Non-mycosis Fungoides Cutaneous T-cell Lymphomas
Non-mycosis Fungoides Cutaneous T-cell Lymphomas
Primary cutaneous CD30+ T-cell lymphoproliferative disorders are the second most common group of CTCLs, accounting for approximately 30% of the cases. Three types are recognized in the 2008 WHO classification: C-ALCL, lymphomatoid papulosis (LyP), and borderline lesions. These entities seem to represent a spectrum of related conditions that show overlapping clinical and histologic features and may coexist in individual patients Table 2. The disease must be distinguished from systemic ALCL with cutaneous involvement, which represents a separate disease with different clinical features, treatment, and outcome.
C-ALCL is defined as a CTCL composed of large anaplastic, pleomorphic cells with expression of CD30 in more than 75% of tumor cells. It occurs predominantly in the adult/elderly population, is rare in children, and shows a male predominance (2:1). Clinically, most patients have solitary or localized nodules, tumors, or papules that frequently become ulcerated. Partial or complete regression may occur, but skin relapses are often observed. Multifocal lesions are reported in approximately 20% of the patients, as was demonstrated in 3 of the submitted cases (cases 13, 266, and 299). Regional lymph nodes may be involved in 10% of the patients (see the end of this section and the section on lymphomatoid papulosis).
Morphologically, the infiltrate is usually diffuse and involves the dermis and subcutaneous tissue characterized by cohesive sheets of large, pleomorphic, and sometimes multinucleated giant cells resembling Reed-Sternberg cells Image 1A, Image 1B, Image 1C, and Image 1D. Although epidermal invasion may occur, epidermotropism is rare. A discrete inflammatory background may be present; however, when the inflammatory background is abundant, a differential diagnosis with LyP or even reactive conditions may be considered. Two cases submitted illustrated this differential diagnosis. In case 13, the extensive granulomatous reaction delayed the diagnosis of C-ALCL for several months. In case 254, the ulcerated lesion with numerous admixed neutrophils (neutrophil-rich or pyogenic variant) masked the neoplastic lymphoid population, mimicking an inflammatory condition such as pyoderma Image 1E and Image 1F. In both cases CD30 staining highlighted the atypical cells in the background and together with the clinical presentation helped to render the correct diagnosis Image 1G. Some cases, in addition to an abundant inflammatory infiltrate, are also accompanied by prominent epidermal hyperplasia mimicking squamous cell carcinoma. Case 146 was an example of the so-called pseudoepitheliomatous hyperplasia variant of C-ALCL Image 1H, Image 1I, and Image 1J. Most patients with pseudoepitheliomatous hyperplasia have 1 or several lesions that show spontaneous regression and even complete resolution within a few months of initial diagnosis. Because these cases seem to follow a relatively benign clinical course, a conservative watch and wait approach is recommended to avoid unnecessary surgery and/or radiation.
(Enlarge Image)
Image 1.
Cutaneous anaplastic large cell lymphoma (C-ALCL). A, Low-power view of a skin biopsy specimen showing a dense lymphoid infiltrate in the dermis without epidermotropism (H&E). B, Higher magnification of the skin infiltrate composed of large pleomorphic cells with open chromatin, prominent nucleoli, and abundant, pale to clear cytoplasm (H&E). Immunohistochemical staining show that the tumor cells are CD4+ (C) and strongly CD30+ (D) with a membrane and Golgi pattern (immunoperoxidase). (Case 90, courtesy of Gerald M. Penn, MD, PhD, and Robert H. Davi, MD, PhD.) Neutrophil-rich variant of C-ALCL. E, Low-power view of an ulcerated skin lesion with a mixed inflammatory infiltrate in the dermis (H&E). F, Higher magnification shows numerous neutrophils admixed with large atypical cells with prominent nucleoli and abundant cytoplasm (H&E). G, The large atypical cells are highlighted with CD30 staining (immunoperoxidase). (Case 254, courtesy of Abner Louissaint Jr, MD, PhD, and Judith A. Ferry, MD.) C-ALCL with pseudoepitheliomatous hyperplasia. H, Low-power view of a skin biopsy specimen with pseudoepitheliomatous hyperplasia and a dense lymphoid infiltrate beneath the epidermis (H&E). I, Higher magnification reveals the presence of large, pleomorphic cells with abundant cytoplasm intermingled with reactive small lymphocytes and histiocytes (H&E). J, Immunohistochemical staining with the anti-CD30 antibody highlights the presence of the large cells. Note the strong cell membrane positivity and the Golgi region (immunoperoxidase). K, In this case, the neoplastic cells were positive for CD8 (immunoperoxidase). (Case 146, courtesy of Michele Roullet, MD, and Celeste Bremer, MD.) L, Partial expression of epithelial membrane antigen in the tumor cells of a C-ALCL (immunoperoxidase). (Case 254, courtesy of Abner Louissaint Jr, MD, PhD, and Judith A. Ferry, MD.) M, Interphase fluorescence in situ hybridization (FISH) analysis using dual-color IRF4/MUM1 break-apart probe (BAC clones RP3-416J7, RP5-1077H22, and RP5-856G1) shows 1 allele with a normal colocalized signal (arrowheads) and the second allele with a split of the red and green signal (arrows), indicating an IRF4/MUM1 break. Note that the red signal is lost (FISH). (FISH analysis courtesy of ltziar Salaverria, PhD, and Reiner Siebert, MD, PhD.) (Case 266, courtesy of Muhammad Adeel Raza, MD, and Michelle Bonnett, MD.)
