The vascular marker CD31 also highlights histiocytes and histiocyte-like cells within cutaneous tumors.

OBJECTIVES: While useful in diagnosing angiosarcomas, CD31 can also highlight histiocytes within soft tissue tumors and lead to errors in diagnosis. We sought to determine how often CD31 highlights cutaneous histiocytomas and histiocytoma mimics. METHODS: We examined eight epithelioid cell histiocytomas (ECHs), 12 xanthogranulomas (XGs), nine cases of Langerhans cell histiocytosis (LCH), eight reticulohistiocytomas, 11 xanthomas, 29 atypical fibroxanthomas, nine granular cell tumors, four cases of angiolymphoid hyperplasia with eosinophilia, nine intradermal Spitz nevi, and nine angiosarcomas with antibodies directed against CD31, CD34, CD163, and factor VIII. RESULTS: CD31 marked cells in three of 12 XGs, four of nine cases of LCH, one of eight reticulohistiocytomas, one of 11 xanthomas, 10 of 29 atypical fibroxanthomas, four of four cases of angiolymphoid hyperplasia with eosinophilia, nine of nine angiosarcomas, zero of nine granular cell tumors, and zero of eight ECHs. CD34 and factor VIII were negative in all nonvascular cases. CONCLUSIONS: Our results indicate that CD31 can mark lesional cells and imitate vascular tumors in cutaneous histiocytomas and histiocytoma mimics, an error that can be avoided by using a panel of antibodies.

Diagnostic utility of Fli-1 and D2-40 in distinguishing atypical fibroxanthoma from angiosarcoma.

Although in most cases one can easily distinguish between atypical fibroxanthomas and angiosarcomas, hemorrhagic atypical fibroxanthomas can pose a diagnostic problem. In rare cases, the large atypical cells of atypical fibroxanthoma can stain with CD31, leading to the erroneous diagnosis of angiosarcoma. We elected to further study this conundrum with 2 additional markers of lymphatic and vascular elements, namely D2-40 (podoplanin) and Fli-1, respectively. We studied 26 cases of atypical fibroxanthoma and 20 cases of angiosarcoma with Fli-1 and D2-40. We found that both Fli-1 and D2-40 stained a majority of cases of angiosarcoma (16/20 and 12/20, respectively), although only staining a minority of cases of atypical fibroxanthoma (8/26 for both). In addition, D2-40 staining of atypical fibroxanthoma was usually weak when positive, whereas Fli-1 staining of angiosarcomas was mostly strong and nuclear. Thus, both D2-40 and Fli-1 seem to be useful in distinguishing between atypical fibroxanthomas and angiosarcomas.

Diagnosing epidermolysis bullosa type and subtype in infancy using immunofluorescence microscopy: the Stanford experience.

The natural history of inherited epidermolysis bullosa (EB) varies significantly across subtypes. When confronted with an infant suspected to have EB, rapidly determining the type and subtype is critical in counselling families accurately about the infant's diagnosis and prognosis. Although transmission electron microscopy (TEM) has been considered the criterion standard for EB diagnosis, immunofluorescence microscopy (IFM) using monoclonal antibodies (mAbs) to EB-specific basement membrane zone proteins has several advantages, but few studies have evaluated the diagnostic utility of IFM. We sought to evaluate the clinical utility of IFM using an expanded panel of EB-specific mAbs. This was a retrospective review of pathology reports from infants younger < 1 year old with suspected EB primarily analyzed with IFM by the Stanford Dermatopathology service. Seventy-seven cases were identified for analysis, of which 20 were suboptimal for IFM analysis. Fifty-five cases were diagnosed with EB and classified as follows: EB simplex (n = 5), junctional EB (n = 31), dystrophic EB (n = 19). TEM was available in 36 of 55 cases (65%). IFM with an expanded panel of EB-specific mAbs should be considered the first-line diagnostic test to evaluate infants with clinically suspected EB.

Reconsidering the diagnostic and prognostic utility of LN-2 for undifferentiated pleomorphic sarcoma and atypical fibroxanthoma.

