Assessment of Histological Criteria in the Diagnosis of Mycosis Fungoides

© 2003

The International Society of Dermatology International Journal of Dermatology

2003,

42

, 45–52

45

Abstract

Background

The histologic diagnosis of early mycosis fungoides (MF) can be difficult to

establish in many instances because the subtle changes observed in patches of MF are also

present in many inflammatory dermatoses.

Methods

To assess the frequency and significance of many of these histologic parameters, we

retrospectively reviewed 50 slides from patients with documented MF in patch, plaque, and

tumor stages. The diagnosis of MF was unequivocally established either by the progression of

patients to advanced stages of the disease or by indubitable histologic findings. In the second

phase of the study, we compared the histologic parameters observed in 24 patch stage MF

patients with those in 24 non-MF patients. The non-MF group were patients whose pathologic

pattern was suspicious for MF, but who definitely did not have MF on clinical grounds. The two

groups were matched by histologic pattern. Two different observers evaluated the slides and the

intensities of 32 histologic parameters were graded on a four-point scale to minimize the

subjective variability in the histologic reports.

Results

On univariate analysis, the following parameters achieved significance in

distinguishing MF from non-MF: Pautrier’s microabscesses, haloed lymphocytes, disproportionate

epidermotropism, epidermal lymphocytes larger than dermal lymphocytes, hyperconvoluted

lymphocytes in the epidermis and dermis, absence of dyskeratosis, and papillary dermal fibrosis.

None of these features proved to have additional discriminating power on multivariate analysis.

Conclusions

The efficacy of single histologic features in the diagnosis of early MF is generally

poor and, to discriminate MF from its inflammatory simulators, a combination of cytologic and

architectural features must be used.

Blackwell Science, LtdOxford, UKIJDInternational Journal of Dermatology0011-9059Blackwell Science, 2002?

Mehregan Supplement Report

Histologic criteria in mycosis fungoidesNaraghi et al.Supplement

Assessment of histologic criteria in the diagnosis of mycosis fungoides

Zahra Safee Naraghi,

MD

, Hassan Seirafi,

MD

, Mahin Valikhani,

MD

, Forshad Farnaghi,

MD

, Susan Kavusi,

MD

, and Yahya Dowlati

From Razi Hospital, Tehran University of Medical Sciences, Tehran, Iran

Correspondence

Zahra Safee Naraghi,

MD

Razi HospitalTehran University of Medical SciencesTehranIranE-mail: [email protected]

Introduction

The histologic diagnosis of early stages of mycosis fungoides(MF) may be difficult in many instances: early patches ofMF may show only subtle histologic changes that can easilybe confused with those of inflammatory dermatoses.

1

Inapproximately half of patients, biopsies are diagnostic ofMF from the outset, but, in many patients, histologic findingsare only suggestive of MF and repeat biopsies over time withclinicopathologic correlation will clarify the correct diagnosis.

2,3

Another problem is the low agreement rate in reportingbiopsies suggestive of MF among dermatopathologists due toa lack of specifically defined histologic criteria.

4

Immunophenotyping and T-cell receptor gene rearrangementstudies by polymerase chain reaction have been utilized asadjunctive techniques to provide additional useful informationin difficult cases; however, the immunophenotypic aberrations

seen in advanced stages of MF may not be found in patches,

2,5

and molecular biopsy techniques can only detect T-cellreceptor gene rearrangement in 50–80% of patches andplaques of MF.

1,6–8

A combination of clinical course, histologicfindings, and these adjunctive techniques remains the goldstandard in the diagnosis of MF.

2,3

A number of recent studies have emphasized the importanceof many of the subtle histologic findings as clues to the correctdiagnosis of MF.

