Search

CD34-negative solitary fibrous tumors (SFTs) are rare and have not been comprehensively studied. We retrospectively reviewed all cases of SFT confirmed with STAT6 immunohistochemistry and/or STAT6 gene fusion between 2013 and 2020 and collected pertinent clinicopathologic parameters. Of a total of 244 cases, 25 (10%) lacked CD34 expression by immunohistochemistry. Compared with CD34-positive SFT, CD34-negative SFT are more likely to arise in the head and neck area (32% vs. 24%, P=0.02) and present as metastatic disease (28% vs. 1%, P<0.0001). A significantly higher percentage of CD34-negative SFT exhibit high-grade cytologic atypia (hypercellularity, round cell or anaplastic morphology, nuclear pleomorphism, etc.) (48% vs. 22%, P=0.0073). There are no significant differences in the distributions of age, sex, tumor size, mitotic count, tumor necrosis, or risk stratification between CD34-negative and CD34-positive SFT. In addition, only 56% of CD34-negative SFT display a typical hemangiopericytoma-like vascular pattern. Special histologic features among CD34-negative SFT include prominent alternating hypercellular or fibrous and hypocellular myxoid areas with curvilinear vessels mimicking low-grade fibromyxoid sarcoma, pulmonary edema-like microcystic changes, and prominent amianthoid collagen fibers. In conclusion, compared with their CD34-positive counterparts, CD34-negative SFT is more likely to present as metastatic disease, show high-grade nuclear atypia, and lack the characteristic hemangiopericytoma-like vasculature, posing a unique diagnostic challenge. The use of STAT6 immunohistochemistry and/or molecular studies may be prudent in soft tissue tumors that appear CD34 negative and lack conventional SFT histopathologic characteristics., Competing Interests: Conflicts of Interest and Source of Funding: The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article., (Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.)

Undifferentiated carcinoma of the esophagus and gastroesophageal junction is a recently recognized entity in the fifth edition of the World Health Organization Classification of Digestive Tumors and is diagnostically challenging, particularly on small biopsies. SMARCA4 and SMARCA2 are chromatin remodeling genes with key roles in oncogenesis. We retrieved 14 cases of SMARCA4/SMARCA2-deficient undifferentiated carcinoma of the gastroesophageal junction and esophagus from the authors' institutions. The tumors showed similar histologic findings: the sheet-like proliferation of tumor cells characterized by discohesion, large nuclei, and prominent macronucleoli with many tumor cells exhibiting a rhabdoid appearance. In 8 cases, adjacent specialized intestinal metaplasia was noted and 3 cases exhibited adjacent high-grade dysplasia. Immunohistochemically, tumors variably expressed keratins and disclosed loss of expression of SMARCA4 in 12 and SMARCA2 in 7 cases. In 2 cases SMARCA2 alone was lost without SMARCA4 loss. A mutant p53 immunohistochemical pattern was seen in 4 of 4 cases, 3 of which showed diffuse, strong nuclear expression, and 1 case displayed a complete loss of nuclear expression of p53, including invasive carcinoma and associated dysplasia, when present. Limited clinical follow-up was available, but 3 patients died of disease within 0.6, 2, and 7 months of diagnosis. We present the first series of undifferentiated carcinoma of the esophagus and gastroesophageal junction with this characteristic morphology associated with loss of SMARCA4 and/or SMARCA2 expression. This tumor type likely arises from dedifferentiation of a lower grade carcinoma in some cases, and Barrett esophagus and appears to be associated with an aggressive clinical course., Competing Interests: Conflicts of Interest and Source of Funding: The authors have disclosed that they have no significant relationships with, or financial interest in, any commercial companies pertaining to this article., (Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.)

Metastatic invasive lobular carcinoma (mILC) may masquerade as primary diffuse gastric adenocarcinoma (PDGA) by demonstrating significant clinical and pathologic overlap. Accurate distinction is of therapeutic and prognostic significance. On the basis of anecdotal cases of mILC that lacked estrogen receptor and/or GATA3 expression, we analyzed the cytoarchitectural features of 28 mILC and 44 PDGA specimens obtained from women to assess features that would help in this distinction and prompt ancillary work-up. In addition to performing an interobserver agreement analysis among 3 pathologists, we also evaluated SATB2 expression in this setting. Eighteen of 20 (90%) patients had a history of ILC. The mean interval between initial diagnosis of breast cancer and metastasis was 7.3 years (range: 1 to 36 y). Compared with mILC, PDGA was significantly associated with full-thickness mucosal involvement (47% vs. 80%; P=0.015), a nested/sheet-like growth pattern (32% vs. 68%; P=0.004), anastomosing cords (0% vs. 100%; P=0.001), multivacuolated cells (0% vs. 61%; P<0.0001), pleomorphic nuclei (4% vs. 70%; P<0.0001) and enlarged nuclei (4% vs. 70%; P<0.0001). Single file growth pattern (P<0.0001) and superficial lamina propria involvement (P=0.009) were more common in mILC. Estrogen receptor and GATA3 were expressed in all but 5 mILC cases; SATB2 was only seen in 30% of PDGA cases. Our results demonstrate that in a biopsy specimen, careful morphologic assessment can be extremely helpful in distinguishing mILC from PDGA and guiding ancillary work-up, especially when a history of breast cancer may not be readily available or when the neoplasm lacks expression of conventional breast markers.

We recently encountered a case of primary pulmonary angiomatoid fibrous histiocytoma (AFH), which was initially misdiagnosed as inflammatory myofibroblastic tumor (IMT) based in part on anaplastic lymphoma kinase (ALK) expression by immunohistochemistry (IHC). Prompted by this experience, we evaluated ALK expression in 11 AFH, 15 IMT, and 11 follicular dendritic cell sarcomas using 3 different antibody clones: D5F3, 5A4, and ALK1. ALK IHC positive cases were analyzed with fluorescence in situ hybridization (FISH) using dual color ALK break-apart probe kit. The majority of AFH cases studied were positive for ALK IHC with at least 1 antibody (9/11 D5F3, 6/9 5A4, 1/9 ALK1), most demonstrating moderate to strong cytoplasmic staining. AFH with positive ALK IHC showed no ALK gene rearrangement by FISH (0/8) with ALK copy number ranging from 1.6 to 2.1. Sixty-seven percent of IMT were ALK positive by IHC (10/15 D5F3, 8/15 5A4, 7/15 ALK1), and 9 of the 10 cases were positive for ALK gene rearrangement by FISH. All follicular dendritic cell sarcomas were negative for ALK by IHC (D5F3 and 5A4). Our results indicate that ALK expression in AFH is common, particularly with the highly sensitive D5F3 and 5A4 antibodies and enhanced detection systems, and may be a potential source of diagnostic confusion with IMT. The underlying mechanism of ALK expression in AFH is unclear, although it does not appear to be from ALK rearrangement or amplification.