The topic of distinguishing atypical fibroxanthoma (AFX) from undifferentiated pleomorphic sarcoma (UPS), formerly malignant fibrous histiocytoma, is highly controversial. Although their clinical behavior is disparate, AFX and UPS commonly appear nearly identical on routine histopathologic examination. Although conceptually useful, subcategorization of UPS into superficial (confined to the dermis and subcutaneous tissue) and deep (involvement of fascia and deeper structures) types has not improved our ability to differentiate UPS from AFX. Numerous authors have purported LN-2 (CD74) immunopositivity as able to distinguish UPS from AFX and to predict those rare AFX likely to behave aggressively, although only a single prior study has been dedicated to evaluating this marker. We performed LN-2 staining of 14 AFX, 8 superficial UPS, and 65 deep UPS specimens using an identical protocol as described by prior authors. Of the 73 total UPS specimens, only 1 (1.4%) stained strongly with LN-2, as compared with 3 of 14 (21%) AFX (P = 0.012). One of 2 (50%) clinically aggressive AFX tumors that later exhibited both local recurrence and metastasis stained strongly for LN-2, whereas 2 of 12 (17%) of the more indolent tumors stained strongly with this marker (P = 0.40). Our data do not replicate prior reports of LN-2 as a sensitive and specific marker for UPS, or as indicative of prognosis for AFX, and therefore does not support the use of LN-2 as either a diagnostic or prognostic marker.

Clonal identity and differences in primary cutaneous B-cell lymphoma occurring at different sites or time points in the same patient.

Primary cutaneous B-cell lymphomas (PCBCL) are rare. Marginal zone lymphomas and follicle center lymphomas (FCL) represent a majority of these cases, and a significant number of cases present with multiple lesions. It is unclear whether multiple lesions in PCBCL represent dissemination of a single clone or multiple new primary lymphomas. In the current study, we analyzed paired samples from 20 PCBCL patients at more than 1 site (16) or at the same site at different time points (4) and 12 patients with benign lymphoid infiltrates to investigate for the presence or absence of a clone, and if present, whether the clones were identical. Both IGH@ and IGK@ rearrangements were tested using the BIOMED-2 protocol. We identified a clone (IGH@ and/or IGK@) in 19 of 20 (95%) PCBCL patients and 2 of 12 (17%) benign lymphoid infiltrate patients. The B-cell clones were proven to be identical in 11 of 20 (55%) PCBCL patients, including 7 of 16(44%) biopsies from patients with 2 different sites and 4 of 4 biopsies (100%) from patients at the same site but different time points. In 4 cases (3 FCL and 1 marginal zone lymphoma), different clones were detected at different sites, suggesting the possibility of a second simultaneous primary lymphoma. Our results indicate that the presence of identical clones is highly suggestive of lymphoma. To our knowledge, this is the first report to investigate the detection of identical clones in 2 distinct biopsies in PCBCL patients. Although the study is small and the results need to be confirmed in a larger study, these findings suggest that a subset of PCBCL at different sites may represent different primary tumors rather than occurrence of a single disease.

Discordant immunophenotypic profiles of adhesion molecules and cytokines in acute myeloid leukemia involving bone marrow and skin.

We investigated the role of adhesion molecules in skin involvement by acute myeloid leukemia (AML) using immunohistochemical analysis. Ten paired cases of skin and bone marrow biopsy specimens from patients with myeloid leukemia cutis (MLC) and 15 bone marrow biopsy specimens from patients without MLC were studied with antibodies directed against CD29, CD34, CD54, CD62-L, CD183, and cutaneous lymphocyte antigen (CLA). CLA was expressed in all cases of leukemia whereas CD54 was negative within blasts. CD62-L was expressed in 4 of 10 specimens of marrow infiltrates with MLC and 6 of 10 specimens of matching skin infiltrates; in marrows without MLC, only 2 of 15 were positive. CD29 was expressed in 1 of 10 marrow infiltrate specimens with MLC and 4 of 10 matching skin infiltrate specimens; in marrows without MLC, only 1 of 15 were positive. CD183 was expressed in 1 of 10 marrow infiltrate specimens with MLC and 4 of 10 matching skin infiltrate specimens; in marrows without MLC, CD183 was negative. The gain of CD62-L, CD29, and CD183 expression in bone marrow and skin infiltrates in leukemia cutis, relative to bone marrow infiltrates of cases without MLC, suggests a role for these markers in AML homing to skin.

