Zubair W. Baloch, MD, PhD, FASCP Virginia A. LiVolsi, MD ... · Zubair W. Baloch, MD, PhD, FASCP...

www.ascp.org/ascp2014

1006 An Integrated Clinicopathologic Approach to Thyroid Pathology

Zubair W. Baloch, MD, PhD, FASCP

Virginia A. LiVolsi, MD, MASCP Professor of Pathology & Laboratory Medicine

Perelman School of Medicine, University of Pennsylvania

Philadelphia, PA

www.ascp.org/ascp2014

Speaker Disclosure

In the past 12 months, I have not had a

significant financial interest or other relationship

with the manufacturer(s) of the product(s) or

provider(s) of the service(s) that will be discussed

in my presentation.

www.ascp.org/ascp2014

Follicular Patterned Lesions

• Diagnosis and Classification

–Cytology

–Surgical Pathology

• Histologic Diagnosis

–Ancillary techniques

Ultrasound/autopsy

Palpation

Mazzaferri, 1993

Prevalence of Thyroid Nodules

Case of an Encapsulated Follicular Patterned Lesion

Thyroid Lesions Controversial?

Follicular patterned lesions

The term “Follicular”

• Cell of origin

– Capable of producing thyroid hormone and thyroglobulin

• Architecture / growth pattern

– Follicular – totally or >95% of the lesion displays a follicular growth pattern.

– Microfollicular

– Macrofollicular

Follicular Patterned Lesions

Cytologic Diagnosis & Classification

Follicular Patterned Lesions of Thyroid

• Cytology – Reality Check – FNA is a screening test for follicular patterned

nodules

– Cannot differentiate between follicular adenoma and carcinoma

– Most are diagnosed as “Follicular Lesion / Neoplasm”

– Up to 80% of cases diagnosed as such are benign on histologic examination (hyperplastic nodule or adenoma)

– Approximately half of malignant cases are follicular variant of papillary carcinoma

FNA Diagnosis Follicular Lesion/Neoplasm

Morphologic Criteria

The Usual Teaching

• Monolayer sheets of follicular cells – Benign

• Microfollicles – Neoplasm / Lesion

– Micro-follicular lesion

• Atypical Follicular cells – Neoplasm / Lesion

Is It That Easy

Don’t Think So

Monolayer Sheet

• Sheet of cells arranged in a layer or loosely cohesive group

• Due to presence of large follicles

– Goiter

– Papillary carcinoma

• Follicular variant

– Follicular carcinoma

Microfollicles

• Define and measure the size of follicle

Microfollicles I Don’t Think So

Microfollicles

• Inter-observer Agreement on Microfollicles – Renshaw AA et al. (Arch Pathol Lab Med 2006)

– 12 cytopathologists were shown 45 small groups of follicular cells • 20 Microfollicles

• 7 Macrofollicles

• 18 Indeterminate

– <15 cells arranged in circle that is at least two-thirds complete, should be classified as microfollicles.

Microfollicles

• Mowschenson PM et al (Surgery 1994)

• FNA of normal thyroid tissue may result in the misdiagnosis of micro-follicular lesions

– 42 cases

• 9 unremarkable

• 18 microfollicular

• 3 mixed macromicrofollicular

• 1 Hurthle cell

• 1 Papillary carcinoma

Cytomorphology: Follicular Neoplasm

Diagnosis Follicular Lesion / Neoplasm

Diagnosis Follicular Lesion / Neoplasm

The Bethesda System for Reporting Thyroid Cytopathology:

Implied Risk of Malignancy and Recommended Clinical Management

Diagnostic Category Risk of Malignancy

(%)

Usual Management

Nondiagnostic or Unsatisfactory Repeat FNA with ultrasound

guidance

Benign 0-3% Clinical follow-up

Atypia of Undetermined Significance or

Follicular Lesion of Undetermined

Significance (AUS/FLUS)

~ 5-15% Repeat FNA

Follicular Neoplasm or Suspicious for a

Follicular Neoplasm (Specify if Hurthle type

or Oncocytic)

15-30% Surgical lobectomy

Suspicious for Malignancy 60-75% Near-total thyroidectomy or

surgical lobectomy

Malignant 97-99% Near-total thyroidectomy

Follicular Lesion Histology

• Controversial Diagnosis

– Adenoma vs. Follicular carcinoma vs. Follicular variant of papillary carcinoma

– Lloyd RV et.al. AJSP 2004 (28) • A concordant diagnosis of FVPCA was made by all 10

reviewers with a cumulative frequency of 39%.