While superficial (cutaneous/subcutaneous) solitary fibrous tumor (SFT) have been described, definitive diagnosis is difficult due to overlapping features with other tumors. We describe the largest series to date of superficial SFT. For inclusion, SFT had to arise in dermis or subcutis. Twenty-six cases were identified. Patients ranged from 16 to 80 years (mean, 47 y) with a marked female predominance (19 F; 7 M). Tumors involved the head (11), thigh (7), back (3), shoulder (2), upper arm (1), ankle (1), and great toe (1). Mean size was 2.9 cm (range, 1.0 to 7.0 cm). The majority (n=19) had typical histologic features (cellular SFT) with irregular fascicles of spindled cells, staghorn-like blood vessels, and variable amounts of collagen. Necrosis was evident in 3 cases (all <25%). Mitotic activity ranged from 0 to 10 mitotic figures/10 high-power fields (mean, 2 mitotic figures/10 high-power fields). Seventeen of the 18 were positive for STAT6, whereas 21/22 expressed CD34. All were low risk (23/23) by proposed criteria (Demicco and colleagues), including 2 cases with malignant histology. Three could not be risk stratified due to lack of information on tumor size. Follow-up, available on 7 cases, showed no recurrence/metastasis (mean follow-up, 100 mo; range, 2 to 241 mo). Cutaneous SFT are more common in women and most often involve the head. They are usually low risk by proposed criteria and appear to behave in an indolent manner, though larger studies are needed to confirm this. Recognition that SFT may present as a superficial mass will avoid misclassification as other CD34-positive neoplasms that frequently arise in the skin and subcutaneous tissue.

Spindle cell lipomas (SCL) are typically tumors of the upper back/neck (shawl region) of men (80% to 90%). In general, there is a frequent tendency to restrict the diagnosis to this specific clinical scenario and a hesitancy to diagnose SCL in women. We hypothesized that SCL in women have a more varied presentation. A total of 395 SCL were diagnosed at our institution over the last 11 years. The diagnosis of SCL in women was confirmed by re-review. Immunohistochemical stains for CD34, desmin, estrogen receptor, and p16 were performed. In a subset, fluorescence in situ hybridization to detect Retinoblastoma1 (RB1) gene deletion was performed. Of 395 SCLs, 331 (86%) occurred in men; 53 (14%) occurred in women (11 cases excluded). Of the 64 SCL in women, 58 had available material. In total, 53 of 58 were confirmed as SCL. Women were younger at diagnosis (median, 51 y; range, 5 to 76 y) compared with men (64 y; range, 23 to 98 y), P<0.0001, t test. SCL in women typically occurred outside the shawl distribution (36/53, 68%) compared with men (95/331, 29%) (P<0.001), including extremities (16/53, 30% vs. 32/331, 10%) and face (11/53, 21% vs. 47/331, 14%). Dermal SCL in women were also relatively common (16/53, 30%). The cases demonstrated varying proportions of bland spindled cells, ropey collagen, myxoid matrix, and adipocytes. By immunohistochemistry, 46/46 were CD34, 48 of 48 were desmin negative, 33 of 42 were estrogen receptor negative, and 29 of 42 had loss of p16 expression. In total, 12 of 14 showed RB1 loss by fluorescence in situ hybridization. SCL in women frequently occurs in unconventional locations and in at a slightly younger patient age.

Of the serrated polyps, the origin, morphologic features, molecular alterations, and natural history of traditional serrated adenomas (TSAs) are the least understood. Recent studies suggest that these polyps may arise from precursor lesions. The frequencies of KRAS and BRAF mutations vary between these studies, and only 1 small study has measured CpG island methylation using current markers of methylation. Mutations in GNAS, a gene commonly mutated in colorectal villous adenomas, have not been fully evaluated in TSAs. Finally, the expression of annexin A10 (ANXA10), a recently discovered marker of sessile serrated adenomas/polyps, has not been studied in these polyps. To further characterize these polyps, 5 gastrointestinal pathologists reviewed 55 left-sided polyps diagnosed as TSA at a single institution. Pathologists assessed various histologic features including cytoplasmic eosinophilia, ectopic crypt foci, presence of conventional dysplasia, and presence of precursor serrated lesions. KRAS, BRAF, and GNAS mutational analysis was performed, as well as CpG island methylation and ANXA10 immunohistochemistry. Ectopic crypt foci were seen in 62% of TSAs. Precursor lesions were seen in 24% of the study polyps, most of which were hyperplastic polyps. KRAS and BRAF mutations were common and were present in 42% and 48% of polyps, respectively. GNAS mutations occurred in 8% of polyps, often in conjunction with a BRAF mutation. Unlike sessile serrated adenomas/polyps, TSAs rarely had diffuse expression of ANXA10. Importantly, BRAF-mutated TSAs had more widespread methylation of a 5-marker CpG island panel compared with KRAS-mutated polyps. However, ectopic crypt foci, a proposed defining feature of TSA, were not associated with any specific molecular alteration.

Pleomorphic hyalinizing angiectatic tumor (PHAT) is a rare, locally aggressive tumor of the distal extremities with a proclivity for local recurrence. PHATs contain characteristic ectatic, thin-walled vessels, lined by fibrin, and are surrounded by groups of variably pleomorphic spindled to epithelioid neoplastic cells. The putative precursor lesion of PHAT, originally termed "early PHAT" shares many clinicopathologic features with hemosiderotic fibrolipomatous tumor (HFLT). HFLT, myxoinflammatory fibroblastic sarcoma (MIFS), and tumors showing hybrid features of HFLT and MIFS often show TGFBR3 and MGEA5 gene rearrangements. To date, only a small number of PHATs has been tested for either rearrangement; all have been negative. We hypothesized that PHATs contain TGFBR3 and/or MGEA5 rearrangements. Cases of PHAT (all containing areas of HFLT) (N=10), HFLT (N=7), MIFS (N=6), hybrid HFLT/MIFS (N=3), and PHAT-like undifferentiated pleomorphic sarcomas (N=7) were retrieved from our institutional and consultation archives and analyzed for TGFBR3 and MGEA5 rearrangements using a break-apart probe strategy for FISH. Six of 10 PHATs harbored TGFBR3 and/or MGEA5 gene rearrangements: 4 cases had both TGFBR3 and MGEA5 rearrangements, and 2 cases contained MGEA5 rearrangements. Two of 7 HFLTs were positive: 1 case had a TGFBR3 rearrangement, and 1 case had an MGEA5 rearrangement. One of 6 MIFSs had an MGEA5 rearrangement. All 3 hybrid HFLT/MIFS cases were positive: 2 cases had both TGFBR3 and MGEA5 rearrangements, and 1 case had a TGFBR3 rearrangement. All PHAT-like undifferentiated pleomorphic sarcomas were negative. We report, for the first time, the presence of TGFBR3 and/or MGEA5 rearrangements in tumors showing mixed features of HFLT and PHAT. The presence of such rearrangements strongly suggests that HFLT is related to both PHAT and MIFS and that the latter 2 tumors may represent morphologic variants of a single, genetically defined entity in which only MIFS has acquired the capacity to metastasize.

