasani OS et al (2004).26

Though progress in molecular oncology has significantly advanced our knowledge on

tumorogenesis; yet the practical applications of these genetic markers remain unresolved in

detecting oral dysplasia. Although many molecules have been studied as intermediate markers,

not many studies have focused correlating the findings with the clinical attributes of terminal

outcome. Combined with routine histopathology studies, the broad based studies of molecular

markers could have a great potential for the determination of prognosis of PMD.

The aim of the present study was to correlate the molecular changes in the epithelium and

basement membrane zone of lamina propria of PMD. An attempt is made in the present study to

correlate the histopathological features with IHC markers for CK-19, Ki-67 and MMP-9 so as to

substantiate the observation of dysplasia observed in H&E stained tissue sections.

Cytokeratins (CK) are epithelial specific intermediate filament proteins. They are 20

different types and exhibit distinct patterns of expression in specific epithelial tissues. The CK

polypeptides have common epitopes against which antibodies are developed and these tend to

cross react. A large number of monoclonal and polyclonal antibodies have been developed that

22

are being used for diagnosis using IHC. CK are connected to the nuclear envelope on one side,

while they interact with plasma membrane proteins. These in turn interact with ECM proteins.

Thus CK are involved in transduction of signals and transport of nutrients from inside to outside

of cells and vice a versa. CK-19 is expressed in simple epithelium and basal cells of non-

keratinizing squamous epithelium reported by Takeda TK et al (2006).27

Basal cells situated along the base of the epithelial ridge have a flattened cell surface in

contact with connective tissue. These non-serrated cells contain only a few cytoplasmic

organelles and appear to be least differentiated cells. These represent the stem cells and transit

amplifying cells. They are characterized by a high nucleus to cytoplasmic ratio, expression of K-

19, a relative lack of keratin filament bundle and high levels of β1 integrins.28

In the present study for all 100 cases the intensity of CK-19 immunoexpression in the

epithelial cells and in the different cell layers were observed. The immunoexpression of Ck-19

was found to be statistically significantly higher (p<0.001) in the ‘risky’ cases as compared to

‘non-risky’. In the ‘risky’ group all the cases showed CK-19 expression whereas 20 ‘non risky’

cases had no CK-19 expression. In the ‘risky’ group, 44 cases out of 59 had more than 26% of

cells with immunoexpression and the intensity of expression ranged from light brown to dark

brown suggesting increase in the number of proliferating cells. 21 ‘non-risky’ cases had immune

expression of CK-19 with maximum having less than 26% of immunopositive cells with

intensity of expression being light brown suggesting the presence of proliferating cells in this

group. The intensity of expression of CK-19 was significantly higher in the ‘risky’ group as

compared to the ‘non-risky’ (p < 0.001). We also found that 6 out of the 8 cases (75%) which

were categorized as high risk with scores more than 45 in the ‘risky’ group had high expression

(+++, >50%) of CK-19. These findings suggest disturbance in the distribution of proliferating

cells and the stem cells during proliferation in the epithelium.

Lindberg K.(1989) 29 in their study suggested CK-19 as a marker of premalignancy when

the supra basal cells of dysplastic epithelium were immunopositive. This is in accordance with

our observation. Takeda TK et al (2006)27 reported CK-19 and p63 over expression in their series

of cases and suggested that CK-19 positive cells are transient applying cells located in the

parabasal layers. Loss of asymmetric cell division leads to increase in number of stem cells in

oral dysplasia. This disturbance of stem cells and asymmetry in cell division can be recognized

as dispolarity of basal cells, an essential histological feature of dysplasia. Thus an increase in

23

suprabasal cells could be an indicator of higher grade of dysplasia. They also suggested that the

decrease in CK-19 expression could be due to alteration of stem cell function and that these cells

could be replaced by proliferating cells in the basal layer of dysplastic epithelium. Hence this

feature could be a useful index in grading.

The expression of terminal differentiation products differ from keratinized to non-

keratinized SSE in which K4 and K13 keratins are expressed in the suprabasal cells, whereas the

suprabasal cells of cornified SSE express high molecular weight keratins (K1, K2, K10 & K11).