• In this series, 24.1% of the patients had metastatic disease (n = 21). In the cases with metastatic disease, a diagnosis of FVPCA was made by all 10 reviewers with a cumulative frequency of 66.7%

– Elshiekh TM et al. AJCP 2008 • 15 cases of FVPCA – unanimous agreement in 2 (13%)

Follicular Carcinoma Diagnosis

Capsular invasion only

• Associated with metastatic disease – Khan & Perzin; Evans et al

• Tumors with capsular invasion also demonstrate foci of vascular invasion on extensive sectioning of the tumor capsule – Yamashina et al

Follicular carcinoma

• Diagnosis based on invasive characteristics

– Capsular invasion

– Vascular invasion

Follicular Carcinoma Diagnosis

Capsular Invasion – Diagnosis

• Tumor cells invading into and through the capsule into the surrounding thyroid parenchyma.

Oncocytic (AKA Hurthle cell) & Follicular Carcinoma Diagnosis

Capsular Invasion – Diagnosis

• Tumor cells invade into the tumor capsule only without invasion into surrounding thyroid. – Tumor cells mushroom out into the capsule

– Tumor cell invade into a hook-like pattern, usually horizontally into the capsule.

Oncocytic (AKA Hurthle cell) & Follicular Carcinoma Diagnosis

• Vascular invasion with or without capsular invasion

– Angio-invasive Oncocytic Follicular / Follicular carcinoma

ADENOMA

MIN INV FOLL-CA

ANGIO-INV FOLL-CA

POST-FNA INV

ENCAPSULATED FOLLICULAR LESIONS

Follicular Adenoma vs. Carcinoma

Follicular Variant of Papillary Carcinoma (FVPTC)

First described by Lindsay in 1960

Chen and Rosai in 1977 • Proposed the term FVPTC • Biologic similarities to conventional Pap Ca • Follicle formation with nuclear features of Pap Ca • Capsular and/or vascular invasion (20%)

FVPTC

Diagnostically challenging cases

• Encapsulated, no invasive features

• Seen in a background of nodular goiter

• Show admixture of micro and macro follicles

• Consists of areas diagnostic of papillary carcinoma and areas that appear benign

Well-differentiated Tumor of Uncertain Malignant Potential Williams et al

Encapsulated follicular patterned lesions

• Minor nuclear changes of PTC

• Minor capsular penetration without nuclear changes of PTC

• Borderline diagnosis

• Extremely good prognosis

Well-differentiated Tumor of Uncertain Malignant Potential Williams et al

• Advantages

–Conservative treatment

• Adequate resection should be enough

• Avoidance of radioactive iodine treatment

Well-differentiated Tumor of Uncertain Malignant Potential Williams et al

Disadvantages

• Lack of commitment on the part of pathologist

• Treatment ???

– Lack of data (long term follow-up)

–Published cases which have shown distant metastasis • Khan and Perzin, Evans, Baloch & LiVolsi

FVPTC Histologic Scenarios

FVPTC Histologic Scenarios

FVPTC Histologic Scenarios

FVPTC Histologic Scenarios

Important Questions/Issues?

• Is the current management of encapsulated FVPTC too aggressive?

– Yes

• Is the diagnosis “WDTUMP” justifiable?

– ?

• Role of immunohistochemical and molecular markers in the diagnosis of PTC

Clinicopathologic Data

Modify the current management of Encapsulated FVPTC

FOLLICULAR VARIANT OF PAPILLARY CARCINOMA

• Liu et al (2006) Memorial series

• Two Types:

– Infiltrative

– Encapsulated

FOLLICULAR VARIANT OF PAPILLARY CARCINOMA

• Statistically significant differences between two types:

• Lymph node mets 65% vs 5%

• ETE 65% vs 5%

• Positive margins 50% vs 2%

• NO NONINVASIVE ENCAPSULATED TUMORS WERE AGGRESSIVE EVEN WITH JUST LOBECTOMY (11 years f/u)!!

Encapsulated FVPTC

• Memorial study (Thyroid 2009) – 43 Encapsulated classical PTC

– 63 Encapsulated FVPTC

– Median F/U 8.3 yrs

• Distant metastasis in 4 FVPTC with extensive capsular and vascular invasion.