The immunophenotype of C-ALCL is characterized by a uniform membrane and Golgi expression of CD30 in most tumor cells (Image 1D, Image 1G, and Image 1J). The neoplastic cells show an activated CD4+ T-cell phenotype (Image 1C) with variable loss of CD2, CD5, CD7, and/or CD3, and expression of cytotoxic granule–associated proteins granzyme B, perforin, and/or TIA-1. Three cases submitted to the workshop represented examples of the classic phenotype of C-ALCL (cases 13, 90, and 299). Fewer than 5% of the cases have a cytotoxic CD8+ T-cell phenotype, as was illustrated in case 146 Image 1K. Interestingly, although CD4/CD8 double negative phenotype is rare, 5 cases submitted to the workshop revealed this phenotype (cases 30, 174, 220, 254, and 266). Of note, 3 of these cases developed under special conditions such as seroma associated in a patient with dermatomyositis (case 30), in an HIV-positive patient (case 174), and in a post-transplant setting (case 220). Although several reports describe C-ALCL in transplant recipients and HIV+ patients, neither conditions are recognized as an associated or predisposing clinical feature of conventional C-ALCL. Interestingly, a relatively high frequency of the neutrophilrich variant of C-ALCL has been reported in patients with compromised immune systems. Unlike systemic CD30+ lymphomas, most C-ALCL lesions express the cutaneous lymphocyte antigen (CLA), but are negative for the epithelial membrane antigen (EMA). Nevertheless, weak and focal heterogeneous expression of EMA may be observed and does not preclude the diagnosis of C-ALCL Image 1L. Staining for the anaplastic lymphoma kinase (ALK) protein, indicative of the t(2;5) chromosomal translocation or its variants is negative (exceptions described later in this article).
Genetically, most cases show clonal rearrangement of T-cell receptor (TCR) genes. Recently, a recurrent translocation involving IRF4/MUM1 oncogene locus at chromosome 6p25.3 was initially described to occur in 57% of C-ALCLs, as opposed to systemic ALCL or other peripheral T-cell lymphomas, not otherwise specified, where this translocation was rarely observed. Two further studies corroborated the presence of IRF4/MUM1 translocation almost exclusively in C-ALCL; however, the frequency was lower—20% in one study and 26% in the second study. The panel performed fluorescence in situ hybridization (FISH) analysis in 8 C-ALCL submitted cases (cases 13, 30, 90, 146, 220, 254, 266, and 299) using a break-apart IRF4 probe (BAC clones RP3-416J7, RP5-1077H22, and RP5-856G1). One case was positive (case 266) Image 1M, accounting for 13% of the cases analyzed. Taken together, all the cases reported in the literature– and the cases analyzed in this workshop, the frequency of IRF4 translocation in C-ALCL is estimated to be around 26% (24/90 cases). The expression of IRF4/MUM1 protein does not correlate with the presence of the translocation and cannot be used as surrogate marker. Interestingly, despite the absence of IRF4/MUM1 translocation in systemic ALCL cases, the vast majority of cases show a strong expression of IRF4/MUM1 protein, suggesting that the expression of this protein might be important for the pathogenesis of cutaneous and systemic ALCL cases.
The prognosis of C-ALCL is favorable, with 5-year survival rates between 76% and 96%. Spontaneous tumor regression, either partial or complete, is observed in 23% to 44% of the cases after a median period of 2 months (1 week to 6 months). The disease recurs in the skin in 39% of the cases. Importantly, regional lymph node involvement does not change the good prognosis of the disease. As initial therapy for patients presenting with solitary or few localized lesions, surgical excision or radiotherapy is recommended. With both therapeutic approaches, 95% complete remission can be achieved. Low-dose methotrexate has been proposed as therapy for multifocal C-ALCL. The use of multiagent chemotherapy is no longer recommended, except for rare cases of systemic disease.
Isolated cases of C-ALCL without systemic disease and cytoplasmic ALK protein expression in the tumor cells characteristic of a variant translocation have been reported.– These cases demonstrate the diagnostic pitfalls that may be encountered when differentiating C-ALCL from systemic ALCL with cutaneous involvement. Before and during the workshop, the significance of ALK protein expression in "primary cutaneous" CD30+ LPD indicative of ALK translocation was extensively discussed by the panel. Four cases submitted to the workshop highlighted the problems in classifying these cases (cases 42, 233, 238, and 288). Two cases presented as a solitary lesion, 1 case was associated with MF, and 1 case developed in a clinical setting resembling MF. The patients included 3 men between 27 and 70 years of age and a 64-year-old woman, with follow-up periods ranging between 1 and 20 years and no systemic disease. FISH analysis using a break-apart probe set flanking the ALK locus confirmed the presence of an ALK translocation in all 4 cases. An example of the complexity of these cases was demonstrated by case 288 of a 31-year-old patient who developed a scaly skin rash when he was 11 years old, which was diagnosed as parapsoriasis and treated with psoralen/UVA (PUVA). He presented again at age 21 years with dry scaly patches on his hands, arms, and legs and was lost to follow-up before further treatment. At age 31 years, the patient presented again with a 2-year history of an 8-mm ulcerated nodule in his popliteal fossa. Elsewhere, the skin showed a widespread poikilodermatous red rash with several interspersed scaly patches that was considered clinically to be poikilodermatous MF. Review of his first skin biopsy findings from age 11 years showed a mild lymphocytic infiltrate without atypia in the epidermis and upper dermis, which was retrospectively demonstrated to be cytoplasmic ALK+. A biopsy of the skin nodule 20 years later showed the typical morphology and phenotype of C-ALCL, whereas a biopsy of the skin rash revealed histologic features suggestive of MF; both biopsies showed cytoplasmic ALK positivity. In all 3 lesions, an ALK break indicative of a translocation was demonstrated with FISH. Case 238 was very similar and corresponded to a 70-year-old man with a 20-year history of MF and new ulcerated nodules in the foot. A biopsy of these nodules showed the classic features of C-ALCL, with cytoplasmic ALK positivity and ALK break demonstrated with FISH. The patient experienced a recurrence of similar lesions 7 years later on the back and buttocks, which were clonally related to the skin nodules in the foot, as demonstrated by identical clonal product size peaks in the TCR gene rearrangement studies. Unfortunately, the clonal relatedness of the MF lesion and the ALK+ C-ALCL could not be demonstrated, leaving open the question of whether these represented 2 independent lesions or if the ALK+ C-ALCL was the result of transformed MF. However, neither of these patients developed a systemic disease. In case 233, a 27-year-old man with a solitary lesion, the ALK staining was nuclear and cytoplasmic, characteristic of a t(2;5) chromosomal translocation. This patient had been followed up for only 1 year, and the possibility that he might develop a systemic disease cannot be excluded.
