Best Practices in Diagnostic Immunohistochemistry: Workup of Cutaneous Lymphoid Lesions in the Diagnosis of Primary Cutaneous Lymphoma

Clinical and morphologic criteria play a very important role in the subclassification of cutaneous lymphoid lesions.⁶  The presence or absence of multiple sites of involvement, antecedent history, peripheral blood involvement, tempo and pattern of progression (aggressive or chronic), and association with other local or systemic diseases (contact dermatitides or autoimmune disease such as lupus) are important considerations not only for the diagnosis but also to distinguish the lesion from pseudolymphoma (reactive lymphoid proliferations that very closely mimic lymphoma). In addition, many types of cutaneous lymphoma can have similar histomorphologic features—for example, some tumoral lesions of mycosis fungoides (MF) with large cell transformation cannot be reliably distinguished from anaplastic large cell lymphoma or lymphomatoid papulosis by hematoxylin-eosin sections or even IHC. Thus, clinical correlation is very important in the accurate diagnosis and classification of cutaneous lymphoma, much more so than in the diagnosis of other extranodal or nodal lymphomas.

For optimal morphologic assessment, sections should be 4 to 5 μm. A low-magnification examination should be first performed to assess the architecture of the infiltrate (superficial, superficial and deep, deep or subcutaneous), distribution (perivascular, interstitial, diffuse or nodular), and involvement of epithelial structures (epidermotropism, appendageal structures—eg, pilotropism or syringotropism). This examination should be followed by an intermediate-magnification assessment to look for cellular composition (monomorphous versus mixed population; presence of other inflammatory cells) and other features (eg, presence of mucin, angioinvasion, destruction). A high-magnification assessment of the cellular morphology is then performed to evaluate features such as cellular shape, cytoplasm, nuclear-cytoplasmic ratio, nuclear contours, and nuclear chromatin.

The above morphologic clues will provide the basic framework for the selection of appropriate antibody panels. For example, a superficial diffuse dermal infiltrate with extensive epidermotropism and epidermal “microabscesses” is most likely to be a T-cell lymphoid proliferation. In contrast, a deep or superficial and deep dermal infiltrate with a nodular pattern is likely to be a B-cell lymphoid proliferation. Nonetheless, at times when this distinction is difficult, a screening panel should be used; this panel could include B-cell markers such as CD20, T-cell markers such as CD3, and CD30 (to identify CD30⁺ lymphoproliferative disorders), and may be useful to provide an initial indication as to lymphocyte lineage. An overview of T-cell marker utilization is given below (also see Table 2); B-cell markers will be described within individual entities (also see Table 3).

T-cell abnormalities are best characterized by immunohistochemical stains for T-cell–specific antigens (Table 2). Normal T cells show expression of pan–T-cell antigens (most commonly CD2, CD3, CD5, and CD7). Precursor stage differentiation within the thymus will give rise to CD4 (also called helper) and CD8 (also called cytotoxic) subset of T cells. All T cells express either CD4 or CD8 except for rare γδ cells that lack both CD4 and CD8. Loss of any of the pan–T cell antigens, as well as coexpression of both CD4/CD8 (so-called double positive) or lack of both CD4/CD8 (so-called double negative), are considered abnormal, indicating a lymphoproliferative process, but may be seen rarely in reactive conditions, particularly autoimmune disorders such as lupus⁷  or rheumatoid arthritis.⁸  Along with these antigenic markers, T cells also undergo rearrangements of the T-cell receptor genes, resulting in unique cell surface receptors (T-cell receptors; TCRs). T-cell receptors are of 2 types. Most circulating T cells (representing 95% or greater of peripheral blood T cells) and tissue T cells express αβ type, with the remainder belonging to the γδ T-cell subtype. The immunohistochemical profile of an αβ T-cell and a γδ T-cell is distinguished most often by a single antigen—recognized by antibody clone βF1. This antibody clone βF1 maps to an epitope within the constant region of the β chain of T-cell receptor, present only within αβ T cells, not seen in γδ T cells, and can be identified in paraffin sections by routine IHC.⁹  Newer antibodies that cover the γ and δ epitopes have been developed in recent years, but are not widely used yet. For example, an antibody to the TCR δ epitope has been described recently,¹⁰  but the experience with it is relatively limited. T-cell receptor γ (TCRG or γ-M1) is available commercially.^11,12 