This normal pattern of keratin expression is altered in hyperplastic and cancerous lesions of the

oral mucosa. The appearance of K8, K18 and K-19, typical of simple highly proliferative

epithelial cells, becomes prominent in oral leukoplakia and stratified squamous cell carcinoma

was reported by Lan Su et al (1996).30

Ki-67 is a nuclear antigen present in the proliferating cells. It is present in all phases of

cell cycle, where its expression is low in G1 phase and increase to peak levels in the G2 and

mitotic phase. In our study ki-67 expression and the area involved in the epithelial cell layer

showed a significantly marked increase of immunoexpression in the ‘risky’ group as compared

to the ‘non-risky’ (p<0.001). In the ‘risky’ group 48 cases had immunoexpression in more than

26% of the epithelial cells and in 6 cases these cells were extending above the suprabasal layer.

Whereas in the ‘non-risky’ group 17 cases were immune positive for Ki-67, maximum of which

had less than 25% of immune positive cells and were limited to the basal layer. The 8 cases with

scores of more than 45 (high risk in the risky group) in the proposed classification had more

intensely stained and more than 50% immunopositive cells in the upper layers of dysplastic

epithelium. As Ki-67 is located within nucleoli of proliferating cells only, it may be suggested

that the antigen may regulate nucleolar metabolism by increasing rates of ribosomal synthesis

required by rapidly dividing cells which can be histologically interpreted as basilar hyperplasia.

Our findings are in accordance to observations of Takeda TK et al27, they found increase in ki-67

expression in the basal and parabasal cells of dysplasia group as compared to normal oral

epithelium. They did not notice any significant difference between the mild and moderate

dysplasia cases. Takashi Saku31 suggested that differentiation of keratinocytes happens in two

directions : 1) towards the surface – keratinization and 2) towards basal cells. Thus basal cells

are neither the germinal cells nor are they a source for squamous epithelial regeneration, but are

24

terminally differentiated cells. The cells with proliferating potential are located in the parabasal

cell layer which may be regarded germinal center of oral squamous epithelium.

When epithelial cell nuclei are Ki-67 positive they are regarded as cells in proliferation.

In normal epithelium these are located in parabasal cells. In ‘risky’ dysplasia group in our study,

ki-67 positive cells were seen up to the middle third of epithelium. With the presence of

dyskeratosis or sudden keratinization in the lower layer of the epithelium the ki-67 positive cells

could indicate that the keratinocyte differentiation is not well regulated that could lead to lack of

basal cell alignment and loss of stratification.

Oral cancer may result from mutation of ras oncogene and a simultaneous deactivating

mutation of p53, a tumors suppresser gene in oral keratinocyte. Oral cancer may result from

mutation in other oncogenes and in EGFr / ras / kinase cascade / c-myc signaling pathway.

Authors correlated this with the stages of carcinogenesis from normal mucosa, to squamous

hyperplasia (9p), to dysplasia (3p, 17p), to carcinoma in situ (11q, 13q, 14q), to invasive

carcinoma (6p, 8q, 4q).1

In dysplasia the oral keratinocytes proliferation is limited within the epithelial

compartment. Eventually the proliferating keratinocytes may break through the epithelial

basement membrane and expand into the underlying connective tissue resulting in carcinoma.

Frank invasion into the stroma by cancer cells are obvious histologically. However in some cases

the invasiveness is not clearly determined. This so called micro invasion cannot be ruled out in

carcinoma in-situ cases. Takashi Saku31 suggests that invasion of cancer cells is accompanied by

a neoplastic stromal change around the cell nests. On H&E stained sections, the newly formed

neoplastic stroma is more diffusely eosinophilic than the lose connective tissue of the lamina

propria but more basophilically myxoid than surrounding fibrous granulation tissue. Carcinoma

in-situ does not induce such diffuse neoplastic stroma. To evaluate the basement membrane and

stromal changes MMP-9 was considered in the present study.

MMPs are a family of structurally related but genetically distinct enzymes that degrade

extracellular matrix and basement membrane components and regulate cell–matrix composition.

MMPs may be secreted by tumour cells, fibroblasts, endothelial cells and macrophages as well as

mast cells and neutrophils.32

In our study MMP-9 immunoexpression was studied in the epithelium and connective

tissue in all 100 cases of PMD for grading dysplasia. The intensity and area of expression in the

25

epithelium was found to be statistically significant when compared between the ‘risky’ and ‘non-

risky’ cases (Fisher Exact Test, Intensity: p<0.001, Area: p<0.002). The intensity of MMP-9

expression in the epithelium was golden brown in majority of the ‘risky’ and ‘non-risky’ cases.