Immunophenotypic analysis of myeloperoxidase-negative leukemia cutis and blastic plasmacytoid dendritic cell neoplasm.

Myeloid leukemia cutis (LC) and blastic plasmacytoid dendritic cell neoplasm (BPDCN) are morphologically indistinguishable malignancies that frequently manifest in the skin. Separating myeloperoxidase-negative LC from BPDCN may be particularly challenging. We identified a panel of immunohistochemical stains to distinguish myeloid LC (23 cases) from BPDCN (12 cases): myeloperoxidase, which stained 7 cases (30%) of LC and 0 cases (0%) of BPDCN; CD56, which stained 12 cases (52%) of LC and all 12 cases (100%) of BPDCN; CD4, which stained 2 cases (9%) of LC and all 12 cases (100%) of BPDCN; CD123, which stained 4 cases (17%) of LC and 10 cases (83%) of BPDCN; and Tcl-1, which stained 2 cases (9%) of LC and 9 (82%) of 11 cases of BPDCN. It is interesting that CD33 was not helpful; it stained 18 (78%) cases of LC and 11 cases (92%) of BPDCN. Our results indicate that a panel that includes CD4, CD56, CD123, and Tcl-1 can appropriately distinguish between these 2 entities.

A case of sebaceous carcinoma diagnosed in an adolescent male.

Sebaceous carcinoma is an uncommon and potentially aggressive malignancy that exhibits sebaceous differentiation. Approximately 75% of cases arise in the periocular region. Sebaceous carcinoma is rare in the pediatric population and its presentation in this age group is not well documented in the dermatopathology literature. We report the case of a 15-year-old male with sebaceous carcinoma who was first seen with a nodular lesion involving the skin of the left orbit/temporal area. A shave biopsy was performed which showed an infiltrative proliferation of basaloid cells that focally exhibited sebaceous differentiation, including the formation of incipient sebocytes. Immunohistochemically, the tumor cells expressed epithelial membrane antigen (EMA) and CK5/6, while a lack of Ber-EP4 was observed. Based upon these attributes, the diagnosis of sebaceous carcinoma was rendered. Subsequent immunohistochemical analysis for a possible DNA mismatch repair enzyme defect revealed that all four mismatch repair gene products showed retained expression, thereby providing no support for the presence of underlying Muir-Torre syndrome. Sebaceous carcinomas are exceptional in the pediatric age group and are rarely documented in the dermatopathology literature. Knowledge that this adult carcinoma can occur mostly in the pediatric age group may aid in the recognition of this uncommon malignancy.

CD163 expression is present in cutaneous histiocytomas but not in atypical fibroxanthomas.

CD163, a hemoglobin scavenger receptor, is expressed in monocytes and macrophages. Recent work has shown that this marker is specific for neoplasms of histiocytic differentiation. Our aim was to test the ability of CD163 to separate cutaneous histiocytomas from their morphologic mimics. We tested the expression of CD163 in 78 cases, including 19 xanthogranulomas, 16 atypical fibroxanthomas, 6 reticulohistiocytomas, 8 epithelioid cell histiocytomas, 9 cases of Langerhans cell histiocytosis, 10 xanthomas, and 10 intradermal Spitz nevi. CD163 expression was seen in all xanthogranulomas and reticulohistiocytomas, 4 epithelioid cell histiocytomas, 2 cases of Langerhans cell histiocytosis, and 8 xanthomas but was absent in atypical fibroxanthomas and Spitz nevi. CD163 is an excellent marker for confirming histiocytic differentiation and is useful in eliminating morphologic mimics such as Spitz nevi from the differential diagnosis. The lack of CD163 in atypical fibroxanthomas argues against a histiocytic origin for this tumor.