Mycosis fungoides is the most common cutaneous T-cell lymphoma.¹³  It has 3 distinct clinical stages (patch, plaque, and tumor). Most cases of typical MF are chronic, slowly evolving lesions with a long-standing course. Lymphomas with similar histomorphology but aggressive clinical course are usually classified in other categories (eg, cytotoxic T-cell/natural killer [NK] cell lymphoma). Mycosis fungoides is characterized by a proliferation of small to medium-sized pleomorphic lymphocytes and often by prominent epidermotropism. The nuclear contours of the lymphocytes are convoluted and can be appreciated in MF cells circulating in the peripheral blood (sometimes referred to as “cerebriform”). Collections of intraepidermal lymphocytes termed Pautrier microabscesses (originally described by Pautrier as Darier nests) are considered characteristic (Figure 1, A) but are seen only in a minority of lesions, usually patch and plaque stage. The epidermotropic lymphocytes have slightly enlarged nuclei compared to the intradermal lymphocytes and may show a convoluted nuclear contour (Figure 1, B). These cells are approximately the size of the nucleus of an epithelial cell. The presence of lymphocytes aligned along the basal layer of the epidermis is a useful histologic feature to recognize. In advanced-tumor-stage lesions the epidermotropism and the Pautrier microabsceses are often absent. Interstitial fibrosis, most commonly seen in the upper dermis as bands of fibrosis interrupted by abnormal lymphocytes (“fettuccine fibrosis”), is a common finding in MF. Other histologic features may include psoriasiform epidermal hyperplasia and presence of intraepidermal mucin sometimes mimicking an inflammatory process. There is usually little or no epidermal spongiosis. In the cases with spongiosis there is a disproportionate amount of lymphocytic epidermotropism relative to the degree of spongiosis.

Some cases of MF may not need immunohistochemical characterization, if the histomorphology and clinical features are characteristic. In cases with less definitive histology, and if the clinical history is not reliable or not consistent with MF, immunohistochemical studies may be performed. Mycosis fungoides is characterized by T cells with αβ T-helper memory phenotype. Their typical immunohistochemical profile is CD3⁺, βF1⁺, CD4⁺, CD5⁺ and CD8⁻. Often there is aberrant loss of CD7. An initial IHC profile demonstrating a CD3⁺ lymphocytic population (Figure 1, C) with an increased CD4:CD8 ratio is helpful in establishing a diagnosis.^14,15  A high epidermal CD4:CD8 ratio (averaging 3.5) has been associated with MF. The ratio was lower in atypical (1.6) and benign (1) cases in this study. Double-positive and double-negative (CD4 and CD8) MFs have been described.^16,17  CD8⁺ MFs tend to occur in pediatric cases and in hypopigmented variant lesions, and are prognostically similar to the usual CD4⁺ MFs.^18,19  Mycosis fungoides in plaque or tumor stage or with large cell transformation may have increased numbers of CD30⁺ cells. Cytotoxic markers such as TIA-1 may show positivity in advanced stages. Very rarely, these T cells may aberrantly express CD20 or CD79a,²⁰  but typically do not express other B-cell antigens such as PAX5.