After extensive discussions, the panel arrived at the conclusion that cases with ALK positivity, such as the 4 cases submitted to the workshop, should be diagnosed as ALK+ C-ALCL until data from more cases can show whether these cases represent true ALK+ C-ALCL cases or whether some cases represent transformed MF with secondary ALK translocation. Further proof that primary cutaneous ALK+ ALCLs exist was provided by a recent series of pediatric ALK+ ALCL, in which 6 children presented with a single skin lesion without developing systemic disease after a median follow-up of 7 years (ratio, 1-8 years). Interestingly, in contrast to the cases in the workshop where 3 of 4 cases expressed cytoplasmic ALK protein indicative of a variant translocation, in the pediatric population 5 of 6 cases showed nuclear-cytoplasmic ALK staining characteristic of the t(2;5) chromosomal translocation. Indeed, the question regarding how these patients should be treated still remains. In the reported cases, as well as in the cases in the workshop, complete resection of the lesion with or without local radiation resulted in complete remissions. Conservative therapy seems justified after systemic disease is thoroughly excluded in these cases.
LyP is defined as a chronic, recurrent, self-healing skin disease composed of large anaplastic, immunoblastic, or Hodgkin-like cells in a marked inflammatory background. LyP occurs mainly in adults (median age, 45 years) but children may also be affected. A male predominance has been found, with a male-to-female ratio of 2 to 3:1. Clinically, LyP is characterized by the presence of papular, papulonecrotic, and/or nodular skin lesions at different stages of development affecting predominantly the trunk and extremities. Skin lesions regress spontaneously within 3 to 12 weeks, and may leave superficial, hyperpigmented scars. The duration of the disease may vary from a couple of months up to many years.
The 2008 WHO classification recognizes 3 histologic subtypes of LyP, types A, B, and C Image 2. These histologic subtypes represent a spectrum of overlapping features that in part correlates with the age of the skin lesion biopsy specimen, and may be present in individual patients at the same time, as was demonstrated in case 135. Immunophenotypically, the CD30+ large atypical cells are CD4+ in most cases, but CD8+ cases, especially in children, have been reported and represent about 5% of LyP types A-C. Although clonally rearranged TCR genes have been detected in 60% of the lesions, from a clinical perspective LyP is not considered a malignant disorder. Ten cases were submitted to the workshop with the diagnosis of LyP (cases 69, 87, 128, 135, 152, 162, 178, 222, 256, and 273) representing the wide morphologic spectrum of the disease. LyP type A lesions are characterized by a wedge-shaped infiltrate with scattered or small clusters of large, sometimes multinucleated or Reed-Sternberg–like, CD30-positive cells intermingled with numerous inflammatory cells, such as histiocytes, small lymphocytes, neutrophils, and/or eosinophils (Image 2A, Image 2B, and Image 2C). LyP type A is the most common histologic presentation (cases 87, 135, and 256). LyP type B is uncommon (<10%) and is characterized by an epidermotropic infiltrate of small atypical cells with cerebriform nuclei that histologically resembles MF (Image 2D and Image 2E). Although originally it was reported that the tumor cells in LyP type B are CD30 negative, further experience has shown that many cases are CD30 positive, as demonstrated by case 152 (Image 2F). LyP type C lesions are composed of a monotonous population of cohesive sheets of CD30+ large anaplastic lymphoid cells with only few admixed reactive inflammatory cells mimicking C-ALCL (Image 2G). The differential diagnosis with C-ALCL is not always easy and requires a careful clinicopathologic correlation. This was seen in case 69, a 68-year-old man who had papules on the right foot that resolved over 8 weeks. The lesions recurred 2 months later in the legs and earlobes and regressed after 3 to 4 weeks. Morphologically, the lesions were characterized by a diffuse, monotonous infiltration of large, cohesive CD30+, CD3+, CD4–, CD8– cells involving mainly the dermis (Image 2H and Image 2I). Clonal analysis revealed that the different lesions were clonally related. Interestingly, FISH analysis revealed a break in the IRF4/MUM1 locus indicative of a 6p25.3 translocation characteristic of C-ALCL. Over a period of several years the patient has continued with the same clinical findings. Despite the striking morphology and presence of the IRF4/MUM1 break, the panel agreed that because of the waxing and waning clinical picture, this case was better classified as LyP type C.
(Enlarge Image)
Image 2.
Histopathologic features of lymphomatoid papulosis, type A. A, Low-power view of the skin biopsy specimen with a wedge-shaped perivascular lymphoid infiltrate without epidermotropism that extends into the adnexae in the deep dermis (H&E). B, Higher magnification reveals that the infiltrate is composed of scattered large cells intermingled with numerous inflammatory cells such as histiocytes, small lymphocytes, and eosinophils (H&E). C, CD30 staining highlights the large atypical cells (immunoperoxidase). (Case 87, courtesy of Nicky Leeborg, MD, and Katalin Kelemen, MD, PhD.) Histopathologic features of lymphomatoid papulosis, type B. D, Skin biopsy specimen with an epidermotropic infiltrate of small atypical cells resembling mycosis fungoides (H&E). E, CD3 staining highlights the epidermotropism (immunoperoxidase). F, Some of the tumor cells are CD30+ (immunoperoxidase). (Case 152, courtesy of Cordelia Sever, MD, and Emily DeSantis, MD.) Histopathologic features of lymphomatoid papulosis, type C. G, Low-power view of a skin biopsy specimen with a dense lymphoid infiltrate in the dermis. Higher magnification shows that the infiltrate is composed of large anaplastic cells with abundant cytoplasm. Note that frequent mitoses are observed (inset) (H&E). The tumor cells are strongly CD30+ (H) and CD4+ (I) (immunoperoxidase). (Case 69, courtesy of Eric D. Hsi, MD, and Andrew Feldman, MD.) Histopathologic features of lymphomatoid papulosis, type D. J, Skin biopsy specimen with a striking epidermotropic infiltrate characterized by medium-sized, pleomorphic lymphocytes. The tumor cells are strongly CD30+ (K) and CD8+ (L) mimicking primary cutaneous CD8+ aggressive epidermotropic cytotoxic T-cell lymphoma (immunoperoxidase). (Case 178, courtesy of Carlos A. Torres-Cabala, MD, Victor G. Prieto, MD, PhD, Francisco Vega, MD, PhD, and Madeleine Duvic, MD.)