Where in more than 26% of epithelial cells had enhanced MMP-9 immunoexpression in 37 cases

of ‘risky’ dysplasia. The immunoexpression in stromal cells did not show any significant

correlation / association in the intensity and area between the two study groups (Fisher Exact

Test, Intensity: p<0.551 NS, Area: p<0.700 NS).

The expression of MMP-9 in ‘risky’ group was in accordance with the study on OSCC

where it was noted that MMP-9 is expressed in carcinoma and inflammatory cells around

carcinoma islands.33,34 For tumors to invade and metastasize, neoplastic cells must be capable of

degrading the extracellular matrix and accessing blood vessels and lymphatics. Mounting

evidence supports the view that extracellular proteinases, such as the MMPs, mediate many of

the changes in the microenvironment during tumor progression.35

Richard et al36 in his retrospective study examined changes in MMP 1, 2 and 9 using

polymerase chain reaction in oral dysplasia and oral squamous cell carcinoma and found higher

levels of MMP-1 and 9 mRNA that was significantly associated with dysplasia that progressed to

oral cancer as compared with those that did not. Peschos et al37 checked expression of MMP-9 in

benign, premalignant and malignant laryngeal lesions. They suggested a two-step model of up-

regulation of MMP-9 expression, first when a dysplastic lesion evolves and the next when the

dysplasia progresses to invasive carcinoma. MMP-9 expression was related neither to survival

nor to the other available clinicopathological parameters.

In the present study the stromal reaction of MMP-9 did not show any significant

difference between the cases. Though the stromal reaction in the high risk category cases from

the ‘risky’ dysplasia group were more pronounced as compared to the other cases. Recent studies

have shown that MMPs contribute to tumor progression, through its effect on cell proliferation,

survival and angiogenesis.38

In addition to carcinoma cells, cancers consist of tumor-associated stromal cells, which

include fibroblasts, endothelial cells, leukocytes, macrophages, nerve cells, and adipocytes. The

cancer cells crosstalk with stromal components during cancer progression and they are mediated

by transmembrane receptors, which are expressed on cancer cells and stromal cells.39 Tumor-

associated cells promote angiogenesis, inflammation, invasion, and ECM modeling through cell-

26

cell contact and the production of growth factors, hormones, cytokines, and proteinases such as

MMPs.38

MMP 9 is detected in malignant transformation of various cells and is associated with

tumor metastasis and poor prognosis. In the aggressive human tongue squamous cell carcinoma

cell line, MMP-2 was only found in its latent form, whereas MMP-9 was found in its active

form. In an oral squamous cell carcinoma cell line SCC-25, increase expression of MMP-9, in

vitro, this was thought to occur via fibronectin integrin v 6 pathway. MMP-9 may not be the

only factor for OSCC invasion, but it may be important for indirect cell signaling by controlling

the bioavailability and bioactivity of molecules that target specific receptors, which regulate cell

growth, migration, inflammation, and angiogenesis.39

Gao J (2010)40 studied the expression of MMP-2/MMP-9 in normal oral mucosa, lymph

node-negative tongue cancers, lymph node-positive tongue cancers and their metastasized

tumors in cervical lymph nodes. They hardly found MMP-2/MMP-9 expression in normal

epithelium. In lymph node-negative tongue cancer, 45% and 40% of these primary tumors were

positively stained for MMP-2/MMP-9. Importantly, in lymph node-positive tongue cancer, 71%

and 79% of these primary tumors were positive for MMP-2/MMP-9, respectively. Over

expression of MMP-2/MMP-9 was present in the metastatic lymph nodes. Their result implied

the significance of MMP-2 and / or MMP-9 in predictive value for the actual or potential

presence of cervical metastases. Such activity has prognostic value, and provides impetus for

further development of biotherapies targeted at specific inhibition of MMP activities.



detecting oral dysplasia. None of the molecular markers single or in combination appear to be

ready for use in routine clinical diagnostic practices, although many molecules have been studied

as intermediate markers.9

Furthermore, as suggested by Gayani P. et al (2009)9 recent advances in the use of

quantitative methods has enabled the profiling of gene expression (microarray), protein

expression (proteomics), screening of epigenetic changes, e.g. by pyro-sequencing and single

nucleotide changes (SNP) of hundreds of pre-selected genes simultaneously. These techniques

which require minute amount of tissue when carried out in controlled way and using carefully

selected samples with proper validation will help the potentially malignant and malignant tissue

27

samples. Combined with routine histopathology studies, these broad based studies of gene and

their products appear to have a great deal of potential for the identification of diagnostic and

prognostic markers of pre-cancer, and data from multicenter studies using microarray technology

may be engaged in the future for a better understating of aetiopathogenesis of pre-cancer.