Sezary syndrome is a rare disease characterized by erythroderma, lymphadenopathy, and a neoplastic T-cell population involving skin, lymph nodes, and peripheral blood. Cutaneous lesions of Sezary syndrome share many of the histologic features of mycosis fungoides, and in the United States, MF and Sezary syndrome are often combined together into 1 entity. The WHO 2008 classification distinguishes these as 2 separate entities. In the EORTC scheme, peripheral blood involvement in a patient with MF is not distinguished as Sezary syndrome, even though some pathologists preferentially note this as “Sezary syndrome preceded by MF,” and distinct chromosomal changes have been observed in these cases. The primary distinguishing feature of Sezary syndrome is the presence of blood involvement; however, the diagnosis (which leads to treatment decisions and further defines clinical end points) of Sezary syndrome requires a set of specific criteria,^21,22  as listed in Table 4. Peripheral blood often shows medium-sized Sezary cells, with convoluted cerebriform nuclei, similar to those seen in tissues. The tumor cells have a CD3⁺, CD4⁺, and CD8⁻ phenotype similar to MF. Often, these peripheral blood cells show loss of T-cell antigens, most commonly CD7 and CD26²³  (demonstrable by flow cytometry), but may also show loss of CD2, CD3, CD4, and CD5. Expression of the antigen PD-1 (programmed death–1) is reported as a reliable marker of Sezary cells.²⁴  Of note, PD-1 is a marker of T-helper cells of follicle origin and is seen in other T-cell lymphomas such as angioimmunoblastic T-cell lymphoma and a subset of peripheral T-cell lymphoma, not otherwise specified.²⁵ 

Primary cutaneous CD30⁺ T-cell lymphoproliferative disorders^26,27  are among the commonest cutaneous T-cell lymphomas (second to MF), accounting for about 30% of all cases and encompassing 3 diseases: lymphomatoid papulosis (LyP), cutaneous anaplastic large cell lymphoma (C-ALCL), and borderline cases. Despite demonstrable clonality, LyP is not considered a fully malignant disorder, whereas C-ALCL is regarded as bona fide T-cell lymphoma.²⁸  The borderline cases share clinical and morphologic features with the 2 distinct entities.²⁹  Of note, clinical findings are very important in distinguishing between C-ALCL and LyP, since morphology, immunophenotype, and clonality studies alone are not enough to distinguish between these 2 entities, in some cases.

The histology of C-ALCL often shows a diffuse proliferation of medium to large “anaplastic” cells, usually sparing the epidermis. The neoplastic infiltrate is typically composed of large cells that show variable pleomorphism, sometimes interspersed with mononuclear Hodgkin-like cells (Figure 2, A). The nuclear contours are often described as “wreathlike”; the medium-sized anaplastic cells that have abundant, clear, eosinophilic or amphophilic cytoplasm, with a kidney-shaped nucleus and an eosinophilic paranuclear region, are termed hallmark cells. The infiltrate may be dominated by neutrophils or may show panniculitis-like changes, especially in immunosuppressed patients.^30,31  The lesional cells are by definition CD30⁺, which is expressed by a majority (>75%) of the neoplastic infiltrate (Figure 2, B). They are usually CD3⁺, CD4⁺, and CD8⁻. Unlike Hodgkin lymphoma cells, they are CD15⁻ and do not express PAX5.

Lymphomatoid papulosis shares the exact antigenic profile of C-ALCL, with CD4⁺ CD8⁻ expression and also expression of CD30. Three common histologic patterns have been classically described (types A, B, and C) with a fourth one (type D) being added recently.²⁹  Of the three, type B resembles MF and can show extensive epidermotropism. Of note, in type B LyP, the cells are frequently CD30⁻. Rare cases of both C-ALCL and LyP can express CD56 and other NK cell markers. The fourth pattern characteristically is of a CD8 phenotype with pagetoid changes.^32,33  About a third of LyPs or CD30⁺ lymphoproliferative disorders express cytotoxic markers such as TIA-1.³⁴ 