Ossifying fibromyxoid tumor of soft parts (OFMT) is a rare soft tissue neoplasm of uncertain differentiation. Very recently recurrent rearrangements of the PHF1 gene have been reported in OFMT, including typical, atypical, and malignant variants. We sought to validate and extend these findings in a larger series of well-characterized OFMT, in particular malignant variants. Slides and blocks from 41 OFMT were retrieved, rereviewed, and classified as typical, atypical, and malignant using previously published criteria. Interphase fluorescence in situ hybridization (FISH) was performed on paraffin-embedded sections of each case using a break-apart probe strategy, with direct-labeled FISH probes designed from bacterial artificial chromosomes. The 41 tumors occurred in 23 men and 18 women with a mean age of 55 years and involved the head and neck, trunk, and upper and lower limbs. The tumors were classified as typical (n=14), atypical (n=6) and malignant (n=21). PHF1 rearrangements were detected in 20 of 41 cases (49%) including 43% typical, 50% atypical, and 52% malignant cases. The results of our study confirm previous findings, with PHF1 rearrangements present in nearly 50% of OFMT, including roughly similar percentages of typical, atypical, and malignant tumors. These results support our previous hypothesis that OFMT might represent a translocation-associated tumor, underscore the likely importance of PHF1 rearrangements in the pathogenesis of these lesions, confirm the relationship between typical and malignant OFMT, and suggest a role for PHF1 FISH in the diagnosis of morphologically challenging cases.

The presence of high-grade dysplasia (HGD) or villous component (VC) defines an advanced adenoma (AA) in patients with 1 or 2 adenomas <1 cm in size. Current consensus guidelines recommend that patients with AA undergo more intense postpolypectomy surveillance. In these clinical situations, the interobserver reliability in determining VC and HGD would play a major role in the credibility of these consensus guidelines. Therefore, the purpose of this study was to evaluate interobserver variability of VC and HGD in polyps <1 cm before and after the development of consensus criteria among gastrointestinal (GI) pathologists. Five GI pathologists independently evaluated 107 colorectal adenomas <1 cm, and classified them into tubular adenomas or adenomas with a VC (A-VC) and into low-grade dysplasia or HGD. Then a consensus conference was held and consensus criteria for VC and HGD were developed by group review. The same set of 107 slides were rereviewed independently by the same 5 GI pathologists. Interobserver variability using κ statistical analysis before and after the application of consensus criteria was assessed. A 1-sided z-test was used to determine whether κ scores increased after the consensus conference. Interobserver agreement before and after the consensus conference was poor for assessment of A-VC, HGD, and AA. These data calls into question the validity of basing clinical decisions on this distinction.

Paneth cells have been reported in colorectal adenomas and adenocarcinomas; however, the frequency of colonic Paneth cell-containing adenomas is unknown as are their clinicopathologic features. A total of 152 consecutive colorectal adenomas from 103 patients (57 males and 46 females) were reviewed. The frequency of Paneth cells in this cohort of adenomas was determined and correlated with patient demographics. Twenty-six adenomas (17.1%) from 22 (21.4%) patients harbored Paneth cells, which were not limited to the base of the crypts but aberrantly located throughout the crypts. Patient age, adenoma size, villous features, and grade of dysplasia were not different between these 2 groups. Not surprisingly, Paneth cell-containing adenomas were more likely to occur in the proximal colon (84.6% vs. 55.6%; P=0.006). There was a strong association between male sex and Paneth cell-containing adenomas, as 23 of 26 (88.5%) of these adenomas occurred in male individuals compared with 71 of 126 (56.3%) non-Paneth cell-containing adenomas (P=0.002). Upon review of an additional 460 adenomas from 200 patients with varying numbers of adenomas (68 with 1 adenoma, 68 with 2 adenomas, and 64 with 3 or more adenomas), the risk of harboring synchronous adenomas was associated with villous morphology, proximal location, and the presence of a Paneth cell-containing adenoma. Thus, the presence of a Paneth cell-containing adenoma may be a marker for increased risk of developing colorectal neoplasia.

The occurrence of adenocarcinoma after ileal pouch-anal anastomosis for ulcerative colitis (UC) is an infrequent but potentially lethal complication. Neither histomorphologic nor molecular features of pouch adenocarcinoma after ileal pouch-anal anastomosis have been fully investigated. We report the largest series of 12 pouch and peripouch adenocarcinomas and compared them with 58 randomly selected UC-associated adenocarcinomas. The mean age of patients with pouch/peripouch adenocarcinoma was 55.2 years (SD 14.8), which was not significantly different from that of controls (P=0.52). Pouch/peripouch adenocarcinoma and UC-associated adenocarcinoma had a comparable frequency of tumor-infiltrating lymphocytes, lack of dirty necrosis, mucin differentiation, signet ring cell differentiation, heterogeneity, and well differentiation (P>0.05 for all). Pouch/peripouch adenocarcinoma was more likely to show Crohn-like reaction compared with UC-associated adenocarcinoma (P=0.047). Loss of at least 1 mismatch repair protein was noted in 9% of pouch/peripouch adenocarcinomas and 9.6% of UC-related adenocarcinomas (P=1.0). There was no significant difference in the frequency of p53 overexpression (36.4% vs. 61.1%, P=0.184) or nuclear immunoreactivity for β-catenin (9% vs. 7.4%, P=0.99) in pouch/peripouch versus UC-associated adenocarcinomas, respectively. Pouch/peripouch and UC-associated adenocarcinoma had a comparable positive rate for CK7 (54.5% vs. 55.5%, P=0.99), CK20 (100% vs. 98.1%, P=0.99), and CDX2 (72.8% vs. 72.2%, P=0.99) by immunohistochemistry. In summary, pouch and peripouch adenocarcinoma can occur in patients without colorectal neoplasia and in those with idiopathic inflammatory bowel disease, can be potentially lethal, and has histomorphologic and molecular features similar to those of UC-associated adenocarcinoma.