CONCLUSION

Oral epithelial dysplasia : a histopathologic marker of premalignancy, is a diagnostic term used

to describe the histopathological changes seen in progressive chronic and premalignant disorders

of the oral mucosa.. It may also be seen in verrucous or papillary leukoplakia or in the margins of

chronic mucosal ulcers. It is also consistently seen in the apparently normal mucosa adjacent to

the squamous cell carcinoma. Clinically it may present as leukoplakia, erythroplakia or

leukoerythroplakia 41

A histological dysplasia classification system should ideally meet two basic

requirements. First it should be easily applicable in daily routine practice with low inter and intra

observer variability. Secondly, it should allow a clear separation and give clear suggestions to the

clinician to distinguish between patients who need treatment to prevent progression and those

whom no treatment is needed and should be kept on routine observation.

We propose a new binary classification system consisting of two categories where the

individual architectural and cytological features are given a score. This new classification has

been validated and has shown good to almost perfect agreement for both inter and intra observer

variation. This classification has been designed with the following objectives : a) to reduce the

inter and intra observer variability, b) to make it objectively reproducible, c) transmit

maximum information possible from interpretation of routine H&E stained sections, d) give a

clear diagnosis for clinicians to make critical treatment decision.

From the observations of the present study, ten important morphological and architectural

features are identified to be considered for categorizing potentially malignant disorder

histologically as ‘risky’. Abnormal superficial mitosis and marked cellular and nuclear

pleomorphism were two important features that could predict malignant transformation are found

to be statistically significant and that should be identified and considered for risk categorization

and malignant transformation.

The IHC markers using Ki-67 for observing the irregular epithelial cell proliferation, Ck-

19 for loss of polarity and basal cell hyperplasia in the suprabasal layers and MMP-9 for stromal

28

changes, micro-invasion and changes in the basement membrane and lamina propria have

substantiated the H&E observation.

Our proposed system of grading dysplasia is designed for histopathological evaluation

using H&E stained sections. It was designed to extract a larger set of data with detailed

observations of individual features of dysplasia based on architectural and cytological changes.

Our aim was to generate objective data to correlate with defined patterns of dysplastic features.

Grading dysplasia is not an exact science and pathologists are doing their best to reach an

optimal result. In this study we have tried to weigh individual features by defining their

importance and scoring them for OED grading process obtaining a common consensus on the

features. However, data reveals that the proposed classification system can be used to improve or

develop simpler and easy diagnostic methods.

The outcome of ‘risky’ dysplasia for malignant transformation could not be assessed in

the study due to lack of follow up data. Within the limitations of the analysis we have shown that

many individual features of dysplasia can be assessed reproducibly as they were categorized

under high risk with a definite score. Reducing features to definable components increases

objectivity and insight, though it makes it difficult for comparison with the published data.

Future studies should compare lesional responses related to patient population and their

geographic areas of origin. As well as specific carcinogenic factors, diagnostic usefulness of aids

like Toluidine blue application, oral brush biopsy techniques and spectroscopy and the natural

history of the potentially malignant disorders should be correlated.

Relevance of the observations and future scope of this study

• The proposed classification is objective and is easy to understand. As the inter and intra

observer agreement is statistically significant and hence is reproducible.

• The proposed classification has been assigned with scores for specific features that has

helped to minimizes the inter and intra observer variability.

• Reporting and grading of oral epithelial dysplasia in potentially malignant disorders has

been made easy and valid with the use of routine H&E stain section from any

histopathology diagnostic service, center or laboratory.

• Larger sample size studies, at various oral pathology centers, in different geographical

locations could help in developing this proposed classification universally acceptable.