An updated approach to the diagnosis of myeloid leukemia cutis.

The diagnosis of myeloid leukemia cutis can be difficult, particularly in the context of an initial skin biopsy with a malignant hematopoietic neoplasm. We studied the immunohistochemical characteristics of 33 cases of myeloid leukemia cutis diagnosed at Stanford University Medical Center, Stanford, CA, 1996-2007, and compared them with the corresponding bone marrow blast immunophenotype and World Health Organization classification (2008). In the skin, CD43 marked 97% of cases (32/33), myeloperoxidase marked 42% (14/33), CD68 marked 94% (31/33), CD163 marked 25% (7/28), and CD56 marked 47% (14/30). CD34 and CD117 were predominantly negative. In 19 cases in which myeloperoxidase was negative, all marked with CD68 and CD43. The flow cytometric immunophenotype of the leukemic blasts in the bone marrow was discordant with the immunohistochemical profile in the skin in all cases, showing loss or gain of at least 1 antigen. Given the immunophenotypic differences between skin and bone marrow blasts, we provide an updated immunohistochemical approach to the diagnosis of myeloid leukemia cutis.

Cutaneous peripheral T-cell lymphoma associated with a proliferation of B cells.

Although the new World Health Organization-European Organization for Research and Treatment of Cancer classification focuses on providing uniformity in the diagnosis of cutaneous lymphomas, cutaneous peripheral T-cell lymphoma (PTL) remains a poorly defined subgroup. As follow-up to a study of systemic PTL complicated by a proliferation of B cells, we studied 16 cases of cutaneous PTL that contained morphologically atypical T cells associated with a significant infiltrate of B cells (about 20%-50%). A clonal T-cell receptor gamma chain gene rearrangement was present in all cases. In contrast, a clonal immunoglobulin heavy chain gene rearrangement was present in only 1 case. Clinical staging in 14 cases identified systemic involvement in 2. At last follow-up, both patients with systemic involvement had died of disease, and the majority of patients with primary cutaneous disease were alive (11/12). The presence of numerous atypical B cells and T cells caused diagnostic confusion in these cases. Comprehensive pathologic studies, coupled with clinical staging, are necessary for the accurate diagnosis of this unusual manifestation of cutaneous PTL.

A malignant cutaneous neuroendocrine tumor with features of Merkel cell carcinoma and differentiating neuroblastoma.

We report an unusual primary cutaneous neuroendocrine carcinoma that shows histologic and immunohistochemical features of ganglioneuroblastoma/differentiating neuroblastoma. The neoplasm is composed predominantly of small atypical neoplastic cells embedded in distinct clusters of immature and mature ganglion cells with associated neuropil. The neoplastic cells show strong perinuclear staining for cytokeratin 20 (CK20) with a dot-like pattern, supporting our contention that this is an unusual variant of Merkel cell carcinoma. To the best of our knowledge, ganglioneuroblastoma-like differentiation has not been previously described in Merkel cell carcinoma.

Evaluation of B-cell clonality using the BIOMED-2 PCR method effectively distinguishes cutaneous B-cell lymphoma from benign lymphoid infiltrates.

Primary cutaneous B-cell lymphomas (CBCL) are a diverse group of lymphomas that are limited to the skin at the time of diagnosis. Recently, standardized polymerase chain reaction protocols for immunoglobulin (Ig) rearrangement in nodal malignancies using the BIOMED-2 method have been studied extensively. However, reports of investigations of Ig clonality in CBCL using the BIOMED-2 method have been scant. We hypothesized that clonality detection in CBCL with the BIOMED-2 method could effectively distinguish malignant from benign B-cell-rich infiltrates in the skin. Formalin-fixed tissue samples from 26 patients with CBCL and 23 with benign lymphoid infiltrates were analyzed for Ig clonality using standardized BIOMED-2 polymerase chain reaction protocols. The (14;18) translocation was also assessed. A clone was detected in 22 (85%) of the 26 patients with CBCL [12/15 (80%) marginal zone B-cell lymphoma; 10/11 (91%) follicle center lymphoma] and in 1 (4%) of the 23 patients with benign infiltrates. The (14;18) translocation was present in 3 (12%) of the 26 patients with CBCL [1/15 (7%) marginal zone B-cell lymphoma; 2/11 (18%) follicle center lymphoma]. Our preliminary data indicate that Ig clonality can be detected in formalin-fixed samples of CBCL with meaningful sensitivity (85%) and high specificity (96%) using the BIOMED-2 method. This study forms the basis for further investigating the role of Ig clonality in distinguishing CBCL from benign lymphoid infiltrates that may pose a challenge in morphologic diagnosis.