Recently, cases with clinical features of LyP but with histopathologic changes mimicking primary aggressive epidermotropic CD8+ cytotoxic T-cell lymphoma have been recognized and the term LyP type D has been proposed for this unusual variant. This variant is characterized by medium-sized, pleomorphic lymphocytes infiltrating the epidermis with a pagetoid reticulosis-like pattern (Image 2J). The neoplastic cells express CD30, CD8 (Image 2K and Image 2L), and βF1 positivity and at least 1 cytotoxic protein, TIA1 and/or granzyme B. Interestingly, CD56 might be positive; however, Epstein-Barr virus (EBV) is always negative. The diagnostic challenge of these cases was well demonstrated by cases 38 and 178, which originally were diagnosed as primary cutaneous CD8+ aggressive epidermotropic cytotoxic T-cell lymphoma. Both cases had a clinical presentation typical of LyP. Although CD30 expression has been reported sporadically in primary aggressive epidermotropic CD8+ cytotoxic T-cell lymphoma, CTCLs with strong and uniform expression of CD30 and CD8 most likely belong to the spectrum of primary cutaneous CD30+ LPD. Features that favor a diagnosis of LyP type D, besides the homogeneous expression of CD30, are the relative young age at onset, the clinical presentation with waxing and waning erythematous papules or small nodules that resolve spontaneously within several weeks, and the indolent behavior. The differential diagnosis of LyP includes not only other CTCLs but also various inflammatory and infectious disorders that might have large, atypical, activated CD30+ cells (discussed in the article by Sarantopoulos et al in this issue of the Journal).
In 5% to 20% of patients, LyP may be preceded by, associated with, or followed by another type of malignant (cutaneous) lymphoma, generally MF, C-ALCL, or Hodgkin lymphoma. Case 128 was particularly instructive because LyP developed 5 years after a systemic ALK-ALCL. Molecular analysis was necessary in this case to demonstrate that the skin lesions were clonally unrelated to the systemic ALCL. An additional challenge in the differential diagnosis with systemic ALCL is the fact that 5% to 10% of LyP cases may involve regional lymph nodes, which does not make it a systemic disease, as was shown in cases 222 and 273. Useful in the differential diagnosis is the fact that the nodal lesions are in the lymphatic drainage of the dominant cutaneous lesion. Without clinical information, it is not possible, based on morphology alone, to distinguish between lymphatic drainage or nodal involvement by systemic ALCL. Both might show prominent subcapsular and sinusoidal infiltration. To complicate things more, recently it was reported that nodal involvement by C-ALCL may mimic classic Hodgkin lymphoma, not only morphologically but also phenotypically with coexpression of CD30 and CD15 by the tumor cells that may lead to a mistaken diagnosis of classic Hodgkin lymphoma. In these cases molecular analysis is necessary to demonstrate a common T-cell clone. Of note, in some cases the regional lymphadenopathy might regress spontaneously. Patients with regional lymph node involvement have a similar prognosis to those who only have skin lesions. LyP has an excellent prognosis; only 2% of the patients died of systemic disease after a median follow-up duration of 77 months.
According to the 2008 WHO classification, the term "borderline" refers to cases in which, despite careful clinicopathologic correlation, a definite distinction between C-ALCL and LyP cannot be made. However, follow-up will generally disclose whether such patients have C-ALCL or LyP. It is important to stress that the borderline diagnosis should only be used at the time of first diagnosis and not for individual lesions in the follow-up of the disease. Case 209 exemplified the difficulties that might be encountered in daily practice. An 11-year-old girl presented with a 6-week history of a 1.5-cm nodule in the right proximal forearm and a large indurated mass in the right arm. Biopsies of both lesions were performed. The nodule had the typical morphology of C-ALCL with sheets of cohesive large, pleomorphic CD30+ cells and pseudoepitheliomatous hyperplasia of the skin. The indurated mass had the morphology of LyP with scattered large CD30+ cells in an inflammatory background. One lesion was surgically removed and the other underwent spontaneous resolution after 4 months. A diagnosis of C-ALCL was proposed because the large, rapidly growing lesion suggested an aggressive lymphoma. However, the panel felt that because of the age of the patient and the morphologic discordance between the 2 lesions, this case was better classified as "borderline." Nevertheless, despite the clinical presentation, children and young patients (under 20 years) usually have LyP. In these cases, a watch and wait period is recommended to avoid unnecessary treatment. In most cases, the larger skin lesions disappeared completely within 2 to 4 months, as was well illustrated in this case.
The differential diagnosis of a cutaneous CD30+ LPD includes MF with transformed CD30+ large cells and skin localization of a systemic CD30+ ALCL. Both systemic ALK+ and ALK– ALCLs frequently involve lymph nodes and extranodal sites including skin, soft tissues, and bone. One of the major problems in the differential diagnosis is that there are no reliable histologic criteria to differentiate between a primary and secondary cutaneous CD30+ ALCL. It has been suggested that to exclude a secondary cutaneous CD30+ ALCL, adequate staging in all patients with a cutaneous CD30+ ALCL should be done. Seven cases of systemic ALCL with skin involvement were submitted to the workshop. In 2 of 3 systemic ALK– ALCLs submitted (cases 103 and 107) the skin was the initial manifestation of the disease. In general, systemic involvement occurs in less than 6 months, as occurred in these 2 cases (2 and 4 months, respectively). More importantly, patients with secondary cutaneous CD30+ ALCL have a significant difference in survival, with a survival of 25% at 5 years vs more than 90% in patients with primary CD30+ C-ALCL. Three of the 4 ALK+ ALCL cases (20, 73, and 167) were typical examples of the small cell variant with CD8+ phenotype. The small cell variant of ALK+ ALCL is frequently associated with widespread disseminated disease and the initial biopsy may come from skin or other extranodal sites, contributing to the difficulty in making this diagnosis. This variant represents approximately 10% of ALK+ ALCL, and the skin is involved in 78% of these cases. The patients are often young, as demonstrated in cases 20 and 73, and morphologically, the dermis is infiltrated by numerous small lymphocytes with irregular nuclei, which are often CD30–; only a minor population of large cells are CD30+. Although a particularly useful diagnostic feature in the small cell variant is the tendency of ALK+, CD30+, large cells to be distributed in a perivascular pattern rather than randomly scattered, in the cases submitted to the workshop only randomly scattered cells were CD30+ and ALK+.