Long-standing inflammatory bowel disease (IBD), either ulcerative colitis or Crohn disease, is associated with a high risk of developing colorectal adenocarcinoma (CAC). However, histomorphology of IBD-associated CAC has not been thoroughly examined, and it is unclear whether and how these patients should be screened for Lynch syndrome (LS). We evaluated the demographic and morphologic features of 108 IBD-associated CACs, including ulcerative colitis-associated (n = 95) and Crohn disease-associated CACs (n = 13), against 93 control cases of sporadic microsatellite-stable (MSS) CAC, 20 cases of sporadic microsatellite instability high (MSI-H) CAC, and 23 CAC cases of LS. The mean age of patients with IBD-associated CAC was 50 years, which was lower compared with the mean age of 63.7 years of the sporadic MSS controls and 76.5 years of the sporadic MSI-H group but not statistically different from that of the LS patients. Synchronous CACs were noted in 20.4% of the IBD patients and 13% of LS patients but in only 2.1% of the sporadic MSS controls and in none of the MSI-H patients. Right-sided CACs were significantly less frequent in the IBD group than in sporadic MSS controls, MSI-H group, and LS patients (P < 0.05 for all). In contrast to sporadic MSS CAC, IBD-associated CACs are characterized by lack of tumor necrosis, Crohn-like reaction, tumor histologic heterogeneity, the presence of mucin, and signet ring cell differentiation and tumor well differentiation. The histomorphologic similarity among IBD-associated and MSI-H tumors, either sporadic MSI-H or LS-related, is independent of MSI status. The young age of patients with IBD-associated CAC and the morphological similarities among IBD-associated, sporadic MSI-H, and LS-related CAC suggest that an age-based and morphology-based strategy before the screening test for LS may be less effective in IBD patients than in the non-IBD population.

Pleomorphic liposarcoma (PL) is an uncommon form of liposarcoma that rarely occurs in the skin and subcutis. As its behavior in this setting is incompletely characterized, we undertook a study of a series of superficial PLs, defined as those arising or based primarily in the dermis and/or subcutis without involvement of deep structures. In addition, MDM2 gene amplification, a diagnostic signature of well-differentiated/dedifferentiated liposarcoma (WDL/DL), was evaluated to address the recent observation that this gene is amplified within PL-like areas in DL. PLs were obtained from institutional and consultation files (n=29). Cases were evaluated with respect to age, sex, location (dermis, dermis and subcutis, subcutis), size, predominant pattern (pleomorphic spindled or epithelioid), extent of lipogenic differentiation, and tumor necrosis. MDM2 amplification was analyzed using FISH on formalin-fixed, paraffin-embedded material in 26 cases. Patients ranged in age from 5 to 93 years (M:F=1.4:1). Tumors were located on the extremity (n=15), trunk (n=7), and head and neck (n=7) and involved the dermis (n=4), dermis and subcutis (n=10), and subcutis (n=15). Tumor size ranged from 0.8 to 15 cm (median=2 cm). All were mitotically active high-grade sarcomas [FNCLCC grade 2 (n=23) or 3 (n=6)] with either a pleomorphic spindled (n=24) or an epithelioid pattern (n=5) with variable extent of lipogenic differentiation [<25% (n=15), 25% to 50% (n=9), >50% (n=5)]. Necrosis was present in 3 cases. MDM2 gene amplification was present in 3 of 26 cases. Follow-up information in 24 cases (range=1 to 192 mo; median=48 mo; mean=59 mo) revealed local recurrences (4/24) but no metastasis or death from disease. We conclude that cutaneous and subcutaneous PLs, despite their high grade, have a much more favorable outcome compared with their deep-seated counterparts, most likely attributed to their small size and superficial location. The low incidence of MDM2 gene amplification in our series indicates that most superficial PLs are unrelated to WDL/DL. PL likely evolves by way of more than 1 molecular pathway.

Composite intestinal adenoma and microcarcinoid is a rare intestinal neoplasm consisting of intermingled adenomatous and well-differentiated neuroendocrine components. A few case reports and small series have suggested an indolent clinical course for this entity. We reported 7 cases of composite intestinal adenoma-microcarcinoid, including their morphologic features and clinical follow-up, both in biopsy and resection specimens. We identified 7 cases of composite intestinal adenoma-microcarcinoid from our pathology database. Five were from the large intestine, and 2 were in the duodenum. Morphologically, all microcarcinoids exhibited low-grade cytologic atypia and were devoid of significant pleomorphism, necrosis, and mitotic activity. Among the 7 lesions, 6 had a lobular architecture with smooth borders and mucosa-confined microcarcinoids; none had neuroendocrine carcinoma in subsequent resections. However, 1 colonic case had carcinoid cells penetrating the muscularis mucosae into the submucosa with an infiltrative border, and the resection showed metastatic high-grade neuroendocrine carcinoma in 1 lymph node. Composite intestinal adenoma-microcarcinoid is extremely rare. Although composite mucosa-confined adenoma-microcarcinoid is likely to have an indolent behavior, submucosal invasion by the neuroendocrine component may be associated with aggressive behavior.

Carcinoma cuniculatum, a unique variant of well-differentiated squamous cell carcinoma, has been only rarely reported in the esophagus. We report 9 cases of esophageal carcinoma cuniculatum diagnosed on esophagectomy specimens in 7 men and 2 women during a 20-year period. All but 1 of the patients presented with persistent or progressive dysphagia. All patients had an esophageal mass or lesion on endoscopic examination. In 8 cases (88.8%), the tumor was located in the distal esophagus. Burrowing was noted on the tumor surface in 2 recent cases on macroscopic examination. All carcinomas were invasive either at the mucosa (n=2), submucosa (n=1), muscularis propria (n=4), adventitia (n=3), or adventitia and lung (n=1). All carcinomas demonstrated a common histologic pattern characterized by hyperkeratosis, acanthosis, dyskeratosis, abnormal keratinization, keratin-filled cyst/furrows, koilocyte-like cells, intraepithelial neutrophils, and focal cytologic atypia. In situ hybridization for human papillomavirus subtypes was negative in all 10 tumors tested. None of the cases showed lymph node metastasis. Two patients died postoperatively due to complications. The remaining patients were followed up for a median duration of 84 months (48 to 214 mo). During the follow-up period, 3 patients died 49, 66, and 214 months after esophagectomy at the ages of 66, 68, and 91 years, respectively; death in these 3 cases was not related to recurrence/metastases of esophageal cancer. Four patients were alive without disease at 48, 49, 84, and 87 months after curative resection. Our report identifies a common histomorphologic pattern of esophageal carcinoma cuniculatum and supports the fact that surgical resection of the tumor by esophagectomy provides long-term survival even in patients with T3 tumor.