1,3,9

Ackerman

10

noted that four histologicpatterns in the context of infiltrates composed mainly oflymphocytes should suggest a diagnosis of patch/plaque ofMF: spongiotic–lichenoid, psoriasiform–lichenoid, spongiotic–psoriasiform–lichenoid, and psoriasiform–nodular. He alsoreported four parameters as diagnostic criteria of early MF:exocytosis with a paucity of spongiosis, solitary lymphocyteslined up along the basal layer, epidermal lymphocytes largerthan dermal lymphocytes, and papillary dermal fibrosis

International Journal of Dermatology

2003,

42

, 45–52 © 2003

The International Society of Dermatology

46 Mehregan Supplement Report

Histologic criteria in mycosis fungoides

Naraghi

et al.

within a lichenoid infiltrate. Shapiro and Pinto

3

observedsparse, superficial, perivascular or interstitial infiltrate oflymphocytes in the early stages of MF, in addition to slightepidermotropism with minimal spongiosis and lymphocytesalong the basal layer of the epidermis. They also emphasizedthe importance of lymphocytic atypia within the epidermis.Smoller

et al

.,

9

in 1995, stated that haloed lymphocytes werethe most robust factor indicative of MF. Other significantparameters included: Pautrier’s microabscesses, dispropor-tionate epidermotropism, epidermal lymphocytes larger thandermal lymphocytes, hyperconvoluted epidermal lymphocytes,and lymphocytes aligned within the basal layer. They concludedthat a combination of specific histologic criteria may beused to establish a diagnosis of MF without the need forimmunophenotyping in the majority of cases.

The European Organization for Research and Treatmentof Cancer (EORTC)

1

recently declared that the only featurewhich yielded independent diagnostic information in MF wasmedium to large cerebriform lymphocytes in the epidermis orin clusters in the dermis, because the major clue to the diag-nosis of MF was the identification of neoplastic cells. Theyalso noted that the efficacy of single histologic criteria in thediagnosis of MF was generally poor and concluded that a con-fident diagnosis of MF could be achieved using a constellationof cytoarchitectural features. In 1999, Guitart

et al

.

4

proposeda grading system based on the sequential evaluation of majorand minor criteria at different magnifications in order toincrease the agreement rate among dermatopathologists inreporting biopsies suspicious of MF. In order to assess thefrequency and significance of different histologic parametersin Iranian MF patients, we conducted our study.

Materials and methods

We conducted our study in two phases. In phase I, we assessed

the frequency of a series of histologic parameters in 50 slides from 30

patients in different stages (patch, plaque, and tumor) of MF. In phase

II, we compared these histologic criteria in 24 patients with documented

patches of MF with those in 24 non-MF patients in order to evaluate

the significance of each parameter in the diagnosis of early MF.

The diagnosis of MF in all patients was unequivocally

established by either the progression of patients from the patch

stage of MF to the more advanced plaque or tumor stage, or by

indubitable histologic findings pathognomonic for MF. All patients

included in the study had medical records at Razi Hospital and

were undergoing treatment for their disease.

Each slide was evaluated separately by our reference

dermatopathologists (Z. S. Naraghi, MD, and S. Kavusi, MD) by

light microscopy, and each histologic parameter was graded on a

four-point scale, as defined in Table 1. In phase II, the slides of MF

and non-MF patients were mixed and reviewed in a blind manner.

We assessed the histologic pattern,

10

density and location of the

dermal infiltrate at low power (

×

10). At intermediate power (

×

40),

all of the cytoarchitectural parameters were evaluated, except for

nuclear detail and atypia of lymphocytes, which were reviewed at

high power (

×

100). Cell counts in the epidermis and dermis (for

example, number of haloed lymphocytes or eosinophils) were

performed in five different fields from the most dense areas and the

mean value was recorded for each. Features such as spongiosis,

dermal edema, parakeratosis, and papillary dermal fibrosis were

graded as the percentage of specimen involved.

9

Disproportionate

epidermotropism was graded according to the ratio of the highest

degree of epidermotropism to spongiosis.

