The clinical, histologic, immunophenotypic, ultrastructural, and molecular features of a distinctive gastrointestinal tumor are described. Sixteen patients, 8 women and 8 men aged 17 to 77 years (mean age, 42 y; 63% less than 40 y) presented with abdominal pain, intestinal obstruction, and an abdominal mass. Mean tumor size was 5.2 cm (range, 2.4 to 15.0 cm). The tumors arose in the small bowel (10), stomach (4), and colon (2) and were histologically characterized by a sheet-like or nested population of epithelioid or oval-to-spindle cells with small nucleoli and scattered mitoses. Five cases showed focal clearing of the cytoplasm. Scattered osteoclast-type multinucleated giant cells were present in 8 cases. The tumor cells were positive for S-100 protein, SOX10, and vimentin in 100% of cases, for CD56 in 70%, for synaptophysin in 56%, for NB84 in 50%, for NSE in 45%, and for neurofilament protein in 14% of cases. All cases tested were negative for specific melanocytic, gastrointestinal stromal tumors, epithelial, and myoid markers. Ultrastructural examination of 5 cases showed features of primitive neuroectodermal cells with clear secretory vesicles, dense-core granules, occasional gap junctions, and no evidence of melanogenesis. EWSR1 gene rearrangement was assessed by fluorescence in situ hybridization in 14 cases. Twelve cases (86%) showed split EWSR1 signal consistent with a chromosomal translocation involving EWSR1. One case showed extra intact signals, indicating that the nuclei possessed either extra copies of the EWSR1 gene or chromosome 22 polysomy. Only 1 case showed no involvement of the EWSR1 gene. Six cases demonstrated rearrangement of the partner fusion gene ATF1 (46%), and 3 showed rearrangement of CREB1 (23%); 2 cases lacked rearrangement of either partner gene. Clinical follow-up was available in 12 patients and ranged from 1.5 to 106 months. Six patients died of their tumors (mean survival, 32 mo; 83% less than 24 mo). At last follow-up, 4 patients were alive with regional, lymph node, and liver metastases, and 2 patients were alive with no evidence of disease. The tumor described here is an aggressive form of neuroectodermal tumor that should be separated from other primitive epithelioid and spindle cell tumors of the gastrointestinal tract. The distinctive ultrastructural features and absence of melanocytic differentiation serve to separate them from soft tissue clear cell sarcomas involving the gastrointestinal tract. The designation "malignant gastrointestinal neuroectodermal tumor" is proposed for this tumor type.

The distinction of metastatic carcinomas to the skin from poorly differentiated primary cutaneous carcinomas and sometimes primary benign adnexal tumors can pose a significant diagnostic challenge. The purpose of this study was to evaluate the role of p63 and podoplanin (D2-40) immunoreactivity for separating primary skin tumors vs. cutaneous metastases of carcinomas from internal organs. Thirty seven primary tumors and 42 cutaneous metastatic adenocarcinomas were evaluated. The 37 primary cutaneous tumors included 14 cases of benign adnexal tumors, 9 malignant skin adnexal neoplasms, and 14 primary squamous and basal cell carcinomas. The 42 metastatic adenocarcinomas all corresponded to metastases from patients with a well-documented history of a primary tumor at another location. We found variable positivity with podoplanin in all primary cutaneous neoplasms including spiradenoma (6/6), hidradenoma (2/4), cylindroma (3/3), desmoplastic trichilemmoma (1/1), poorly differentiated squamous cell carcinoma (4/4), sebaceous carcinoma (1/1), basal cell carcinoma (4/10), trichilemmal carcinoma (2/2), eccrine carcinoma (3/3), microcystic adnexal carcinoma (1/1), adnexal carcinoma NOS (1/1), and porocarcinoma (1/1). In contrast, all metastatic carcinomas were negative (0/42) for podoplanin. In regards to p63, all cases of primary cutaneous tumors were positive for p63 (37/37); in contrast, all cutaneous metastatic carcinomas were negative (0/42). Sensitivity, specificity, and positive and negative predictive values of podoplanin and p63 immunoreactivity to separate primary skin neoplasms from metastatic carcinomas were 78.4, 100.0, 100.0 and 84.0% for podoplanin, respectively, and 100.0, 100.0, 100.0 and 100.0% for p63, respectively. The differences in p63 and podoplanin immunohistochemical expression between primary skin tumors and metastatic carcinomas to the skin were statistically significant (p < 0, 0001). The results of our study suggest that the combined expression of p63 and podoplanin are a useful adjunct for the diagnosis of skin tumors in the clinical setting of a questionable metastasis and may be relatively specific for distinguishing primary skin tumors from metastatic carcinomas to the skin.