• No recurrence in non-invasive FVPTC – 30 patients treated by lobectomy and no RAI

FVPTC F/U HUP Study (Baloch et al. Endo Practice 2010)

• 42 cases of FVPTC 1997-1978

• 25 cases with mean F/U 7.7 yrs

• 1 recurrence (4%)

– Rib metastasis after 7.3 yrs – encapsulated FVPTC no invasion.

The Quest for a Magic Marker?

The Quest for a Magic Marker?

Immunohistochemistry

Thyroid Tumor Microarray Study Coexpression and concurrent absence of expression of immunohistochemical stain panels

Co expression

Malignant

(n=37)

Benign

(n=53)

Sensitivity for

Carcinoma

Specificity for

Carcinoma

HBME+, CK19+, GAL+ 20 0 54% 100%

P16+, ERK+, RET+ 19 6 51% 89%

Concurrent Absence Sensitivity for

Benign

Specificity for

Benign

HBME-, CK19-, GAL- 0 20 38% 100%

P16-, ERK-, RET- 0 15 28% 100%

Role of Molecular Markers in The Diagnosis of Thyroid Lesions

The Challenge for Diagnostic Pathologists

Practicing Morphology in the Molecular Age

What else is available?

• Protein expression:

– Galectin-3

– proteomics

• Genetic analysis:

– Oncogene mutation analysis

• RNA Expression:

– Micro RNA

– Gene expression classifier (AfirmaTM)

What else is available?

• Protein expression:

– Galectin-3

– proteomics

• Genetic analysis:

– Oncogene mutation analysis

• RNA Expression:

– Micro RNA

– Gene expression classifier (AfirmaTM)

Nikiforov Nature Rev Endocrinol 2011 7:569

PAX8/PPARg

1-5%

PTC 70-75% of PTC

Tall cell and classic papillary

cancer

Propensity for dedifferentiation

20-40% poorly differentiated

30-40% anaplastic

Classic papillary cancer

Younger age at presentation

~20% of sporadic PTC

40-70% of radiation induced PTC

Frequent lymph node metastases

Younger age at presentation

10-20% of sporadic PTC

40-70% of radiation induced PTC

(RET/PTC3 assoc with solid variant)

Frequent lymph node metastases

Follicular variant of papillary cancer

More frequent encapsulation

FC

FC

FA

~70% Follicular CA

PAX8/PPARg

RAS HRAS, NRAS codon 61 KRAS codons 12,13

20-40%

FA

30%

40%

2-13%

More Molecular Analysis of FVPTC

• Rivera M & Ghossein RA. 2010 (Mod Pathol) – The infiltrative (un-encapsulated) tumors are more

likely to harbor BRAFT1799A (BRAFV600E) mutations, and have a higher prevalence of lymph node metastases and local recurrences.

– Encapsulated FVPTC are more likely to harbor mutations in the RAS family of oncogenes and exhibit a low recurrence rate in the absence of capsular or vascular invasion.

• Radkay et al (UPMC Group) – 78% of FVPTC showed mutation in NRAS Codon 61

Molecular Analysis of Cytology Specimens

0

10

20

30

40

50

60

70

80

90

100

Follicular neoplasm Susp for malignancy FLUS

Sen

siti

vity

(%

)

Nikiforov (n=513) Cantara (n=95) Moses (n=137)

Nikiforov J Clin Endocrinol Metab 2011 96:3390; Cantara J Clin Endocrinol Metab 2010 95:1365; Moses W J Surg 2010 34:2589

Mutation panel (BRAF, RAS, RET/PTC, PAX8/PPARg):

HIGH False Negative Rate 20-40%

NOT optimal sensitivity

0

10

20

30

40

50

60

70

80

90

100

Follicular neoplasm Susp for malignancy FLUS

Spec

ific

ity

(%)

Nikiforov (n=513) Cantara (n=95) Moses (n=137)

HIGH specificity

Nikiforov J Clin Endocrinol Metab 2011 96:3390; Cantara J Clin Endocrinol Metab 2010 95:1365; Moses W J Surg 2010 34:2589

Mutation panel (BRAF, RAS, RET/PTC, PAX8/PPARg):

Very Low Frequency of

Mutation Markers

in Atypia/FLUS and Follicular

Neoplasm

Very Low Frequency of Mutation Markers

in Atypia/FLUS and Follicular Neoplasm

2% 1%

19%

0%

20%

40%

60%

80%

100%

8%

16% 15%

0% 0% 2% 0% 1% 2%

g

Nikiforov J Clin Endocrinol Metab 2011 96:3390

Mutation present: AUS/FLUS—10% Foll neoplasm—18%

Mutation Panel Analysis High Specificity -High Positive Predictive Value

• Confirm Malignancy

• Not to Avoid Surgery

1Yip J Clin Endocrinol Metab 2012 97:1905

What else is available?