9

Pautrier’s

microabscess was defined as a cluster of at least four atypical

lymphocytes within a single intraepidermal vacuole. The dermal

infiltrate was monomorphous if more than 75% of the infiltrate was

composed of lymphocytes and granulomatous if clusters of

epithelioid histiocytes or giant cells were detected. Lymphocytes

were graded as small (nucleus, < 5

µ

m), medium (nucleus,

5–9

µ

m), or large (nucleus,

≈

10–12

µ

m). Mitosis was present if

more than three mitoses were detected per five high power fields

(> 3/5

×

100

×

). Vasculopathy was defined as the presence of

extravasated red blood cells and/or swollen endothelial cells

and/or thick vessel walls.

Our non-MF group was composed of patients who definitely

did not have MF on clinical grounds and follow-up, but whose

histologic pattern was suspicious for MF. Patients in the MF

and non-MF groups were matched by histologic pattern

10

(i.e.

lichenoid, spongiotic–lichenoid, psoriasiform).

Data were entered on a computer file and statistically analyzed

using the SPSS package to obtain frequency tables and cross

tables. A univariate analysis was performed on each criterion in

phase II by chi-squared and Fischer’s exact test, and the sensitivity

and specificity were calculated for each parameter. A multivariate

analysis was performed on factors that gained significance in the

univariate analysis by logistic regression.

Results

The results of the histologic evaluation of 50 slides from 30patients in patch, plaque, and tumor stages of MF are summa-rized in Table 2.

We reviewed 24 slides from patch stage MF, 18 fromplaque stage MF, and eight from tumor stage MF; 60% of thepatients were males.

11

The patients’ ages in the differentgroups are given in Table 3. There were five cases below30 years of age in patch stage MF, confirming that MF is nota disease exclusively of older individuals, but also affectsyoung patients.

12

The results of the second phase of our study are summarizedin Table 4. The

P

value for each parameter is recorded and thesensitivity and specificity of all criteria are noted in Table 4.

In univariate analysis, we found that the degree ofepidermotropism, Pautrier’s microabscesses, haloedlymphocytes, disproportionate epidermotropism, largerepidermal lymphocytes, hyperconvoluted lymphocytes in the

© 2003


2003,

42

, 45–52

47

Naraghi

et al. Histologic criteria in mycosis fungoides


Table 1

Grading of histologic parameters

Parameter 1+ 2+ 3+ 4+

Atrophy None Mild Moderate SevereAcanthosis None Mild Moderate SevereParakeratosis None < 10% 10–50% > 50%Spongiosis None < 10% 10–50% > 50%Intensity of epidermotropism None 1–5/40

×

6–10/40

×

> 10/40

×

Pautrier’s microabscesses None Present/40

×

– –Haloed lymphocytes None 1–5/40

×

6–10/40

×

> 10/40

×

Disproportionate None Mild Moderate SevereepidermotropismEpidermal > dermal None 10% larger/100

×

10–50% larger/100

×

> 50% larger/100

×

lymphocytesHyperconvoluted None 1–5/100

×

5–10/100

×

> 10/100

×

epidermal lymphocytesDyskeratosis None 1–5/40

×

6–10/40

×

> 10/40

×

Intensity of Sparse Mild perivascular Dense lichenoid Very dense or deepdermal infiltrate or lichenoidEosinophils None 1–5/40

×

6–10/40

×

> 10/40

×

Plasma cells None 1–5/40

×

6–10/40

×

> 10/40

×

Histiocytes None 1–5/40

×

6–10/40

×

> 10/40

×

Hyperconvoluted None 1–5/100

×

6–10/100

×

> 10/100

×

dermal lymphocytesPapillary fibrosis None < 10% 10–50% > 50%Dermal edema None < 10% 10–50% > 50%Telangiectasia None < 1 vessel 1–3 vessels > 3 vesselsPerieccrine lymphocytes None 1–5/40