In the most recent classification, subcutaneous panniculitis-like T-cell lymphoma is recognized as a rare form of primary cutaneous T-cell lymphoma with distinct morphologic and immunohistochemical features. Subcutaneous panniculitis-like T-cell lymphoma (SPTCL) preferentially involves the subcutaneous fat in a lobular pattern. Fat necrosis, karyorrhexis, and admixed histiocytes with vacuolated cytoplasm, containing apoptotic debris and erythrocytes, are commonly seen (Figure 3, A), and this may superficially mimic other reactive conditions, such as lupus.^35,36  The lymphocytes are small to medium, have irregular nuclei, and often show rimming around individual fat cells (Figure 3, B). In the WHO 2008 and EORTC classification, this entity is restricted only to the αβ T cells. The characteristic immunohistochemical profile³⁷  is CD3⁺, CD8⁺ (Figure 3, C and D), with expression of cytotoxic markers (including granzyme-B, TIA-1, and perforin) and of α/β receptor (Figure 3, E). CD56 usually shows negativity. Those lesions with predominantly γδ T cells (βF1⁻) are considered an entity separate from SCPTCL owing to their significantly worse prognosis and are classified as cutaneous γδ T-cell lymphoma (see below).

These lymphoproliferative lesions with a γδ phenotype are clinically aggressive³⁸  (median survival of ∼15 months) and consist of activated cytotoxic T cells with variable morphology. The variability in morphology is described within 3 major histologic patterns—primarily epidermotropic, dermal, and subcutaneous.¹⁰  The neoplastic cells are medium to large with elongated or folded nuclei and show frequent mitotic activity. The tumor cells are T cells that express CD3 (Figure 4, A) and CD2. They show loss of CD5 (Figure 4, B) or CD7. The cytotoxic proteins (TIA-1 and granzyme) are expressed, as is CD56 (Figure 4, C). The cells can show loss of CD4 (Figure 4, D) and CD8 (Figure 4, E). A subset of cases is CD8⁺ but only rare cases express CD4. CD30 and Epstein-Barr encoded RNA expression may be seen in rare cases. The cells are of the TCR-γδ phenotype, by definition. When available, this phenotype can be demonstrated by IHC staining for TCR-δ or TCR-γ (Figure 4, F) with appropriate detection methods. However, in routine practice, so far, absence of βF1 expression (Figure 4, G) is considered an inferential surrogate for the γδ phenotype.

Also known as Berti lymphoma, this is an aggressive T-cell lymphoma with a cytotoxic phenotype.^39,40  The histology and neoplastic cytomorphology are frequently indistinguishable from other epidermotropic T-cell lymphoproliferative disorders such as CD8⁺ variants of MF, pagetoid variants of cutaneous γδ T-cell lymphoma, or type D LyP.^10,32  Distinction from other CD8⁺ T-cell lymphomas with epidermotropism is based mostly on clinical presentation. Unlike conventional MF, the cells are CD8⁺ and unlike γδ T-cell lymphoma, the neoplastic cells express βF1, indicating an αβ-TCR phenotype. Thus, the neoplastic cells of this entity are CD3⁺, βF1⁺, CD7⁺, CD8⁺ with cytotoxic proteins (TIA-1 and granzyme). CD7 is usually preserved among the T cells. Among the pan–T-cell antigens there is variable loss of CD2, and, more frequently, loss of CD5 may be seen. The median survival of patients, as originally described in the Berti series,⁴⁰  is approximately 32 months.

This is a new and provisional entity in the WHO-EORTC.⁵  It is usually a solitary lesion, reported mostly in the head and neck region. Most of the lesion is composed of small to intermediate atypical cells (Figure 5, A and B) that are positive for CD4 and lack CD8 or other cytotoxic markers.¹⁰  Antigenic loss or aberrancy is only rare. More recently, PD-1 expression, a marker of follicular T-helper cells, has been found in this entity.^25,41  A population of small reactive lymphocytes, histiocytes, and occasional eosinophils is usually noted. Since there are no major distinguishing features of this infiltrate from benign or reactive lymphoid population outside of cytologic features, clonality studies are usually helpful and should yield positive findings. The localized solitary lesions have a very good prognosis, with a 5-year survival rate of more than 75%. The correct classification of this entity as a true lymphoma or as a pseudolymphoma is still disputed.

WHO-EORTC classification includes 4 distinct T- and NK-cell lymphomas that involve the skin as a secondary site. These are extranodal NK-/T-cell lymphoma, nasal type, hydroa vacciniforme-like lymphoma, adult T-cell leukemia/lymphoma, and angioimmunoblastic lymphoma.¹⁰  Since these are rare entities, or not primarily cutaneous, they are not described further in this review.