Primary Cutaneous CD30+ T-cell Lymphoproliferative Disorders
Primary cutaneous CD30+ T-cell lymphoproliferative disorders are the second most common group of CTCLs, accounting for approximately 30% of the cases. Three types are recognized in the 2008 WHO classification: C-ALCL, lymphomatoid papulosis (LyP), and borderline lesions. These entities seem to represent a spectrum of related conditions that show overlapping clinical and histologic features and may coexist in individual patients Table 2. The disease must be distinguished from systemic ALCL with cutaneous involvement, which represents a separate disease with different clinical features, treatment, and outcome.
Primary C-ALCL
C-ALCL is defined as a CTCL composed of large anaplastic, pleomorphic cells with expression of CD30 in more than 75% of tumor cells. It occurs predominantly in the adult/elderly population, is rare in children, and shows a male predominance (2:1). Clinically, most patients have solitary or localized nodules, tumors, or papules that frequently become ulcerated. Partial or complete regression may occur, but skin relapses are often observed. Multifocal lesions are reported in approximately 20% of the patients, as was demonstrated in 3 of the submitted cases (cases 13, 266, and 299). Regional lymph nodes may be involved in 10% of the patients (see the end of this section and the section on lymphomatoid papulosis).
Morphologically, the infiltrate is usually diffuse and involves the dermis and subcutaneous tissue characterized by cohesive sheets of large, pleomorphic, and sometimes multinucleated giant cells resembling Reed-Sternberg cells Image 1A, Image 1B, Image 1C, and Image 1D. Although epidermal invasion may occur, epidermotropism is rare. A discrete inflammatory background may be present; however, when the inflammatory background is abundant, a differential diagnosis with LyP or even reactive conditions may be considered. Two cases submitted illustrated this differential diagnosis. In case 13, the extensive granulomatous reaction delayed the diagnosis of C-ALCL for several months. In case 254, the ulcerated lesion with numerous admixed neutrophils (neutrophil-rich or pyogenic variant) masked the neoplastic lymphoid population, mimicking an inflammatory condition such as pyoderma Image 1E and Image 1F. In both cases CD30 staining highlighted the atypical cells in the background and together with the clinical presentation helped to render the correct diagnosis Image 1G. Some cases, in addition to an abundant inflammatory infiltrate, are also accompanied by prominent epidermal hyperplasia mimicking squamous cell carcinoma. Case 146 was an example of the so-called pseudoepitheliomatous hyperplasia variant of C-ALCL Image 1H, Image 1I, and Image 1J. Most patients with pseudoepitheliomatous hyperplasia have 1 or several lesions that show spontaneous regression and even complete resolution within a few months of initial diagnosis. Because these cases seem to follow a relatively benign clinical course, a conservative watch and wait approach is recommended to avoid unnecessary surgery and/or radiation.
(Enlarge Image)
Image 1.
Cutaneous anaplastic large cell lymphoma (C-ALCL). A, Low-power view of a skin biopsy specimen showing a dense lymphoid infiltrate in the dermis without epidermotropism (H&E). B, Higher magnification of the skin infiltrate composed of large pleomorphic cells with open chromatin, prominent nucleoli, and abundant, pale to clear cytoplasm (H&E). Immunohistochemical staining show that the tumor cells are CD4+ (C) and strongly CD30+ (D) with a membrane and Golgi pattern (immunoperoxidase). (Case 90, courtesy of Gerald M. Penn, MD, PhD, and Robert H. Davi, MD, PhD.) Neutrophil-rich variant of C-ALCL. E, Low-power view of an ulcerated skin lesion with a mixed inflammatory infiltrate in the dermis (H&E). F, Higher magnification shows numerous neutrophils admixed with large atypical cells with prominent nucleoli and abundant cytoplasm (H&E). G, The large atypical cells are highlighted with CD30 staining (immunoperoxidase). (Case 254, courtesy of Abner Louissaint Jr, MD, PhD, and Judith A. Ferry, MD.) C-ALCL with pseudoepitheliomatous hyperplasia. H, Low-power view of a skin biopsy specimen with pseudoepitheliomatous hyperplasia and a dense lymphoid infiltrate beneath the epidermis (H&E). I, Higher magnification reveals the presence of large, pleomorphic cells with abundant cytoplasm intermingled with reactive small lymphocytes and histiocytes (H&E). J, Immunohistochemical staining with the anti-CD30 antibody highlights the presence of the large cells. Note the strong cell membrane positivity and the Golgi region (immunoperoxidase). K, In this case, the neoplastic cells were positive for CD8 (immunoperoxidase). (Case 146, courtesy of Michele Roullet, MD, and Celeste Bremer, MD.) L, Partial expression of epithelial membrane antigen in the tumor cells of a C-ALCL (immunoperoxidase). (Case 254, courtesy of Abner Louissaint Jr, MD, PhD, and Judith A. Ferry, MD.) M, Interphase fluorescence in situ hybridization (FISH) analysis using dual-color IRF4/MUM1 break-apart probe (BAC clones RP3-416J7, RP5-1077H22, and RP5-856G1) shows 1 allele with a normal colocalized signal (arrowheads) and the second allele with a split of the red and green signal (arrows), indicating an IRF4/MUM1 break. Note that the red signal is lost (FISH). (FISH analysis courtesy of ltziar Salaverria, PhD, and Reiner Siebert, MD, PhD.) (Case 266, courtesy of Muhammad Adeel Raza, MD, and Michelle Bonnett, MD.)
The immunophenotype of C-ALCL is characterized by a uniform membrane and Golgi expression of CD30 in most tumor cells (Image 1D, Image 1G, and Image 1J). The neoplastic cells show an activated CD4+ T-cell phenotype (Image 1C) with variable loss of CD2, CD5, CD7, and/or CD3, and expression of cytotoxic granule–associated proteins granzyme B, perforin, and/or TIA-1. Three cases submitted to the workshop represented examples of the classic phenotype of C-ALCL (cases 13, 90, and 299). Fewer than 5% of the cases have a cytotoxic CD8+ T-cell phenotype, as was illustrated in case 146 Image 1K. Interestingly, although CD4/CD8 double negative phenotype is rare, 5 cases submitted to the workshop revealed this phenotype (cases 30, 174, 220, 254, and 266). Of note, 3 of these cases developed under special conditions such as seroma associated in a patient with dermatomyositis (case 30), in an HIV-positive patient (case 174), and in a post-transplant setting (case 220). Although several reports describe C-ALCL in transplant recipients and HIV+ patients, neither conditions are recognized as an associated or predisposing clinical feature of conventional C-ALCL. Interestingly, a relatively high frequency of the neutrophilrich variant of C-ALCL has been reported in patients with compromised immune systems. Unlike systemic CD30+ lymphomas, most C-ALCL lesions express the cutaneous lymphocyte antigen (CLA), but are negative for the epithelial membrane antigen (EMA). Nevertheless, weak and focal heterogeneous expression of EMA may be observed and does not preclude the diagnosis of C-ALCL Image 1L. Staining for the anaplastic lymphoma kinase (ALK) protein, indicative of the t(2;5) chromosomal translocation or its variants is negative (exceptions described later in this article).