×

6–10/40

×

> 10/40

×

Follicular mucinosis None < 10% 10–50% > 50%Follicular infiltration None 1–5/40

×

6–10/40

×

> 10/40

×

Vasculopathy None Present – –Mitosis None Present (> 3/500

×

) – –Pattern of Single lymphs Pagetoid Mixedepidermotropism in basal layerBasal layer Intact Focal vacuolar Diffuse vacuolar –

change changePattern of Perivascular Lichenoid – –dermal infiltrateLocation ofdermal infiltrate

Papillary dermis Papillary and upperreticular dermis

Papillary andreticular dermis

Papillary andreticular dermisand subcutis

Dermal infiltrate Monomorph Polymorph Granulomatous –Dermal lymphocytes Small Small, medium Small, medium, largePattern Spongiotic – – –

Lichenoid – – –Psoriasiform – – –Spongiotic–lichenoid – – –Spongiotic–psoriasiform

– – –

Lichenoid–psoriasiform

– – –

Spongiotic–lichenoid–psoriasiform

– – –

epidermis and dermis, absence of dyskeratosis, and papillarydermal fibrosis were the most significant discriminatorsbetween MF and its inflammatory simulators, emphasizingthe importance of both cytologic and architectural features inthe diagnosis of MF.

3

On multivariate analysis by logistic regression, none of thehistologic criteria found to be important diagnostically byunivariate analysis had additional discriminatory power.

In reviewing the slides of 50 MF patients, we encoun-tered all of the major inflammatory patterns described by


2003,

42

, 45–52 © 2003




Naraghi

et al.

Table 2

Frequency of histologic parameters (phase I)

Parameter Patch % > 2+ Plaque % > 2+ Tumor % > 2+ Total % > 2+

Atrophy 41.7 22.2 25 32Acanthosis 66.7 77.8 75 72Parakeratosis 54.2 77.8 87.5 68Spongiosis 54.2 72.2 75 64Pautrier’s microabcesses 37.5 66 12.5 44Haloed lymphocytes 87.5 94.4 10 92Disproportionate epidermotropism 75 83 50 74Epidermal > dermal lymphocytes 41.7 61.1 25 46Hyperconvoluted epidermal lymphocytes 62.5 83.3 75 72Dyskeratosis 29.8 44.4 0 26Intensity dermal infiltrate 87.5 94.4 100 92Eosinophils 29.2 72.2 87.5 54Plasma cells 25 44.4 62.5 38Histiocytes 75 100 87.5 86Hyperconvoluted dermal lymphocytes 83.3 77.8 100 84Papillary dermal fibrosis 95.8 88.9 62.5 88Dermal edema 66.7 61.1 37.5 60Telangiectasia 100 83.3 87.5 92Perieccrine lymphocytes 13 17.6 83.3 22Follicular mucinosis 0 22.2 37.5 14Follicular infiltration 45.8 55.6 87.5 56Vasculopathy 25 50 50 38Mitosis 0 33.3 62.5 22Pattern of epidermotropism

Single basal 79 33 62.5 60Pagetoid 8 28 25 18Mixed 12.5 39 125 22

Basal layerIntact 12.5 22 62.5 24Focal vacuolar 62.5 44 0 46Diffuse vacuolar 25 33 37.5 30

Pattern of dermal infiltratePerivascular 54 39 0 40Lichenoid 46 61 100 60

Location of dermal infiltratePapillary dermis 54 22 0 34Papillary and upper reticular dermis 21 33 0 22Papillary and reticular dermis 25 33 25 28Papillary and reticular dermis 0 11 62.5 14and subcutaneous fatDeep reticular dermis 0 0 12.5 2and subcutaneous fat

Dermal infiltrateMonomorphous 71 55.5 25 58Polymorphous 29 28 75 36Granulomatous 0 16.5 0 6

Dermal lymphocytesSmall 83 67 25 68Small, medium 17 33 62.5 30Small, medium, large 0 0 12.5 2

PatternSpongiotic 4 0 12.5 4Lichenoid 58 28 25 42Psoriasiform 12.5 17 12.5 14Spongiotic–lichenoid 8 28 12.5 16Spongiotic–psoriasiform 4 5 25 8Lichenoid–psoriasiform 12.5 11 0 10Spongiotic–lichenoid–psoriasiform 0 11 12.5 6

© 2003


2003,

42

, 45–52

49

Naraghi



Ackerman.