29

• Our classification of ‘risky’ and ‘non-risky’ will be helpful for clinical application in

imparting proper treatment, patient education, ‘wait & watch’ policy and follow-up.

Hence, the need is for:

• Application of the proposed binary classification as ‘risky’ and ‘non-risky’ for routine

use.

• Multicentric focused group assessment using the proposed binary classification.

• Longitudinal follow-up study for evaluating the malignant potential of ‘risky’ and

‘non-risky’ cases of potentially malignant disorders.

BIBLIOGRAPHY

1. Sugerman PB, Savage NW. Current concepts in oral Cancer. Australian Dental Journal 1999;

44(3): 147-156.

2. Bouquet JE, Speight PM, Farthing PM. Epithelial dysplasia of oral mucosa diagnostic


3. Stedman's Medical Dictionary. Copyright © 2006 Lippincott Williams & Wilkins.

4. Warnakulasuriya S, Johnson NW, van derWaal I. Nomenclature and classification of

potentially malignant disorders of the oral mucosa. J Oral Pathol Med 2007; 36: 575-80.

5. Mehrotra R, Pandya S, Chaudhary A, Kumar M, Singh M. Prevalence of oral pre malignant

and malignant lesions at a tertiary level hospital in Allahabad, India. Asian Pacific Journal of

Cancer Prevalence 2008(9) ;

6. Pindborg JJ, Reibel J. Subjectivity in evaluating oral epithelial dysplasia, carcinoma in situ

and initial carcinoma. J Oral Pathology 1985; 14: 695-708.

7. Reibel J. Prognosis of Oral Premalignant lesions, Significance of clinical, histopathological

and molecular biological characteristics. Crit Rev Oral Biol Med 2003; 14(1): 47-62.

8. Holmstrup P, Vedtofe P, Reibel J. Long term treatment outcome of oral premalignant lesions.

Oral Oncology 2006; 42: 461-74.

9. 9.Gayani P, Tilakaratne WM, Tavassoli M, Warnakulasuriya S. Molecular markers in oral

epithelial dysplasia: review. J Oral Pathol Med 2009; 38: 737-752.

10. Speight PM. Update an oral epithelial dysplasia and progression to cancer. Head and Neck

Pathol 2007 ; 1 : 61-66.

30

11. Warnakulasuriya S, Reibel J, Bouquot J, Dabelsteen E. Oral epithelial dysplasia

classification systems : predictive value, utility, weakness and scope for improvement. J Oral

Pathol Med 2008; 37: 127-133.

12. Richart RM. Natural history of cervical intra epithelial neoplasia. Clin Obstet Gynecol 1967 ;

10 : 748-84.

13. Fleskeus S, Slootweg P. Grading systems in head and neck dysplasia : their prognostic value,

weakness and utility. Head and Neck Oncology 2009 : 11.

14. Houston WJ. The analysis of errors in orthodontic measurements. Am J Orthod1983 ; 83 :

382-390.

15. Brothwell DJ, Lewis DW, Bradley G, Leong I, Jordan RCK. Observer agreement in the

grading of oral epithelial dysplasia. Community Dent Oral Epidemiol.2003; 31:300-305.

16. Bosman FT. Dysplasia classification: Pathology in disgrace? J Pathol. 2001; 194: 143-144.

17. Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement

or partial credit. Psychol bull. 1968;70 : 213-30.

18. Fleiss LJ, Cohen J. The equivalence of weighted kappa and the intra class correlation

coefficient as measures of reliability 1973 ;33 : 613-619.

19. Kujan O, Khattab A, Oliver RJ, Roberts SA, Thakker N, Sloan P. Why oral histopathology

suffers inter observer variability on grading oral epithelial dysplasia : An attempt to

understand the sources of variation. Oral Oncology 2007 ; 43 : 224-231.

20. Abbey LM, Gunsolle JC, Page DG. Intra examiner and inter examiner reliability in diagnosis

of oral epithelial dysplasia. Oral Surg. Oral Med. Oral Pathol. 1995; 80: 188-191.

21. Sudbo J, Kildal W, Risberg B, Koppang HS. DNA content as a prognostic marker on patients

with oral leukoplakia. New Engl J Med 2001; 344:1270-8.

22. Fischer D, Joel E, Morton TH, Schwartz S. Inter observer reliability in the histopathologic

diagnosis of oral pre malignant and malignant lesions. J Oral Pathol Med 2004; 33:65-70.