Genetically, most cases show clonal rearrangement of T-cell receptor (TCR) genes. Recently, a recurrent translocation involving IRF4/MUM1 oncogene locus at chromosome 6p25.3 was initially described to occur in 57% of C-ALCLs, as opposed to systemic ALCL or other peripheral T-cell lymphomas, not otherwise specified, where this translocation was rarely observed. Two further studies corroborated the presence of IRF4/MUM1 translocation almost exclusively in C-ALCL; however, the frequency was lower—20% in one study and 26% in the second study. The panel performed fluorescence in situ hybridization (FISH) analysis in 8 C-ALCL submitted cases (cases 13, 30, 90, 146, 220, 254, 266, and 299) using a break-apart IRF4 probe (BAC clones RP3-416J7, RP5-1077H22, and RP5-856G1). One case was positive (case 266) Image 1M, accounting for 13% of the cases analyzed. Taken together, all the cases reported in the literature– and the cases analyzed in this workshop, the frequency of IRF4 translocation in C-ALCL is estimated to be around 26% (24/90 cases). The expression of IRF4/MUM1 protein does not correlate with the presence of the translocation and cannot be used as surrogate marker. Interestingly, despite the absence of IRF4/MUM1 translocation in systemic ALCL cases, the vast majority of cases show a strong expression of IRF4/MUM1 protein, suggesting that the expression of this protein might be important for the pathogenesis of cutaneous and systemic ALCL cases.
The prognosis of C-ALCL is favorable, with 5-year survival rates between 76% and 96%. Spontaneous tumor regression, either partial or complete, is observed in 23% to 44% of the cases after a median period of 2 months (1 week to 6 months). The disease recurs in the skin in 39% of the cases. Importantly, regional lymph node involvement does not change the good prognosis of the disease. As initial therapy for patients presenting with solitary or few localized lesions, surgical excision or radiotherapy is recommended. With both therapeutic approaches, 95% complete remission can be achieved. Low-dose methotrexate has been proposed as therapy for multifocal C-ALCL. The use of multiagent chemotherapy is no longer recommended, except for rare cases of systemic disease.
ALK+ Primary C-ALCL
Isolated cases of C-ALCL without systemic disease and cytoplasmic ALK protein expression in the tumor cells characteristic of a variant translocation have been reported.– These cases demonstrate the diagnostic pitfalls that may be encountered when differentiating C-ALCL from systemic ALCL with cutaneous involvement. Before and during the workshop, the significance of ALK protein expression in "primary cutaneous" CD30+ LPD indicative of ALK translocation was extensively discussed by the panel. Four cases submitted to the workshop highlighted the problems in classifying these cases (cases 42, 233, 238, and 288). Two cases presented as a solitary lesion, 1 case was associated with MF, and 1 case developed in a clinical setting resembling MF. The patients included 3 men between 27 and 70 years of age and a 64-year-old woman, with follow-up periods ranging between 1 and 20 years and no systemic disease. FISH analysis using a break-apart probe set flanking the ALK locus confirmed the presence of an ALK translocation in all 4 cases. An example of the complexity of these cases was demonstrated by case 288 of a 31-year-old patient who developed a scaly skin rash when he was 11 years old, which was diagnosed as parapsoriasis and treated with psoralen/UVA (PUVA). He presented again at age 21 years with dry scaly patches on his hands, arms, and legs and was lost to follow-up before further treatment. At age 31 years, the patient presented again with a 2-year history of an 8-mm ulcerated nodule in his popliteal fossa. Elsewhere, the skin showed a widespread poikilodermatous red rash with several interspersed scaly patches that was considered clinically to be poikilodermatous MF. Review of his first skin biopsy findings from age 11 years showed a mild lymphocytic infiltrate without atypia in the epidermis and upper dermis, which was retrospectively demonstrated to be cytoplasmic ALK+. A biopsy of the skin nodule 20 years later showed the typical morphology and phenotype of C-ALCL, whereas a biopsy of the skin rash revealed histologic features suggestive of MF; both biopsies showed cytoplasmic ALK positivity. In all 3 lesions, an ALK break indicative of a translocation was demonstrated with FISH. Case 238 was very similar and corresponded to a 70-year-old man with a 20-year history of MF and new ulcerated nodules in the foot. A biopsy of these nodules showed the classic features of C-ALCL, with cytoplasmic ALK positivity and ALK break demonstrated with FISH. The patient experienced a recurrence of similar lesions 7 years later on the back and buttocks, which were clonally related to the skin nodules in the foot, as demonstrated by identical clonal product size peaks in the TCR gene rearrangement studies. Unfortunately, the clonal relatedness of the MF lesion and the ALK+ C-ALCL could not be demonstrated, leaving open the question of whether these represented 2 independent lesions or if the ALK+ C-ALCL was the result of transformed MF. However, neither of these patients developed a systemic disease. In case 233, a 27-year-old man with a solitary lesion, the ALK staining was nuclear and cytoplasmic, characteristic of a t(2;5) chromosomal translocation. This patient had been followed up for only 1 year, and the possibility that he might develop a systemic disease cannot be excluded.