• Protein expression:

– Galectin-3

– proteomics

• Genetic analysis:

– Oncogene mutation panels

• RNA Expression:

– Micro RNA

– Gene expression classifier (GEC - Commercial test)

Rationale for The GEC

• Development of a molecular diagnostic test with high negative predictive value

• When applied to cytologically indeterminate thyroid nodules, a benign result from this test would accurately predict a non-cancerous nodule.

Development of a gene expression classifier (GEC) to accurately identify benign thyroid nodules in patients with indeterminate FNA cytology

A multidimensional algorithm applied to the complex data set

Whole Transcriptome approach using microarray technology

Suspicious

Molecular Classifier Trained and Validated to Distinguish Benign vs.

Suspicious Nodules

Benign

Final chip 167 genes

Gene Expression Classifier

Alexander, et al. N Engl J Med 2012 367:206

Indeterminate Subtypes n=265

AUS / FLUS: Sensitivity: 90% Specificity: 53% *Negative Predictive Value: 95%

n=129 24% CANCER Positive Predictive Value: 38%

• Disagreement as the calculation of negative predictive value included cases Diagnosed by surgical pathology panel as “tumors of uncertain malignant potential”

Gene Expression Classifier

Alexander, et al. N Engl J Med 2012 367:206

Indeterminate Subtypes n=265

AUS / FLUS: Sensitivity: 90% Specificity: 53% Negative Predictive Value: 95%

n=129 24% CANCER Positive Predictive Value: 38%

Follicular Neoplasm: Sensitivity: 90% Specificity: 49%

Negative Predictive Value: 94% Positive Predictive Value: 37%

n=81 24.5% CANCER

Afirma Gene Expression Classifier

Alexander, et al. N Engl J Med 2012 367:206

Indeterminate Subtypes n=265

AUS / FLUS: Sensitivity: 90% Specificity: 53% Negative Predictive Value: 95%

n=129 24% CANCER Positive Predictive Value: 38%

Follicular Neoplasm: Sensitivity: 90% Specificity: 49%

Negative Predictive Value: 94% Positive Predictive Value: 37%

n=81 24.5% CANCER

Suspicious for Malignancy: Sensitivity: 94% Specificity: 52%

Negative Predictive Value: 85%

n=55 62% CANCER

Positive Predictive Value: 76%

Negative predictive Value Falls at higher disease prevalence

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

Prevalence of malignancy

NP

V 95% Sensitivity

90% Sensitivity

80% Sensitivity

Courtesy of Bryan McIver

NPV and PPV for Afirma GEC

based upon N Engl J Med 2012 367:206

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.2 0.4 0.6 0.8 1

NP

V

Prevalence of malignancy

90% Sensitivity and 52% Specificity

NPV

PPV

0

1

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

PP

V

38%

94%

Courtesy of Bryan McIver

Afirma GEC in action—Mayo clinic 53 nodules with GEC results

13 (24%) BENIGN 40 (76%) SUSPICIOUS

27 surgery 13 no surgery

23 (85%) BENIGN 4 (15%) CANCER

McIver ATA abstract 2012

• Different cancer prevalence in FLUS/FN

– NEJM 25% vs. Mayo 15%

– Lower prevalence lower PPV

• Real world application may vary . . .