Distinguishing Barrett esophagus with high-grade dysplasia (BE-HGD) from intramucosal and submucosal adenocarcinomas on biopsies is challenging, yet important, in the choice of therapy. The current study evaluates preresection biopsies from patients who underwent esophagectomy for at least BE-HGD, to compare the recently published histologic categories by the University of Michigan (UM) and Cleveland Clinic (CC), correlate preresection and final resection diagnosis, and identify histologic features in biopsies that might be predictive of adenocarcinoma on esophagectomy. A total of 112 cases with a consensus biopsy diagnosis (agreement by ≥4 of 7 gastrointestinal pathologists) were statistically analyzed to identify histologic features that predicted adenocarcinoma on resection. Applying the UM criteria to the biopsy series showed excellent agreement with the CC system (κ=0.86) and significant correlation between preoperative and esophagectomy diagnoses (P<0.001). The likelihood of finding carcinoma on resection was significantly higher with the category of HGD with marked glandular distortion cannot exclude intramucosal adenocarcinoma [CC; odd ratio (OR), 2.8; P=0.046] or HGD suspicious for adenocarcinoma (UM; OR, 4.3; P=0.008), compared to HGD alone. The presence of "never-ending" glands (OR, 3.7; P=0.008), sheet-like growth (P<0.001), angulated glands (OR, 8.5; P<0.001), ≥3 dilated glands with intraluminal debris (OR, 2.6; P=0.05), and >1 focus of single-cell infiltration into the lamina propria (OR, 8.9; P<0.001) increased the odds of finding carcinoma on resection. The latter 2 variables remained independent predictors of adenocarcinoma in multivariable analysis. In conclusion, the CC and UM systems show excellent agreement and define histologic categories that can improve prediction of adenocarcinoma on resection.

Colonic perineuriomas are recently described benign mucosal polyps that are composed of a bland spindle cell proliferation surrounding crypts that often demonstrate hyperplastic/serrated epithelial changes. However, the origin of this unique stromal proliferation is still unclear, and the association with serrated polyps, including sessile serrated adenomas, has not been fully described. We evaluated the pathologic and molecular features of colonic polyps associated with perineurial-like proliferations in 2 retrospective cohorts: (1) a series of 198 consecutive sessile serrated adenomas and (2) 20 colonic polyps diagnosed as a perineurioma irrespective of the presence of serrated colonic crypts. Thirteen of 198 (6.5%) sessile serrated adenomas demonstrated a perineurial-like stromal proliferation, with most (12 of 13, 92%) involving the right (9 cases) and transverse colon (3 cases). In all 13 cases, the perineurial-like proliferation surrounded serrated colonic crypts and typically involved only a small area of the sessile serrated adenoma (average 9% of polyp size; range, 2% to 19%). All 11 polyps evaluated for epithelial membrane antigen (EMA) expression demonstrated stromal EMA staining limited to the perineurial-like proliferation. Twelve of 13 (92%) sessile serrated adenomas with perineurial-like proliferations demonstrated a pV600E BRAF mutation. Of the 20 colonic polyps diagnosed as a perineurioma, 18 (90%) demonstrated serrated crypts intimately associated with the perineurial-like proliferation. In 13 of 18 polyps with associated serrated crypts, all serrated crypts were invested with the perineurial proliferation. In 5 cases, serrated crypts were seen away from the perineurial proliferation. Of these 18 polyps, the majority (16 of 18, 89%) were microvesicular hyperplastic polyps involving the left colon. However, 2 (11%) polyps in the right colon demonstrated histologic features diagnostic of sessile serrated adenoma. All 18 polyps with serrated crypts demonstrated a pV600E BRAF mutation. In contrast, the 2 polyps not associated with serrated crypts were negative for a BRAF mutation. Our results show for the first time that perineurial-like stromal proliferations frequently occur in sessile serrated adenomas. The presence of focal perineurial-like stromal proliferations in sessile serrated adenomas and the common finding of serrated crypts in colonic perineuriomas are likely indicative of an epithelial-stromal interaction, possibly related to some factor elaborated by the serrated epithelium.

Low-grade fibromyxoid sarcoma (LGFMS) is a rare soft tissue tumor that typically involves the deep soft tissues of the extremities. However, LGFMS has been described in other locations including the head and neck, retroperitoneum, and mesentery. Translocations involving the FUS gene located at 16p are considered to be highly specific for the diagnosis of this tumor when present in conjunction with the appropriate morphologic appearance. We report 4 cases of LGFMS arising in the small bowel. The patients' ages (3 female, 1 male) ranged from 52 to 71 years (median, 61 y). None of the patients had a history of soft tissue sarcoma. An intramural tumor was presented in all cases, ranging from 4.5 to 14.5 cm in greatest diameter (mean, 10 cm). Microscopically, the tumors consisted of an admixture of hypocellular, heavily collagenized zones and more cellular myxoid nodules, containing a greater number of neoplastic cells arranged around characteristic curvilinear vessels. By immunohistochemistry, all tumors were positive for vimentin and negative for smooth muscle actin, CD117, CD34, and S100 protein. Fluorescence in situ hybridization for rearrangement of the FUS locus at 16p11 was positive in 3 informative cases. Follow-up information was available in 3 cases; 2 patients are alive without disease and another patient died of disease. We conclude that LGFMS should be included in the differential diagnosis of fibrous or myxoid spindle cell tumors of the gastrointestinal tract. On the basis of limited follow-up, the natural history of enteric LGFMS seems to be similar to that of LGFMS in other locations, with potential for late metastases.

Objective: The purpose of this study was to evaluate the relationship between the depth of invasion within the mucosal compartment, with particular attention to the duplicated muscularis mucosae, and survival after esophagectomy performed as treatment for Barrett esophagus-related superficial adenocarcinoma., Methods: A total of 185 patients with pT1 esophageal adenocarcinoma treated by esophagectomy without induction therapy were identified. Depth of invasion was subdivided into invasion into the lamina propria (LP), into the inner muscularis mucosae, between the inner and outer muscularis mucosae, and into the outer muscularis mucosae (OMM), with comparison with tumors invading the inner one third of the submucosa (SM-1). Patient and tumor characteristics were compared among the 5 groups using the χ test or the Kruskal-Wallis test. Survival was estimated using the Kaplan-Meier method and compared using the log-rank test. The prognostic effect of depth of invasion on survival was assessed with Cox proportional hazards analysis., Results: Depth of invasion was LP (n=68), inner muscularis mucosae (n=38), BMM (n=11), OMM (n=33), and SM-1 (n=35). There was no significant difference in sex or age among groups. One of 150 patients with intramucosal adenocarcinoma (0.7%; LP, pN2) and 3 of 35 patients with SM-1 (8.6%; all pN1) had nodal disease. There were no significant differences in survival among the groups., Conclusions: Depth of invasion relative to the duplicated muscularis mucosae for tumors restricted to the mucosal compartment does not affect survival in Barrett esophagus-related superficial adenocarcinoma. Patients with SM-1 tumor had survival similar to those patients with tumor invasion into the OMM.

Signet ring cell change is a benign process that can be mistaken for signet ring cell adenocarcinoma. Signet ring cell change remains confined within the basement membrane and lacks an infiltrative growth pattern. It typically lacks cellular atypia, nuclear hyperchromasia, prominent nucleoli, or mitotic activity. This study illustrates a case of signet ring cell change with mitotic activity in a background of Helicobacter pylori gastritis with a penetrating ulcer. The signet ring cell change shows normal E-cadherin expression, absence of p53 expression, and confinement within glands on reticulin staining. The intact gastric mucosa adjacent to the signet ring cells shows significant mitotic activity, which likely explains the mitoses within the signet ring cell change. Mitotic activity in signet ring cells should be interpreted in the context of adjacent mucosa and should not preclude the diagnosis of benign signet ring cell change. Reticulin staining of the involved glands is helpful for excluding the invasion of the lamina propria in signet ring cell change.