10

The most common pattern was lichenoid(42%), followed by spongiotic–lichenoid and psoriasiform.Focal or diffuse vacuolar change was detected in 76% of slides.The results of our histologic review are recorded in Table 2.

Discussion

In this study, we systemically evaluated many of the criteriapreviously reported to be useful in distinguishing MF from itsinflammatory simulators.

9

By strictly defining the criteriabefore initiating a comparison, we tried to minimize theinter- and intraobserver disagreement that often accompaniesqualitative and semiquantitative histologic evaluations.

4,13,14

Attempts to enhance diagnostic sensitivity with the use ofimmunophenotyping have proven to be of some value,

5,9

butthe loss of T-cell common antigens useful in confirming thepresence of a neoplastic infiltrate most commonly occurs inplaque and tumor stages, stages in which histopathologicdiagnosis is less challenging.

2

Although some authorities have reported the loss of CD-7to be helpful in identifying MF, others have not found theabsence of CD-7 to be a reliable feature in differentiating MFfrom inflammatory diseases.

2

The polymerase chain reactiontechnique is used to assess T-cell clonality in biopsies suspiciousfor early MF. Clonal T-cell receptor-

γ

gene rearrangementswere detected in 77% of suspected MF specimens in a recentreport by Murphy

et al

.

8

T-cell receptor gene rearrangementscould be detected in 50–80% of documented patches andplaques.

15

Moreover, when indeterminate lesions wereanalyzed, the clonality dropped to as low as 19%.

5

Also, thedemonstration of a clonal T-cell proliferation does notnecessarily establish a diagnosis of a malignant process, suchas MF.

2,9

These results indicate that clinicopathologiccorrelation remains the gold standard in the diagnosis of MF.

In our study, upward migration of lymphocytes into theepidermis (epidermotropism) was found in all MF patients,but also in 80% of inflammatory diseases; it is therefore asensitive, but nonspecific, parameter. Disproportionateepidermotropism was significantly more specific (92%),although not as sensitive. Our results are in agreement withthose of Smoller

et al

.,

9

indicating that epidermotropism is acrucial feature for the diagnosis of early MF. The mostfrequent pattern of epidermotropism in the MF group was thelinear arrangement of single cells in the basal layer (seen in80% of cases), but because lymphocytes were also detected in

the lower one-third of the epidermis in 87.5% of non-MFpatients, this pattern was not a specific marker for MF. Incontrast, a pagetoid pattern was not detected in any of theinflammatory disorders, and was therefore a specific, but notsensitive, feature. Haloed lymphocytes were detected in87.5% of patients with MF, but also in 33% of the controlgroup. Prominent haloed cells (> 6/40

×

) were found in only25% of MF patients. Our study supports the impressions ofAckerman

10

and Smoller

et al

.

9

that haloed lymphocytes arehighly indicative of MF. Pautrier’s microabscesses were onlydetected in patches of MF and never in its inflammatory sim-ulators; therefore, it proved to be a significant discriminatoron univariate analysis, similar to haloed lymphocytes and dis-proportionate epidermotropism, although they were detectedin only 37.5% of patches of MF, similar to that reported bySmoller

et al

.

9

The most specific feature in the epidermis was the presenceof lymphocytes that were larger than dermal lymphocytes, inaddition to Pautrier’s microabscesses. This feature was alsotypical of MF, and not seen in any inflammatory condition.