CD34+ pigmented fibrous proliferations: the morphologic overlap between pigmented dermatofibromas and Bednar tumors.

Pigmented dermatofibrosarcoma protuberans (DFSP; Bednar tumor) constitutes 5%-10% of all cases of DFSP and shows morphologic features that overlap with melanocytic and fibrous proliferations. We report 2 unusual cases of pigmented fibrous proliferations that demonstrate features of dermatofibromas and DFSP. The first case is that of a 19-year-old man with a 3-year history of a slowly growing pigmented lesion on the right arm. On clinical exam, the lesion was a 7-mm firm pigmented papulonodular lesion. The second case is that of a 31-year-old woman with a 4- to 5-year history of a slowly enlarging, asymptomatic "dark area" on the right buttock. On clinical exam, the lesion was a 2-cm darkly pigmented flat nodule. Morphologically, both lesions are primarily dermal proliferations of spindled cells admixed with pigmented dendritic melanocytes. The lesional cells trap collagen fibers at the periphery and there is basal cell hyperpigmentation. Adnexal structures are effaced, but significant trapping of subcutaneous fat is not present. By immunohistochemistry, both lesions show focal CD34 positivity but are negative for Factor XIIIa and melanocytic markers. Although overlap between standard dermatofibromas and DFSP is well documented in the literature, pigmented fibrous lesions with features of both entities are not well described.

Silent mutations in KIT and PDGFRA and coexpression of receptors with SCF and PDGFA in Merkel cell carcinoma: implications for tyrosine kinase-based tumorigenesis.

Merkel cell carcinoma is a rare and aggressive form of skin cancer of neuroendocrine origin. Its treatment involves wide excision and radiotherapy but no effective therapy exists for advanced disease. Upregulation of the platelet-derived growth factor receptor family of tyrosine kinases, PDGFRA and KIT, has a crucial role in cancer development. Several studies have shown expression of the tyrosine kinase receptor KIT (CD117) in Merkel cell carcinoma. In this study, we examined the expression and mutational status of KIT and PDGFRA in 14 primary and 18 metastatic Merkel cell carcinoma. The expression of KIT and PDGFRA and their respective ligands, stem cell factor (SCF) and PDGFA, was assessed by immunohistochemistry. In addition, we analyzed KIT exons 9, 11, 13 and 17, and PDGFRA exons 10, 12 and 18 for the presence of activating mutations. We found that only 53% of cases of Merkel cell carcinoma expressed KIT, which was mostly seen as diffuse weak staining, and SCF expression was observed only in 31% of cases. In contrast, 87 and 81% of cases expressed PDGFRA and PDGFA, respectively. We observed coexpression of SCF and KIT in only 5 of 32 cases (16%) whereas 25 of 31 cases (81%) showed coexpression of PDGFRA and its ligand PDGFA. While we documented silent mutations in exon 17 of KIT and exons 10, 12 and 18 of PDGFRA, we were not able to identify any known activating mutations. Our results indicate that there is no correlation between positive immunostaining and occurrence of activating mutations in KIT and PDGFRA. Moreover, the presence of KIT/SCF and PDGFRA/PDGFA coexpression in a proportion of cases may indicate an autocrine/paracrine stimulation loop. We think therefore that imatinib mesylate is less likely to be an effective therapy for Merkel cell carcinoma, unless activating mutations exist in other exons of these receptor kinases.