Recent morphologic and molecular evidence suggests that dysplasia in Barrett esophagus (BE) begins in the bases of the crypts [crypt dysplasia (CD)] and progresses with time to involve the upper portions of the crypts and surface epithelium. The aim of this study was to evaluate the criteria and reproducibility of diagnosing CD among 6 gastrointestinal pathologists, all with research interest in BE. Six gastrointestinal pathologists evaluated 2 clinical study sets, the first consisting of 40 BE cases [BE: 10, CD: 9, low-grade dysplasia (LGD): 10, high-grade dysplasia (HGD); 9, and intramucosal adenocarcinoma; 2] and the second consisting of 63 cases (BE: 16, CD: 15, LGD: 15, HGD: 15, and intramucosal adenocarcinoma: 2), at least 4 months apart. In between evaluations, all of the pathologists met at 1 hospital (consensus conference) to review the areas of disagreement and establish more objective criteria. Overall, the level of agreement for all cases was moderate (κ=0.44), and the level of agreement did not change significantly after evaluation of the second study set. The highest levels of agreement were obtained for lesions at the low and high end of the spectrum (BE without dysplasia and HGD). Overall, the degree of agreement for CD was moderate after both the first and second study set review (κ=0.44 and 0.46, respectively). However, the degree of agreement for CD was higher than that obtained for LGD in both study sets. In the first study set, 4 or more pathologists agreed with the original CD diagnosis in 78% of cases, and this value did not change significantly after review of the second study set. The observers agreed that characteristic features of CD include the presence of unequivocal dysplastic cells, similar in appearance to traditional LGD, involving any part, or all of the length, of the crypt in the absence of intercrypt surface epithelial involvement. Rare cases of CD may show high-grade cytologic features composed of markedly enlarged nuclei with increased nuclear/cytoplasmic ratio, eosinophilic cytoplasm, irregular nuclear membranes, and loss of polarity. The findings in this study suggest that CD can be diagnosed reliably with a moderate level of interobserver agreement. Long-term and multi-institutional studies should be carried out to further determine the biological and clinical significance and natural history of CD in patients with BE.

Perivascular epithelioid cell neoplasms (PEComas) include the common renal angiomyolipoma, pulmonary clear cell sugar tumor, lymphangioleiomyomatosis, and less common neoplasms of soft tissue, gynecologic, and gastrointestinal tracts. Recently, aberrant immunoreactivity for TFE3 protein (a sensitive and specific marker of neoplasms harboring TFE3 gene fusions) has been reported in as many as 100% of PEComas; however, TFE3 gene status in these neoplasms has not been systematically investigated. We used a fluorescence in situ hybridization (FISH) break-apart assay to evaluate for evidence of TFE3 gene fusions in archival material from 29 PEComas. These cases included 2 earlier published TFE3 immunoreactive nonrenal PEComas, 14 additional nonrenal PEComas, and 13 renal angiomyolipomas with predominantly spindle or epithelioid morphology. Four nonrenal PEComas (mean patient age 24 y) showed TFE3 gene rearrangements by FISH, and all 4 of these showed strong positive (3+) TFE3 immunoreactivity using the original validated overnight incubation protocol. Two of these cases had adequate mRNA for RT-PCR analysis, but neither harbored the PSF-TFE3 gene fusion reported earlier in 1 PEComa. In addition, a lung metastasis of a uterine PEComa showed TFE3 gene amplification, an earlier unreported phenomenon. None of the other 24 PEComas (mean patient age 54 y) showed TFE3 gene alterations, though 4 exhibited moderate positive (2+) TFE3 immunoreactivity. In contrast, using an automated stainer, 2 of these 4 cases exhibited strong (3+) TFE3 immunoreactivity. All PEComas with TFE3 genetic alterations immunolabeled strongly for Cathepsin K, similar to other PEComas. In conclusion, a subset of lesions currently classified as PEComas harbors TFE3 gene fusions. Although numbers are small, distinctive features of these cases include a tendency to young age, the absence of association with tuberous sclerosis, predominant alveolar architecture and epithelioid cytology, minimal immunoreactivity for muscle markers, and strong (3+) TFE3 immunoreactivity. Despite significant morphologic and immunohistochemical overlap with other PEComas, PEComas harboring TFE3 gene fusions may represent a distinctive entity.

Diagnosing myxoid soft tissue neoplasms can be challenging because of overlapping histologic features. Distinct chromosomal translocations have been identified in several myxoid sarcomas, including t(12;16)(q13;p11) FUS-DDIT3 in myxoid liposarcoma, t(7;16)(q34;p11) FUS-CREB3L2 in low-grade fibromyxoid sarcoma, and t(9;22)(q31;q12) EWSR1-NR4A3 in extraskeletal myxoid chondrosarcoma. These recurrent chromosomal alterations are attractive targets for diagnostic studies. To that end, dual-color, break-apart fluorescence in situ hybridization (FISH) probes spanning the genomic regions of EWSR1 (22q12), DDIT3 (12q13), and FUS (16p11) (Vysis, Downer's Grove, IL) were evaluated in formalin-fixed, paraffin-embedded tissues from myxoid neoplasms, including intramuscular myxoma (n=10), myxoid liposarcoma (n=18), low-grade fibromyxoid sarcoma (n=10), extraskeletal myxoid chondrosarcoma (n=13), and myxofibrosarcoma (n=8). Of the myxoid liposarcomas, 18/18 cases had a rearrangement of the DDIT3 gene, with 17/18 (94.4%) showing both DDIT3 and FUS gene rearrangements. A FUS gene rearrangement was identified in 7/10 (70%) of low-grade fibromyxoid sarcomas, with no changes involving EWSR1 or DDIT3. An EWSR1 translocation was seen in 6/13 (46.2%) of extraskeletal myxoid chondrosarcomas, without changes in DDIT3 or FUS genes. The remaining neoplasms studied showed no rearrangements involving DDIT3, FUS, or EWSR1 genes. In conclusion, interphase FISH using DDIT3 and FUS probes identifies the characteristic translocation in myxoid liposarcoma. FUS and EWSR1 probes are useful in confirming the diagnosis of low-grade fibromyxoid sarcoma and extraskeletal myxoid chondrosarcoma, respectively. The specificity of the probes is documented as none of the non-translocation-associated myxoid tumors showed genomic abnormalities with the probes tested. FISH is capable of providing specific ancillary information useful in this often difficult differential diagnosis.