9

It was readily detectable (epidermal lymphocytes > 10%larger than dermal lymphocytes) in 30% of patches of MF.Hyperconvolution of intraepidermal lymphocytes alsoproved to have discriminating power, although it was not aparticularly common feature (62.5%).

9

We observed areas of atrophy adjacent to acanthotic areasin 12.5% of patches, a histologic clue suggestive of MF.

3

Spongiosis was more common in the non-MF group and, inthose MF patients in which spongiosis was observed, it wasalmost always of a mild degree. Microvesiculation wasabsent in 96% of patches of MF. It is a sensitive criterion andidentical to the results of the EORTC.

1

We should emphasizethat, although mild degrees of spongiosis do not excludeMF, the presence of microvesiculation is exceedingly rare.The absence of dyskeratosis (79%) proved to be anotherdiscriminating factor.

Eosinophils and plasma cells were uncommon in the dermalinfiltrates of patches of MF, but plasma cells were detectedmore frequently in MF patients than in those with inflamma-tory dermatoses. Dermal hyperconvoluted lymphocytes werecommonly detected in patches of MF (83%), but not in thecontrol group, proving to be a specific discriminating factor inthe univariate analysis. This observation is emphasized by thereport of the EORTC

1

which declared that the presence ofmedium to large cerebriform cells in the epidermis or in clus-ters in the dermis is the single most useful histopathologicmarker in establishing a diagnosis of early MF.

Dermal infiltration was more dense in MF patients. In 96%of patches, the infiltrate was composed mainly of smalllymphocytes, but medium lymphocytes were found in 16%of patients, usually mixed with small cells; 71% showed amonomorphic dermal infiltrate. We obtained a result betweenthat reported by Nickoloff,

16

who observed a monomorphous

Table 3 Age of patients

Age Patch Plaque Tumor Total

Range (years) 17–77 30–78 42–75 17–78Mean (years) 47.2 57.2 58.7 52.7


2003,

42

, 45–52 © 2003




Naraghi

et al.

Table 4

Univariate analysis

ParameterPatch MF > 2+(%)

Non-MF > 2+(%)

Sensitivity(%)

Specificity(%)

P

value

Atrophy 41.7 50 42 50 0.77Acanthosis 66.7 50 67 50 0.38Parakeratosis 54.2 33.3 54 67 0.24Spongiosis 54.2 75 54 25 0.22Pautrier’s microabscesses 37.5 0 37.5 100 0.002Haloed lymphocytes 87.5 33.3 87.5 67 0Disproportionate epidermotropism 75 8.3 75 92 0Epidermal > dermal lymphocytes 41 0 41 100 0.001Hyperconvoluted epidermal lymphocytes 62.5 25 62.5 75 0.019Dyskeratosis 20.8 70.8 21 29 0.001Intensity dermal infiltrate 87.5 75 87.5 25 0.46Eosinophils 29.2 29.2 29 71 1Plasma cells 25 12.5 25 87.5 0.46Histiocytes 75 91.7 75 8 0.24Hyperconvoluted dermal lymphocytes 83.3 12.5 83 87.5 0Papillary fibrosis 95.8 50 96 50 0.001Dermal edema 66.7 87.5 67 12.5 0.168Telangiectasia 100 91.7 100 8 0.489Perieccrine lymphocytes 13 15 13 85 0.712Follicular mucinosis 0 0 0 – –Follicular infiltration 45.8 63.6 46 36 0.14Mitosis 0 0 0 – –Pattern of epidermotropism

Single basal 79 87.5 79 12.5 0.35Pagetoid 8 0 8 100Mixed 12.5 12.5 12.5 87.5

Basal layerIntact 12.5 20.8 12.5 79 0.067Focal vacuolar 62.5 29.1 62.5 71Diffuse vacuolar 25 50 25 50

Pattern of dermal infiltratePerivascular 54 58 54 42 1Lichenoid 46 42 46 58

Location of dermal infiltratePapillary dermis 54 37.5 54 62.5 0.49Papillary and upper reticular dermis 21 25 21 75Papillary and reticular dermis 25 33 25 67Papillary and reticular dermis and subcutis 0 4 0 96