After extensive discussions, the panel arrived at the conclusion that cases with ALK positivity, such as the 4 cases submitted to the workshop, should be diagnosed as ALK+ C-ALCL until data from more cases can show whether these cases represent true ALK+ C-ALCL cases or whether some cases represent transformed MF with secondary ALK translocation. Further proof that primary cutaneous ALK+ ALCLs exist was provided by a recent series of pediatric ALK+ ALCL, in which 6 children presented with a single skin lesion without developing systemic disease after a median follow-up of 7 years (ratio, 1-8 years). Interestingly, in contrast to the cases in the workshop where 3 of 4 cases expressed cytoplasmic ALK protein indicative of a variant translocation, in the pediatric population 5 of 6 cases showed nuclear-cytoplasmic ALK staining characteristic of the t(2;5) chromosomal translocation. Indeed, the question regarding how these patients should be treated still remains. In the reported cases, as well as in the cases in the workshop, complete resection of the lesion with or without local radiation resulted in complete remissions. Conservative therapy seems justified after systemic disease is thoroughly excluded in these cases.
Lymphomatoid Papulosis
LyP is defined as a chronic, recurrent, self-healing skin disease composed of large anaplastic, immunoblastic, or Hodgkin-like cells in a marked inflammatory background. LyP occurs mainly in adults (median age, 45 years) but children may also be affected. A male predominance has been found, with a male-to-female ratio of 2 to 3:1. Clinically, LyP is characterized by the presence of papular, papulonecrotic, and/or nodular skin lesions at different stages of development affecting predominantly the trunk and extremities. Skin lesions regress spontaneously within 3 to 12 weeks, and may leave superficial, hyperpigmented scars. The duration of the disease may vary from a couple of months up to many years.
The 2008 WHO classification recognizes 3 histologic subtypes of LyP, types A, B, and C Image 2. These histologic subtypes represent a spectrum of overlapping features that in part correlates with the age of the skin lesion biopsy specimen, and may be present in individual patients at the same time, as was demonstrated in case 135. Immunophenotypically, the CD30+ large atypical cells are CD4+ in most cases, but CD8+ cases, especially in children, have been reported and represent about 5% of LyP types A-C. Although clonally rearranged TCR genes have been detected in 60% of the lesions, from a clinical perspective LyP is not considered a malignant disorder. Ten cases were submitted to the workshop with the diagnosis of LyP (cases 69, 87, 128, 135, 152, 162, 178, 222, 256, and 273) representing the wide morphologic spectrum of the disease. LyP type A lesions are characterized by a wedge-shaped infiltrate with scattered or small clusters of large, sometimes multinucleated or Reed-Sternberg–like, CD30-positive cells intermingled with numerous inflammatory cells, such as histiocytes, small lymphocytes, neutrophils, and/or eosinophils (Image 2A, Image 2B, and Image 2C). LyP type A is the most common histologic presentation (cases 87, 135, and 256). LyP type B is uncommon (<10%) and is characterized by an epidermotropic infiltrate of small atypical cells with cerebriform nuclei that histologically resembles MF (Image 2D and Image 2E). Although originally it was reported that the tumor cells in LyP type B are CD30 negative, further experience has shown that many cases are CD30 positive, as demonstrated by case 152 (Image 2F). LyP type C lesions are composed of a monotonous population of cohesive sheets of CD30+ large anaplastic lymphoid cells with only few admixed reactive inflammatory cells mimicking C-ALCL (Image 2G). The differential diagnosis with C-ALCL is not always easy and requires a careful clinicopathologic correlation. This was seen in case 69, a 68-year-old man who had papules on the right foot that resolved over 8 weeks. The lesions recurred 2 months later in the legs and earlobes and regressed after 3 to 4 weeks. Morphologically, the lesions were characterized by a diffuse, monotonous infiltration of large, cohesive CD30+, CD3+, CD4–, CD8– cells involving mainly the dermis (Image 2H and Image 2I). Clonal analysis revealed that the different lesions were clonally related. Interestingly, FISH analysis revealed a break in the IRF4/MUM1 locus indicative of a 6p25.3 translocation characteristic of C-ALCL. Over a period of several years the patient has continued with the same clinical findings. Despite the striking morphology and presence of the IRF4/MUM1 break, the panel agreed that because of the waxing and waning clinical picture, this case was better classified as LyP type C.
(Enlarge Image)
Image 2.
Histopathologic features of lymphomatoid papulosis, type A. A, Low-power view of the skin biopsy specimen with a wedge-shaped perivascular lymphoid infiltrate without epidermotropism that extends into the adnexae in the deep dermis (H&E). B, Higher magnification reveals that the infiltrate is composed of scattered large cells intermingled with numerous inflammatory cells such as histiocytes, small lymphocytes, and eosinophils (H&E). C, CD30 staining highlights the large atypical cells (immunoperoxidase). (Case 87, courtesy of Nicky Leeborg, MD, and Katalin Kelemen, MD, PhD.) Histopathologic features of lymphomatoid papulosis, type B. D, Skin biopsy specimen with an epidermotropic infiltrate of small atypical cells resembling mycosis fungoides (H&E). E, CD3 staining highlights the epidermotropism (immunoperoxidase). F, Some of the tumor cells are CD30+ (immunoperoxidase). (Case 152, courtesy of Cordelia Sever, MD, and Emily DeSantis, MD.) Histopathologic features of lymphomatoid papulosis, type C. G, Low-power view of a skin biopsy specimen with a dense lymphoid infiltrate in the dermis. Higher magnification shows that the infiltrate is composed of large anaplastic cells with abundant cytoplasm. Note that frequent mitoses are observed (inset) (H&E). The tumor cells are strongly CD30+ (H) and CD4+ (I) (immunoperoxidase). (Case 69, courtesy of Eric D. Hsi, MD, and Andrew Feldman, MD.) Histopathologic features of lymphomatoid papulosis, type D. J, Skin biopsy specimen with a striking epidermotropic infiltrate characterized by medium-sized, pleomorphic lymphocytes. The tumor cells are strongly CD30+ (K) and CD8+ (L) mimicking primary cutaneous CD8+ aggressive epidermotropic cytotoxic T-cell lymphoma (immunoperoxidase). (Case 178, courtesy of Carlos A. Torres-Cabala, MD, Victor G. Prieto, MD, PhD, Francisco Vega, MD, PhD, and Madeleine Duvic, MD.)