Synovial sarcoma is a soft tissue malignancy defined by the SYT-SSX fusion oncogene. Demonstration of the t(X;18) by cytogenetics, fluorescence in situ hybridization or reverse-transcriptase polymerase chain reaction has become the gold standard for diagnosis, but practical considerations limit the availability of these methods. Gene expression profiling studies performed by several independent groups have consistently identified TLE1 as an excellent discriminator of synovial sarcoma from other sarcomas, including histologically similar tumors such as malignant peripheral nerve sheath tumor. TLE proteins (human homologues of Groucho) are transcriptional corepressors that inhibit Wnt signaling and other cell fate determination signals, and so have an established role in repressing differentiation. We examined the expression of TLE proteins in synovial sarcoma and in a broad range of mesenchymal tumors using tissue microarrays to assess the value of anti-TLE antibodies in the immunohistochemical confirmation of synovial sarcoma. We demonstrate that TLE expression is a consistent feature of synovial sarcoma using both a well-characterized monoclonal antibody recognizing the TLE family of proteins and a commercially available polyclonal antibody raised against TLE1. Both antibodies gave intense and/or diffuse nuclear staining in 91/94 molecularly confirmed synovial sarcomas. Moderate staining is occasionally seen in schwannoma and solitary fibrous tumor/hemangiopericytoma. In contrast, TLE staining is detected much less frequently and at lower levels, if at all, in 40 other mesenchymal tumors. Our findings establish TLE as a robust immunohistochemical marker for synovial sarcoma, and may have implications for understanding the biology of synovial sarcoma and for developing experimental therapies for this cancer.

Campylobacter jejuni (CJ) is the most commonly isolated stool pathogen in the United States. Biopsy findings are typically those of focal active colitis (FAC), a nonspecific pattern usually indicating infection or adverse drug effect that is characterized by focal cryptitis and preservation of crypt architecture. We developed a molecular test for CJ that can be performed on routinely processed gastrointestinal biopsy specimens, and assessed what percentage of patients with biopsy findings of FAC have molecular evidence of CJ infection. One hundred and ten colon biopsies diagnosed as FAC were retrieved from three institutions. Polymerase chain reaction (PCR) was performed following DNA extraction; primers were designed to target a 286-bp fragment of the mapA gene that is specific to CJ. Pure genomic DNA derived from cultures served as the positive control; reagent blanks and 50 normal colon specimens served as negative controls. Nineteen percent (21/110) of the FAC biopsies were positive for CJ DNA by PCR analysis. Fourteen CJ-positive patients presented with diarrhea, 3 presented with gastrointestinal bleeding, and 3 had incidental FAC found on screening colonoscopy. Ten patients had abnormal colonoscopic findings, including erythema (4), ulcers (4), colitis (1), and hemorrhage (1). As CJ is an enteric pathogen that is not present in the gut as a commensal organism, the presence of CJ DNA suggests current or recent previous infection in these patients. CJ infection should be considered in patients with diarrhea and colon biopsies showing FAC. Furthermore, PCR analysis performed on fixed, routinely processed colon biopsies is an excellent diagnostic method for detection of this organism.

More than 85% of Ewing sarcoma/primitive neuroectodermal tumor (ES/PNET), or "Ewing family of tumors" (EFTs), have the translocation, t (11;22) (q24;q12), with others having variant translocations. Identification of these by cytogenetic and/or molecular genetic techniques is specific for EFT and is increasingly recognized as the "gold standard" for diagnosis. However, these techniques are not universally available. We therefore studied a large group of genetically confirmed EFTs to more completely understand the morphologic and immunophenotypic spectrum of this rare sarcoma. Sixty-six cytogenetically, FISH or RT-PCR proven-EFTs were retrieved. In 56 cases, immunohistochemistry (IHC) was performed for pan-cytokeratins (PanCK), high molecular weight cytokeratins (HMWCK), desmin (DES), CD99, CD117, and FLI1 protein using heat-induced epitope retrieval and the Dako Envision system. The cases arose chiefly in children and young adults (median 18 years; range, 3-65 years) of both sexes (male, 32; female, 31; unknown, 3) in a variety of bone (N = 39) and soft tissue (N = 27) sites. Histologically, 46 cases (73%) showed only typical features of ES, 9 cases (16%) showed features of PNET, 3 cases (5%) showed "adamantinoma-like" features, 3 cases (5%) corresponded to "atypical Ewing sarcoma," 3 cases (5%) showed principally intersecting fascicles of spindled cells, and 2 cases had abundant hyalinized matrix. IHC results were as follows: PanCK (18 of 56, 32%), HMWCK (3 of 55, 5%), DES (1 of 56, 2%), CD99 (52 of 52, 100%), CD117 (13 of 54, 24%), and FLI1 (44 of 47, 94%). HMWCK was expressed only in "adamantinoma-like" EFTs, none of which expressed DES. In conclusion, most, but not all, EFTs can be accurately diagnosed using time-honored morphologic criteria and ancillary immunohistochemistry. However, genetic confirmation remains essential for the diagnosis of unusual morphologic variants of EFT, including "adamantinoma-like," spindled, sclerosing, and clear cell/anaplastic variants. Therefore, to exclude or confirm the diagnosis of Ewing's sarcoma in round cell sarcomas having a variety of patterns but not specifically conforming to a tumor of known lineage (eg, rhabdomyosarcoma), cytogenetics, and/or molecular analysis is required.

Using gene microarray expression profiling, we previously found that apolipoprotein D (Apo D) was highly expressed in dermatofibrosarcoma protuberans (DFSP). In this study, we confirm that Apo D is highly and relatively specifically expressed in DFSP using immunohistochemistry. A tissue microarray containing 421 soft tissue tumors was constructed and stained with antibodies against Apo D and CD34. Cytoplasmic immunostaining for Apo D was found in 9 of 10 typical DFSPs. In addition, 3 of 3 Bednar tumors and 2 of 3 giant cell fibroblastomas stained in conventional sections. In contrast, Apo D was immunoreactive in only a very small subset of a diverse collection of other soft tissue tumors, including Malignant Fibrous Histiocytoma (MFH), glomus tumor, neurofibroma, and malignant peripheral nerve sheath tumors. Immunostains for Apo D were negative in conventional sections of 16 fibrous histiocytomas, and an additional 12 variants of fibrous histiocytoma. Digital images of all immunohistochemical and hematoxylin and eosin tissue microarray stains are available at the accompanying website (http://microarray-pubs.stanford.edu/tma&#95;portal/apod/). We conclude that Apo D is strongly expressed in DFSPs and neural lesions and may be useful in differentiating DFSP from fibrous histiocytoma.