Dermal infiltrateMonomorphous 71 62.5 71 37.5 0.76Polymorphous 29 37.5 29 62.5Granulomatous 0 0 0 –

Dermal lymphocytesSmall 83 100 83 0 0.113Small, medium 12.5 0 12.5 100Medium 4 0 4 100

PatternSpongiotic 4 4 1Lichenoid 58 58Psoriasiform 12.5 12.5Spongiotic–lichenoid 8 8Spongiotic–psoriasiform 4 4Lichenoid–psoriasiform 12.5 12.5Spongiotic–lichenoid–psoriasiform 0 0

© 2003


2003,

42

, 45–52

51

Naraghi



infiltrate in 33% of patches and plaques, and that obtained bythe EORTC,1 which detected a monomorphous infiltrate oflymphocytes to be a sensitive criterion (91.7%).

In 75% of patch stage MF, dermal infiltration was confinedto the papillary and upper reticular dermis. No involvementof the subcutis was observed in our MF patients; 37.5% ofpatches showed a purely lichenoid pattern and 20% showedan atrophic lichenoid pattern at low magnification. Othercommonly observed gross patterns in decreasing order werepsoriasiform (12.5%) and lichenoid–psoriasiform (12.5%).Shapiro and Pinto3 reported similar findings. This emphasizesthe common occurrence of interface changes in MF patients,2,17

found in 87.5% of our cases of MF.Dermal edema and vasculopathy were neither sensitive

nor specific for MF. Severe dermal edema was not observedin the MF group. In contrast, papillary dermal fibrosis was asensitive feature (96%), with the observation of either fine orcoarse bundles of collagen, and it achieved statistical signifi-cance as a discriminating factor. This is similar to the resultsreported by Smoller et al.9 and Ackerman.10 They alsopointed out that dermal fibrosis was a feature of late atrophicpatches or plaques and was not encountered in early patches.Mitoses were not detected in patches of MF, showing thatlymphocytic mitotic activity is not increased in early lesions ofMF.9 In addition, none of our MF patients had any evidenceof follicular mucinosis, but folliculotropism was detected inabout one-half of patients. Eccrine gland involvement wasreported in 13% of patch stage MF.

Pautrier’s microabscesses, haloed lymphocytes, dispropor-tionate epidermotropism, larger epidermal lymphocytes,hyperconvoluted lymphocytes in the dermis and epidermis,absence of dyskeratosis, and papillary dermal fibrosis achievedsignificance as discriminating factors between MF and itsinflammatory simulators on univariate analysis; however,none of these features showed additional distinguishing poweron multivariate analysis by logistic regression. It is counter-intuitive that the diagnosis produced by multiple histologicfeatures does not demonstrate a unique criterion for discrim-inating MF from non-MF. There may be two reasons for this.First, covariation of our significant features (on univariateanalysis) and their high correlation may exhaust the discrim-inating power of the other. Second, the relatively small samplesize; with a larger sample size, some features might emerge assignificant factors.

Based on the data presented here, we conclude that thereare some histologic features, such as Pautrier’s microabscesses,haloed lymphocytes, disproportionate epidermotropism,larger epidermal lymphocytes, absence of dyskeratosis,hyperconvoluted dermal and epidermal lymphocytes, andpapillary dermal fibrosis, that can be used to discriminate MFfrom its inflammatory simulators. In order to diagnose MFconfidently, we must use the constellation of cytologic andarchitectural parameters proposed. The gold standard in the

diagnosis of MF is clinicopathologic correlation, and theefficacy of single histopathologic features in the diagnosis ofearly MF is generally poor.1

References

1 Santucci M, Biggeri A, Feller AC, Burg G. Efficacy of histologic criteria for diagnosing early mycosis fungoides. Am J Surg Pathol 2000; 24 (1): 40–50.