Cutaneous marginal zone lymphoma^42,43  is a low-grade B-cell lymphoma composed of nodular or diffuse proliferation of small to medium-sized cells. These cells are classically described as monocytoid lymphocytes and may show conspicuous lymphoplasmacytic differentiation or be admixed with terminally differentiated plasmacytic cells (Figure 6, A). In practice, many lesions consist of small lymphoid cell infiltrate with no specific morphology, indicating a plasmacytic differentiation. Germinal centers are occasionally noted in the lesion, but they may be absent, less conspicuous, distorted, atrophic, or commonly, completely absent. The B cells are typically positive for CD20 (Figure 6, B), CD79a, and PAX5 (the B-cell markers may have variable intensities) and usually lack expression of CD5 and CD10. BCL2 expression is present, and the Ki-67 fraction is not very high in most of the lesions. CD21 stain is useful in highlighting the follicular dendritic cells of lymphoid follicles present in the lesion. The CD21 stains usually show distortion and dissolution of the dendritic cell meshwork, which supports a diagnosis of lymphoma. Follicular colonization is a distinctive feature observed in marginal zone lymphoma, appreciated with many of the immunohistochemical stains used. The presence of follicle center cells (BCL6 positive, BCL2 negative), which are infiltrated by the neoplastic marginal zone cells (BCL6 negative, BCL2 positive), is a useful clue for the diagnosis of marginal zone lymphoma. This impression may also be seen in immunostains for Ki-67, where the highly proliferative germinal center cells are infiltrated by neoplastic cells with a low proliferation index. A population of reactive CD3⁺ T cells is often noted. PD-1 is conspicuously negative.

Demonstration of light-chain expression is a very useful adjunct in evaluation of B-cell neoplasms (see also Table 3). κ and λ light-chain restriction by protein IHC is difficult to demonstrate, even in lesions that have a predominance of plasma cells; both κ/λ IHC (Figure 6, C and D) and in situ hybridization may be attempted to demonstrate proof of clonal restriction. If clear light-chain monotypia is present, then this supports a clonal process. However, absent or equivocal light-chain restriction does not exclude a diagnosis of lymphoma. Since a variety of reactive lymphoid proliferations can mimic lymphoma,⁴⁴  immunoglobulin (Ig) H clonality studies may need to be performed if light-chain restriction studies are not definitive. In a small subset of cases, Borrelia burgdorferi infections are associated as an antigenic stimuli of cutaneous B-cell lymphoma of the marginal zone type, particularly in European populations, but are neither necessary nor easy to demonstrate to make the diagnosis.^45–47  When demonstrated (usually by polymerase chain reaction), antibiotic therapy may be attempted, which is associated with cure of only a small subset of cases.⁴⁸  The t(11;18), which is usually seen in nodal marginal zone lymphomas, is not present in most cases of cutaneous marginal zone lymphoma.

Primary cutaneous follicle center lymphoma (PCFCL) differs significantly from systemic or nodal follicular lymphoma, which partially explains the challenges in the diagnosis.^5,42  The typical case of PCFCL is composed of small to intermediate-sized cells with classically cleaved nuclear contours (centrocytes) admixed with centroblasts. The similarity with nodal follicular lymphoma may end there. The presence of back-to-back follicles or nodules (as seen in the upper dermis of Figure 7, A) is not seen in many cases of PCFCL, where there may be confluence of nodules (bottom half of the biopsy specimen in Figure 7, A) or just diffuse effacement with B cells (Figure 7, B) with centrocyte morphology. BCL2 expression, classically seen and pathognomonic of most nodal follicular lymphomas, is lacking in most cases of PCFCL. Diagnosis of PCFCL requires the demonstration of germinal center–associated markers within the neoplastic B cells. While CD10 is a commonly used marker and shows positivity in a subset of cases⁴⁹  with a follicular pattern, it may be lost in a good portion of PCFLCs and those with a diffuse growth. Other markers, including BCL6, human germinal center–associated lymphoma (HGAL also known as GCET2) antigen,⁵⁰  or GCET1, may be used as alternatives.