Afirma in action—Mayo clinic Positive predictive value Mayo 15% vs. NEJM 38%

Afirma in action – UPENN Med Ctr

Positive Predictive Value

General Category Specific Diagnosis

Malignant

(22 cases; 46%)

- FV-PTC: 16 cases (33%)

- Classic PTC: 3 cases (6%)

- Follicular carcinoma: 3 cases (6%)

Benign

(26 cases; 54%)

- Follicular adenoma: 9 cases (19%)

- Hurthle cell adenoma: 14 cases (29%)

- Adenomatoid nodule: 3 cases (6%)

Diagnosis on surgical pathology follow-up of cases with suspicious Afirma® results (n=48)

Afirma in action – UPENN Med Ctr

Positive Predictive Value

Diagnosis on surgical pathology follow-up of cases with suspicious Afirma® results (n=48)

Malignant Benign

AUS/FLUS

(n=18)

FNA x2

11 (61%) - 8 FV-PTC (72%)

- 2 Classic PTC (27%)

- 1 Follicular carcinoma (1%)

6 (39%) - 5 Follicular adenoma (28%)

- 1 Hurthle cell adenoma (6%)

- 1 Adenomatoid nodule (6%)

FN

(n=17)

9 (53%)

- 9 FV-PTC (53%)

8 (47%) - 4 Follicular adenoma (24%)

- 3 Hurthle cell adenoma (18%)

- 1 Adenomatoid nodule (5%)

FNOF

(n=13)

2 (15%)

- 2 Follicular carcinoma (15%)

11 (85%) - 10 Hurthle cell adenoma (77%)

- 1 Adenomatoid nodule (8%)

• Simultaneous parallel sequencing of thousands to millions of short nucleic acid sequences

• High sensitivity and quantitative assessment

• Large scale analyses required for discovery projects

• BUT targeted panels may be useful for molecular dx of thyroid nodules

Next-Generation Sequencing

Next-Generation Sequencing UPMC group--34 primers for amplification of targeted genomic

areas

Nikiforova J Clin Endocrinol Metab 2013;98:E1852-60

Next-Generation Sequencing UPMC group—WHOLE genome/exome/ transcriptome sequencing in

mutation negative 26 PTC and 16 FTC

Nikiforov, unpublished data World Thyroid Congress Toronto 2013

What about combining BRAF testing with the GEC?

0

10

20

30

40

50

60

70

80

90

100

AUS/FLUS Foll Neoplasm Susp for malig

Sen

siti

vity

(%

)

BRAF + GEC Susp BRAF + GEC Susp

No increase in sensitivity because all GEC false negative specimens were also BRAF negative

Kloos J Clin Endocrinol Metab 2013

Evaluation of thyroid nodules 2014

History, physical examination, TSH

Thyroid US with risk stratification for FNA

FNA cytology sample

Cytology •Benign •Malignant •Nondx

Cytology • Follicular lesion of undetermined significance • Follicular/Hürthle cell neoplasm • Suspicious for maligancy

Result reported Molecular analysis

Result reported

Personalized Management of Thyroid Nodules

The Personalized Approach to the Management of Thyroid Nodules -1

Molecular analysis with HIGH NEGATIVE

PREDICTIVE VALUE Gene expression classifier

FLUS/AUS Follicular/Hürthle cell neoplasm Cancer risk <25%

BENIGN

OBSERVE

SUSPICIOUS

LOBECTOMY

Suspicious for malignancy Cancer risk >65-70%

Molecular analysis with HIGH POSITIVE

PREDICTIVE VALUE Mutational analysis

The Personalized Approach to the Management of Thyroid Nodules -2

POSITIVE

THYRX

NEGATIVE

LOBECTOMY

My View About These Tests?

• Marketing different than reality

– Testing all or reflex

– Local cytopathologists in the driver seat vs. a centralized lab reviewing all specimens

– Cost

Conclusions Pathologist in the front seat

Diagnosis of Thyroid Tumors

Multi-modality Approach

Diagnosis of Thyroid Tumors

• Clinical presentation and radiologic findings

• Cytologic diagnosis

– Molecular analysis of selected cases

• Type of Surgery

– Lobectomy vs. Total thyroidectomy

– Lymph node excision (sampling, dissection)

• Management + Molecular analysis

– Radioactive iodine treatment Yes or no

Abu-al Qasim (936-1013 AD) Kitab al-Tasrif

He described thyroid nodules/enlargements as “this tumor, which is called Elephant of the throat, is a large tumor which commonly occurs in women and is of congenital and acquired types. The congenital type is incurable, whereas, the acquired type is of two types: one resembles sebaceous cyst and other as an arterial aneurysm which is dangerous to incise, so never apply knife to it unless the tumor is small.

Thank you for your attention