Epithelioid sarcoma is a rare, aggressive keratin-positive sarcoma that co-expresses CD34 in 50% of cases and may mimic an angiosarcoma. Recently, we have observed one case of epithelioid sarcoma that labeled for ERG, an ETS family regulatory transcription factor, which is considered to be a reliable marker for vascular differentiation. We investigated the prevalence of nuclear expression of ERG and FLI1, a homologous transcription factor, in these tumors. A formalin-fixed paraffin-embedded tissue microarray of 37 epithelioid sarcomas was examined. Immunohistochemistry was performed using anti-ERG monoclonal antibody to the N-terminus, anti-ERG monoclonal antibody to the C-terminus and anti-FLI1 monoclonal antibody. Comparison was made with CD34, CD31, and D2-40 labeling. The extent of immunoreactivity was graded according to the percentage of positive tumor cell nuclei (0: no staining; 1 þ : o5%; 2 þ : 5-25%; 3 þ : 26-50%; 4 þ : 51-75%; and 5 þ : 76-100%), and the intensity of staining was graded as weak, moderate, or strong. Nuclear staining for the N-terminus of ERG was seen in 19 out of 28 cases: 10 with diffuse(4 to 5 þ ) strong/moderate labeling; 1 with 2 þ moderate labeling and 8 with weak labeling (1 to 4 þ , 2 each). Focal staining for the C-terminus of ERG was seen in only 1 out of 29 cases (2 þ moderate). FLI1 labeling was seen in nearly all (28 out of 30) cases: 16 with diffuse (5 þ ) predominantly moderate labeling, and 8 cases with diffuse(5 þ ) weak labeling. The remainder had variable moderate (1 to 3 þ ) or weak (1 to 4 þ ) FLI1 staining. CD34 was positive in 22 out of 30 cases and D2-40 was found to be positive in 22 out of 31 cases. All cases were negative for CD31 (0 out of 30). Epithelioid sarcoma can label with antibodies to the N-terminus of ERG, FLI1, and D2-40, which may cause diagnostic confusion for a vascular tumor. A panel of other antibodies including SMARCB1 and CD31 should be used in evaluating these tumors. ERG antibody selection is also critical, as those directed against the C-terminus are less likely to label epithelioid sarcoma. Modern Pathology (2014) 27, 496-501; doi:10.1038/modpathol.2013.161; published online 27 September 2013 Keywords: epithelioid sarcoma; ERG; FLI1 ERG and FLI1, members of the ETS family of transcription factors, are markers of endothelial differentiation. The utility of ERG and FLI1 in distinguishing cutaneous and non-cutaneous angiosarcomas from other histologic mimics has been previously described. [1][2][3][4] Recently, a recurrent epithelioid sarcoma at our institution showed histologic features suggestive of vascular differentiation: hemorrhagic at low power, discohesive sheets of large atypical epithelioid cells, and areas of hobnailed-appearing cells in blood-filled cystic spaces mimicking epithelioid angiosarcoma. (Figure 1) Immunohistochemical studies showed the tumor to be reactive for keratin, CD34, and ERG but negative for CD31; SMARCB1 expression was lost. Initial biopsies showed central geographic necrosis with palisading tumor cells, cha...

Primary cutaneous large B-cell lymphomas (PCLBCL) have historically been a matter of debate in the literature. The 2005 World Health Organization-European Organization for Research and Treatment of Cancer (WHO-EORTC) classification scheme segregated cutaneous B-cell lymphomas into 3 groups: primary cutaneous marginal zone B-cell lymphoma, primary cutaneous follicle center cell lymphoma, and primary cutaneous diffuse large B-cell lymphoma (PCDLBCL), "leg type" (PCDLBCL-LT). Additionally, the WHO-EORTC classification scheme utilized the term PCLBCL "other" not otherwise specified (NOS) type for rare cases of PCLBCL not belonging to either the "leg type" or the primary cutaneous follicle center cell lymphoma group. In this study, we retrospectively assessed the histomorphologic features of 79 cases of PCDLBCLs, including those of "leg type" and "other" NOS type, to further categorize the histologic spectrum of these unusual cutaneous neoplasms. The histologic diagnosis of PCLBCL usually poses little diagnostic difficulty; however, some cases may adopt unusual or unfamiliar appearances mimicking other lymphoproliferative disorders or other malignant neoplasms. Seventy-nine cases, occurring in 37 men and 42 women, aged 34-94 years, were analyzed. Fifty-three cases were classified as "leg type" and 26 cases were classified as "other" NOS type using the WHO-EORTC classification. Of the 53 cases classified as "leg type," 33 were women and 20 were men; of the 26 cases of "other" NOS type, 9 were women and 17 were men. In the "leg type" category, 31 cases were located on the lower extremities, 5 cases on the face, 5 cases on the arm, 3 cases on the chest, 2 cases on the shoulder, 2 cases on the back, 1 case on the trunk, 1 case on the buttock, 1 case on the supraclavicular area, 1 case on the head, and 1 case on the flank. Of the "other" NOS type category, 8 cases were located on the face, 5 cases on the shoulder, 3 cases on the head, 2 cases on the abdomen, 2 cases on the chest, 1 case on the trunk, 1 on the vulva, 1 on the axilla, 1 on the back, 1 on the neck, and 1 on the hip. Most cases assessed showed the classic morphologic appearance of PCDLBCL, but cases mimicking Burkitt lymphoma (starry-sky pattern), natural killer-cell (NK) lymphoma, mycosis fungoides (epidermotropism), low-grade B-cell lymphomas, epithelial malignancies, and Merkel cell carcinoma were encountered in this series. The high frequency of these rare histologic patterns can be explained by a bias associated with consultation practice. Careful histologic examination and immunohistochemical stains were used to establish the correct diagnosis. The differential diagnosis of PCDLBCL is broad and difficult to define histologically. Knowledge of these rare histologic variants is necessary to avoid misinterpretation of these cases as nonlymphoid malignancies.

No abstract