In summary, most PCFCLs are monoclonal B cells (typically demonstrated by IgH clonality studies) usually expressing BCL6 (Figure 7, C) but showing variable CD10 expression. BCL2 usually shows negativity or, when present, is faint (or in a subset of cells).⁴²  MUM-1 is not expressed in PCFCL. IGH/BCL2 translocation will be negative in most (59%–90%) cases.⁴² 

The absence of BCL2 in most PCFCLs does not have prognostic implications. However, excluding secondary involvement of skin by systemic/nodal follicular lymphoma should be done in all cases with positive BCL2 expression, since this is more common.

This is an aggressive B-cell lymphoma that is usually, but not always, present in the lower limbs; it is more common in females and can be bilateral.^5,42  The lesion is typically composed of a monomorphic population of large centroblasts and immunoblasts with few admixed cells. Immunohistochemistry shows strong BCL2 expression. BCL6 usually shows positivity but CD10 is variably expressed. MUM-1/IRF4 shows positivity (activated B-cell—ABC profile). The strong BCL2 and MUM-1 expression distinguishes diffuse large B-cell lymphoma (DLBCL), leg type, from follicle center cell lymphoma (follicular lymphoma) with large cell transformation, in most cases.⁴²  IgM and HGAL as markers to aid in the differentiation between these 2 entities have been reported.^50–52  FOXP1 is shown to be expressed in most DLBCLs, leg type.⁵³ 

Most cases of primary cutaneous DLBCLs are either follicle center with diffuse/large cell transformed type, or of the leg type.⁴²  However, the “other” category consists of rarer lesions that can be subcategorized into well-known morphologic entities: T-cell/histiocyte-rich large B-cell lymphoma (TC/HR LBCL), plasmablastic lymphoma, and intravascular lymphoma. The TC/HR LBCL of the skin has a better prognosis compared to its nodal/noncutaneous counterparts. Complete clinical staging is of paramount importance, to distinguish between primary cutaneous and systemic disease.

This new addition to the WHO-EORTC classification scheme is defined by immunohistochemical expression of CD4 and CD56.⁵⁴  This lesion used to be known as blastic NK-cell lymphoma and CD4⁺/CD56⁺ hematodermic neoplasm. The neoplastic cells express CD4, CD43, and CD56. In addition, the plasmacytoid dendritic cell–associated antigens CD123 (interleukin 3α chain receptor), BDCA-2/CD303, TCL1, and CLA, and the interferon α–dependent molecule MxA, are also expressed. CD4, CD56, CD43, and CD123 are commonly used immunostains.

Both T- and B-cell clonality studies form an essential ancillary test⁵⁵  in diagnosis of cutaneous lymphoma for the following reasons: many of the primary cutaneous lymphomas do not have or fail to demonstrate antigenic aberrancies, compared to their nodal or extracutaneous counterparts. Exuberant cutaneous B-cell hyperplasia can mimic cutaneous lymphoma (marginal zone lymphoma), and demonstrating clonality is essential to support a diagnosis of lymphoma. Both B- and T-cell clonality studies, as done currently by using BIOMED-II validated primers against targets such as IGH, IGK, IGL, TRG, and TRB, have several rounds of redundancies and hence are sensitive for capturing most of the cutaneous lymphomas (up to 90%). Two important caveats should be remembered when ordering antigen receptor gene rearrangement studies for cutaneous lymphoid lesions. One, gene rearrangement studies are never a substitute for a thorough morphologic and immunophenotypic evaluation. In fact, clonality studies should be emphasized as valuable “ancillary” studies and should always be interpreted in conjunction with clinical, morphologic, and immunohistochemical data. In those cases where clinical and morphologic features were suggesting a benign lesion and a gene rearrangement study result is positive, a note indicating this discrepancy should be added to the diagnosis. Watchful waiting and clinical follow-up may be indicated in many cases, instead of reclassifying the lesion as a lymphoma solely on the basis of positive clonality study findings alone. A direct conversation with a clinician is appropriate in these circumstances, which may lead to reevaluation of the clinical findings. This is particularly true in the interpretation of TCR clonality studies, which may have a high degree of positivity based on interpretative thresholds and techniques (primer sets) used. The TCR gene rearrangement studies can be positive in benign inflammatory diseases too, but when multiplex polymerase chain reaction primers are used, the false-positive rate is typically low (2.3% in one series).⁵⁶ 