Recently, cases with clinical features of LyP but with histopathologic changes mimicking primary aggressive epidermotropic CD8+ cytotoxic T-cell lymphoma have been recognized and the term LyP type D has been proposed for this unusual variant. This variant is characterized by medium-sized, pleomorphic lymphocytes infiltrating the epidermis with a pagetoid reticulosis-like pattern (Image 2J). The neoplastic cells express CD30, CD8 (Image 2K and Image 2L), and βF1 positivity and at least 1 cytotoxic protein, TIA1 and/or granzyme B. Interestingly, CD56 might be positive; however, Epstein-Barr virus (EBV) is always negative. The diagnostic challenge of these cases was well demonstrated by cases 38 and 178, which originally were diagnosed as primary cutaneous CD8+ aggressive epidermotropic cytotoxic T-cell lymphoma. Both cases had a clinical presentation typical of LyP. Although CD30 expression has been reported sporadically in primary aggressive epidermotropic CD8+ cytotoxic T-cell lymphoma, CTCLs with strong and uniform expression of CD30 and CD8 most likely belong to the spectrum of primary cutaneous CD30+ LPD. Features that favor a diagnosis of LyP type D, besides the homogeneous expression of CD30, are the relative young age at onset, the clinical presentation with waxing and waning erythematous papules or small nodules that resolve spontaneously within several weeks, and the indolent behavior. The differential diagnosis of LyP includes not only other CTCLs but also various inflammatory and infectious disorders that might have large, atypical, activated CD30+ cells (discussed in the article by Sarantopoulos et al in this issue of the Journal).
In 5% to 20% of patients, LyP may be preceded by, associated with, or followed by another type of malignant (cutaneous) lymphoma, generally MF, C-ALCL, or Hodgkin lymphoma. Case 128 was particularly instructive because LyP developed 5 years after a systemic ALK-ALCL. Molecular analysis was necessary in this case to demonstrate that the skin lesions were clonally unrelated to the systemic ALCL. An additional challenge in the differential diagnosis with systemic ALCL is the fact that 5% to 10% of LyP cases may involve regional lymph nodes, which does not make it a systemic disease, as was shown in cases 222 and 273. Useful in the differential diagnosis is the fact that the nodal lesions are in the lymphatic drainage of the dominant cutaneous lesion. Without clinical information, it is not possible, based on morphology alone, to distinguish between lymphatic drainage or nodal involvement by systemic ALCL. Both might show prominent subcapsular and sinusoidal infiltration. To complicate things more, recently it was reported that nodal involvement by C-ALCL may mimic classic Hodgkin lymphoma, not only morphologically but also phenotypically with coexpression of CD30 and CD15 by the tumor cells that may lead to a mistaken diagnosis of classic Hodgkin lymphoma. In these cases molecular analysis is necessary to demonstrate a common T-cell clone. Of note, in some cases the regional lymphadenopathy might regress spontaneously. Patients with regional lymph node involvement have a similar prognosis to those who only have skin lesions. LyP has an excellent prognosis; only 2% of the patients died of systemic disease after a median follow-up duration of 77 months.
Borderline Cases
According to the 2008 WHO classification, the term "borderline" refers to cases in which, despite careful clinicopathologic correlation, a definite distinction between C-ALCL and LyP cannot be made. However, follow-up will generally disclose whether such patients have C-ALCL or LyP. It is important to stress that the borderline diagnosis should only be used at the time of first diagnosis and not for individual lesions in the follow-up of the disease. Case 209 exemplified the difficulties that might be encountered in daily practice. An 11-year-old girl presented with a 6-week history of a 1.5-cm nodule in the right proximal forearm and a large indurated mass in the right arm. Biopsies of both lesions were performed. The nodule had the typical morphology of C-ALCL with sheets of cohesive large, pleomorphic CD30+ cells and pseudoepitheliomatous hyperplasia of the skin. The indurated mass had the morphology of LyP with scattered large CD30+ cells in an inflammatory background. One lesion was surgically removed and the other underwent spontaneous resolution after 4 months. A diagnosis of C-ALCL was proposed because the large, rapidly growing lesion suggested an aggressive lymphoma. However, the panel felt that because of the age of the patient and the morphologic discordance between the 2 lesions, this case was better classified as "borderline." Nevertheless, despite the clinical presentation, children and young patients (under 20 years) usually have LyP. In these cases, a watch and wait period is recommended to avoid unnecessary treatment. In most cases, the larger skin lesions disappeared completely within 2 to 4 months, as was well illustrated in this case.
Secondary Cutaneous CD30+ Anaplastic Large T-cell Lymphoma
The differential diagnosis of a cutaneous CD30+ LPD includes MF with transformed CD30+ large cells and skin localization of a systemic CD30+ ALCL. Both systemic ALK+ and ALK– ALCLs frequently involve lymph nodes and extranodal sites including skin, soft tissues, and bone. One of the major problems in the differential diagnosis is that there are no reliable histologic criteria to differentiate between a primary and secondary cutaneous CD30+ ALCL. It has been suggested that to exclude a secondary cutaneous CD30+ ALCL, adequate staging in all patients with a cutaneous CD30+ ALCL should be done. Seven cases of systemic ALCL with skin involvement were submitted to the workshop. In 2 of 3 systemic ALK– ALCLs submitted (cases 103 and 107) the skin was the initial manifestation of the disease. In general, systemic involvement occurs in less than 6 months, as occurred in these 2 cases (2 and 4 months, respectively). More importantly, patients with secondary cutaneous CD30+ ALCL have a significant difference in survival, with a survival of 25% at 5 years vs more than 90% in patients with primary CD30+ C-ALCL. Three of the 4 ALK+ ALCL cases (20, 73, and 167) were typical examples of the small cell variant with CD8+ phenotype. The small cell variant of ALK+ ALCL is frequently associated with widespread disseminated disease and the initial biopsy may come from skin or other extranodal sites, contributing to the difficulty in making this diagnosis. This variant represents approximately 10% of ALK+ ALCL, and the skin is involved in 78% of these cases. The patients are often young, as demonstrated in cases 20 and 73, and morphologically, the dermis is infiltrated by numerous small lymphocytes with irregular nuclei, which are often CD30–; only a minor population of large cells are CD30+. Although a particularly useful diagnostic feature in the small cell variant is the tendency of ALK+, CD30+, large cells to be distributed in a perivascular pattern rather than randomly scattered, in the cases submitted to the workshop only randomly scattered cells were CD30+ and ALK+.