23. Kujan O, Khattab A, Richard O, Stephen R, Thakker N, Sloan P. Evaluation of a new binary

system of grading oral epithelial dysplasia for prediction of malignant transformation. Oral

Oncology 2006; 42:987-993.

24. Warnakulasuriya S. Histologic grading of oral epithelial dysplasia: revisited. J Pathol. 2001;

194: 294-297.

31

25. Patridge M, Emilion G, Pateromichelakis S, Phillips E, Langdon J. Field cancerization of oral

cavity : comparison of the spectrum of molecular alterations in cases presenting with both

dysplastic and malignant lesions. Eur J Cancer B Oral Oncology 1997 ; 33 (b) : 332-337.

26. Upasani OS, Vaidya MM, Bhisey AN. Database on monoclonal antibodies to cytokeratins.

Oral oncology 2004 ; 40 : 236-256.

27. Takeda TK, Sugihera YH, Hirano M., Tanuma JB, Semba I. Immunohistological evaluation

of ki-67, p63, CK-19 and p63 expression in oral epithelial dysplasia. J Oral Pathol Med 2006;

35 : 369-75.

28. Garrant PR. Oral mucosa in Oral cells and tissues Quintessence Publishing Co. Inc.1st

Edition 2003, Chapter 3 & 4 ; 81 - 121.

29. Lindberg K, Rheinwald J. Suprabasal 40kd (K 19) expression as an immunohistologic

marker of premalignancy in oral epithelium. Am J Pathol. 1989; 134: 89-98.

30. Lan Su, Morgan P, Birgitte L. Keratin 14 and 19 expression in normal, dysplastic and

malignant oral epithelia. A study using in situ hybridization and immunohistochemistry. J

Oral Pathol Med. 1996 ; 25: 293-301.

31. Takashi S. The working Committee for New Histopathological criteria for Borderline

Malignancy of Oral Mucosa, The Japanese Society for Oral Pathology (2004). Proceedings

of symposium of JSOP Tokyo.

32. EA Baker, Leaper DJ, Hayter JP, Dickenson AJ. The matrix metalloproteinase system in oral

squamous cell carcinoma. British Journal of Oral and Maxillofacial Surgery 2006 ; 44 : 482-

486.

33. Verma RP, Corwin H. Matrix metalloproteinases (MMPs): Chemical–biological functions

and (Q)SARs. Bioorganic & Medicinal Chemistry 2007;15 :2223–2268.

34. Vihinen P, Kahari VM. Matrix metalloproteinases in cancer: prognostic markers and

therapeutic targets. Int. J. Cancer 2002 ; 99 : 157–166.

35. Lisa J. Mc Cawley, Matrisian LM. Tumor progression: Defining the soil round the tumor

seed. Current Biology 2001; 11(1):25–27.

36. Richard C, Maricris M, Caroline H, Shiboski. Overexpression of MMP-1 and 9 mRNA is

associated with the progression of oral dysplasia to cancer. Clin Cancer Res. 2004 ; 10 :

6460-65.

32

37. Peschos D. Expression of matrix metalloproteinase-9 (gelatinase B) in benign, premalignant

and malignant laryngeal lesions. Histol Histopathol 2006; 21: 603-608.

38. McCawley LJ, Matrisian LM. Tumor progression: Defining the soil round the tumor seed.

Current Biology 2001; 11(1):25–27.

39. Suvi TV, Tuula S, Timo S, Nyberg P. Fluctuating Roles of Matrix Metalloproteinase-9 in

Oral Squamous Cell Carcinoma. Scientific World Journal Volume 2013.

40. Gao J. Immunoexpression of matrix metalloproteinase-2 and matrix metalloproteinase-9 in

the metastasis of squamous cell carcinoma of the human tongue. Australian Dental Journal

2010 ; 55 : 385-89.

41. Lumerman H, Freedman P, Kerpel S. Oral Epithelial dysplasia and the development of

invasive squamous cell carcinoma. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1995;

79: 321-9.

Documents

Doctor of Philosophy IInn In the Faculty ofIn the …shodhganga.inflibnet.ac.in/bitstream/10603/69120/1/24.pdforal epithelial dysplasia as ‘risky’ and ‘non-risky’ using the