Secondly, antigen receptor gene rearrangement studies should never be used as a surrogate to make a definitive diagnosis when biopsied tissue is limited and should probably not be ordered in the absence of a lymphoid infiltrate. Smaller lymphoid infiltrates, especially a partially sectioned (focal) germinal center, may result in pseudoclonal patterns in electropherograms, owing to limited IGH repertoire amplification or an interpretation bias due to the lack of a good “Gaussian” distribution of reactive lymphocytes in the background. When only a very limited number of B or T cells are present in the sample, these studies may selectively amplify a single cell that might simulate a clone.^57,58  Thus, when sample is limited, instead of requesting gene rearrangement studies, a repeated biopsy is often of better yield to establish an accurate diagnosis.

Because skin biopsies are inherently small, which can result in the abovementioned artifacts, all gene rearrangement studies in this setting should be viewed cautiously owing to the possibility of false-positive results. To reduce false-positive results due to selective amplification in a truly reactive process, repeated studies should be performed on the same sample and the clonal population should be of identical size on the electropherograms (Figure 8). Oligoclonality or false clonality will not usually be reproducible and the finding of the same clone from different amplifications is often reassuring. Even more helpful is testing of samples from more than 1 site. Several studies^59–61  have demonstrated that detecting the same clone from different sites, including a single skin and peripheral blood sample, increases the reliability of these tests. Therefore, when results are questionable, one can request retesting on a new sample for comparison of the possible clone. In T-cell proliferations, the combined detection of both TRG and TRB rearrangements has been shown to increase the reliability of detection of a true clone, but this approach will underdiagnose clonality in some cases.⁶² 

Although the BIOMED-II primers have greatly improved our ability to detect B- and T-cell clones in paraffin-embedded tissues, false-negative results may still occur.^63,64  These may be secondary to DNA degradation due to formalin fixation or due to the inability of the primers used in the test to actually anneal with the clonal rearrangement. Therefore, in the presence of morphologic and immunophenotypic evidence of lymphoma, a negative antigen receptor gene rearrangement study finding should not be interpreted as evidence of a benign proliferation.

Many monoclonal antibody therapies are available either as primary or secondary regimen in the treatment of cutaneous lymphoma; some clinical trials are in progress to establish the efficacies in these disorders, particularly in the areas of late-stage MF, CD30⁺ anaplastic large cell lymphoma, or most low-grade and high-grade B-cell lymphomas. If treatment with a monoclonal antibody is being considered, it is appropriate that IHC be performed to demonstrate the presence of target antigen (especially for CD20 and CD30) and that IHC be done at follow-up to show efficacy of treatment. Examples of such targeted therapies include rituximab, alemtuzumab targeting CD52,³⁴  and brentuximab targeting CD30 (Ber-H2).^22,26  The pathologist should be aware that there are no consensus guidelines in the interpretation of these results following therapy and that presence or absence of the marker may correlate poorly with response to therapy.

The above framework of cutaneous lymphoid lesions based on the WHO-EORTC classification is not complete. This review provides pertinent challenges, focused on the most challenging entities as well as newer entities. As many of these provisional entities evolve, with better-defined markers along with newer techniques such as genomics, it is expected that many of these entities will be further refined for practical diagnosis. In addition, the availability of targeted therapy may make a theranostic classification of these disorders more practical and useful than a pure morphologic or phenotype-based classification in the future.