2 Glusac EJ, Shapiro PE, Mcniff M. Cutaneous T-cell lymphoma: Refinement in the application of controversial histologic criteria. Dermatol Clin 1999; 17 (3): 601–614.

3 Shapiro PE, Pinto FJ. The histologic spectrum of mycosis fungoides/Sezary syndrome. A review of 222 biopsies including newly described patterns and the earliest pathologic changes. Am J Surg Pathol 1994; 18 (7): 645–667.

4 Guitart J, Variakojis D, et al. Histologic criteria for the diagnosis of early mycosis fungoides: Proposal of a grading system to standardize the pathology report (Abstract). Am J Dermatol 1999; 21 (1): 76.

5 Bergman R, Faclieru D, Sahar D, et al. Immunophenotyping and T-cell receptor gamma gene rearrangement analysis as an adjunct to the histopathologic diagnosis of mycosis fungoides. J Am Acad Dermatol 1998; 39: 554–559.

6 Ashton-Key M, Diss TC, Du MQ. The value of polymerase chain reaction in the diagnosis of cutaneous T-cell infiltrates. Am J Surg Pathol 1997; 21 (7): 743–757.

7 Delfau-Larue MH, Patrella T, Lahet C, et al. Value of clonality studies of cutaneous T-lymphocytes in the diagnosis and follow-up of patients with mycosis fungoides. Fungoides J Pathol 1998; 184 (2): 185–190.

8 Murphy M, Signoretti S, Kadin ME, et al. Detection of TCR-gamma gene rearrangement in early mycosis fungoides by nonradioactive PCR-SSCP. J Cutan Pathol 2000; 27 (5): 228–234.

9 Smoller BR, Bishop K, Glusac E, et al. Reassessment of histologic parameters in the diagnosis of mycosis fungoides. Am J Surg Pathol 1995; 19 (12): 1423–1430.

10 Ackerman AB. Histologic Diagnosis of Inflammatory Skin Diseases: An Algorithmic Method Based on Pattern Analysis. Baltimore: Williams & Wilkins, 1997.

11 Morales Suarez-Varela MM, Llopis Gonzalez A, Marquina Vila A, et al. Mycosis fungoides: review of epidermologic observations. Dermatology 2000; 201 (1): 21–28.

12 Elder D. Lever’s Histopathology of the Skin. Philadelphia: Lippincott-Raven, 1997.

13 Burg G, Zwingers T, Feller AC, et al. Interrater and intrarater variabilities in the evaluation of cutaneous lymphoproliferative T-cell infiltrates. Dermatol Clin 1994; 12: 311–314.

14 Santucci M, Biggeri A, Feller AC, et al. Accuracy, concordance, and reproducibility of histologic diagnosis in cutaneous T-cell lymphoma: an EORTC cutaneous lymphoma project group study. Arch Dermatol 2000; 136 (4): 497–502.

International Journal of Dermatology 2003, 42, 45–52 © 2003 The International Society of Dermatology

52 Mehregan Supplement Report Histologic criteria in mycosis fungoides Naraghi et al.

15 Tok J, Szabolcs J, Silvers DN, et al. Detection of clonal T-cell receptor & chain gene rearrangements by PCR and denaturing gradient gel electrophoresis (PCR/DGGE) in archival specimens from patients with early CTCL. Correlation of histologic findings with PCR/DGGE. J Am Acad Dermatol 1998; 38: 453.

16 Nickoloff BJ. Light microscopic assessment of 100 patients with patch/plaque stage mycosis fungoides. Am J Dermatopathol 1988; 10: 469–477.

17 Guitart J, Peduto M, Caro WA, et al. Lichenoid changes in mycosis fungoides. J Am Acad Dermatol 1997; 36 (3): 417–422.

Documents

Assessment of Histological Criteria in the Diagnosis of Mycosis Fungoides