Mesenchymal proliferations presenting as mucosal polyps are relatively uncommon and are represented by gastrointestinal stromal tumors, smooth muscle and neural tumors, and inflammatory fibroid polyps. In this report, we describe the clinicopathologic features of a distinctive type of mucosal polyp composed of cytologically bland spindled cells with fibroblastic features. Fourteen cases with histologic features of"fibroblastic polyps" were identified from our case files from January 2000 to December 2003. The clinical and endoscopic findings were reviewed. Immunohistochemistry using a panel of antibodies (vimentin, smooth muscle actin, desmin, CD31, CD34, Bcl-2, c-Kit, S-100, and epithelial membrane antigen) was performed in all cases, and electron microscopy was performed in two cases. The lesions were solitary in all cases and not associated with an identifiable polyposis syndrome. Associated adenomata and/or hyperplastic polyps at different sites were present in 10 cases and hyperplastic polyps were seen in close association in 3 cases. These polyps were characterized by a monomorphic spindle cell proliferation in the lamina propria, without necrosis or mitotic activity. The lesions were intimately associated with the muscularis mucosae and resulted in wide separation and disorganization of the colonic crypts. Immunohistochemical analysis revealed strong and diffuse positivity for vimentin only. Weak and focal reactivity was noted in 2 cases for CD34 and smooth muscle actin, while staining for other antibodies was negative. Electron microscopy revealed sparse cytoplasmic organelles and many intermediate filaments. The histology and ultrastructural and immunohistochemical findings are suggestive of fibroblastic differentiation of these spindle cells. In summary, these lesions represent a distinctive type of colonic mucosal polyp that should be distinguished from other stromal polyps of the gastrointestinal tract.

Angiosarcoma occurs very rarely in the intestinal tract as either a primary or metastatic malignancy and can present great diagnostic difficulty, especially when it displays epithelioid cytomorphology. Since only isolated case reports have been published, the purpose of this study is to more fully delineate the histopathological and clinical features from a series of 8 angiosarcomas involving the gastrointestinal tract. There were 5 male and 3 female patients whose ages ranged from 25-85 years (median 57). Presenting symptoms included intestinal bleeding, anemia and pain. Five cases involved the small bowel and 3 involved the colon/rectum. Four cases were primary to the intestinal tract, 2 patients initially presented with secondary involvement of the large bowel from occult retroperitoneal primaries, 1 patient presented with disseminated disease including small bowel involvement, and 1 case was metastatic from a breast primary. Seven cases were composed predominantly of sheets of malignant appearing epithelioid cells with subtle areas forming cleft-like spaces suggestive of vascular differentiation. Immunohistochemical studies revealed the lesional cells to be immunoreactive for CD31 (8/8), CD34 (8/8), Factor VIII (8/8), cytokeratins AE1/AE3 (7/8), cytokeratin 7 (2/8), Cam5.2/cytokeratin 8 (5/8), and cytokeratin 19 (5/8). Cytokeratin 20 was negative in all eight cases, which contrasts sharply with the characteristic positivity for cytokeratin 20 in virtually all intestinal carcinomas. One case was weakly and focally positive for EMA and all cases were negative for S-100 protein. Cytokeratin staining was variable and ranged from focal to extensive. Follow-up was available in eight cases and ranged from 1-33 months (median 12.5). Five patients died of disease, between 1 and 33 months (median 6) after diagnosis. One recently diagnosed patient is alive with disease 18 months after diagnosis, and one patient is free of disease 27 months after original diagnosis. Angiosarcomas of the gastrointestinal tract commonly display epithelioid cytomorphology, may be diffusely and strongly positive for cytokeratins and only show subtle signs of vascular differentiation, creating potential diagnostic confusion with primary or metastatic carcinoma. Given the clinically aggressive behavior of angiosarcoma, proper classification and treatment is important. Immunohistochemistry with vascular markers, CK20, and S-100 protein may be helpful in differentiating angiosarcoma from carcinoma and melanoma.

Primary neuroendocrine carcinomas of the colon are rare but highly aggressive malignancies. The recent observations that c-kit protooncogene, a tyrosine kinase, is overexpressed in a subset of small cell lung cancer and that selective kinase inhibitors block the in vitro growth of small cell lung cancer cell lines prompted us to investigate the expression and mutation status of the c-kit gene in colorectal neuroendocrine carcinomas. Sixty-six cases of primary colorectal neuroendocrine carcinoma were collected from 13 institutions, including 36 small cell carcinomas and 30 moderately differentiated neuroendocrine carcinomas. Immunohistochemical studies using a polyclonal antibody against c-kit protein (CD117) demonstrated a strong and diffuse cytoplasmic staining in 15 cases (23%), which were relatively equally distributed in the small cell and moderately differentiated subgroups. As controls, 25 conventional colorectal adenocarcinomas, 26 colorectal adenomas and 19 colorectal carcinoid tumors were all negative, whereas 15 gastrointestinal stromal tumors were all positive, for kit expression. In contrast to gastrointestinal stromal tumors, kit-overexpressing neuroendocrine carcinomas showed no mutations in the juxtamembrane domain (exon 11) of the c-kit gene as determined by mutational analysis. Kaplan-Meier analysis with the log-rank test revealed that the patients with kit-positive tumors did not differ significantly in survival from those with kit-negative tumors (P = 0.77). These results indicate that c-kit overexpression observed in a subset of colorectal neuroendocrine carcinomas may not be mediated via activating mutations, and does not appear to be an initiating event during tumorigenesis because of lack of c-kit expression in other types of colorectal epithelial neoplasms. More importantly, our observations may have potential therapeutic implications since specific tyrosine kinase inhibitors have shown promise in the management of patients with kit-expressing malignancies.

Placental alkaline phosphatase (PLAP) is normally produced by primordial germ cells and syncytiotrophoblasts, and the detection of its expression has been useful in the diagnosis of germ cell tumors. We have recently observed PLAP immunoreactivity in normal human adult and fetal muscle tissue. Based on this observation, we explored the possible role of PLAP in the diagnosis of soft tissue tumors. A total of 271 tumors were studied. These included tumors with myogenic, neural, fibrous, myofibroblastic, lipomatous, neuroepithelial, perivascular, and epithelial differentiation. A formalin-fixed, paraffin-embedded section from each tumor was stained with PLAP monoclonal antibody using standard immunohistochemical methods preceded by antigen retrieval. In addition, western blotting with PLAP monoclonal antibodies was performed on fresh samples from a uterine leiomyoma, grossly normal myometrium, and placenta. Also, formalin-fixed sections of fetal skeletal muscle were labeled with double immunohistochemistry techniques using antibodies to myogenin and PLAP. Cytoplasmic PLAP reactivity was detected in all leiomyomas and rhabdomyosarcomas (100%), 7 of 15 (46%) leiomyosarcomas, 15 of 19 (79%) desmoplastic small round cell tumors, 2 of 15 (13%) gastrointestinal stromal tumors, 1 of 8 (13%) Wilms' tumors, 1 of 9 synovial sarcomas (9%), and 2 of 7 (29%) myofibroblastic tumors. No PLAP reactivity was detected in hyperplastic scars, nodular fasciitis, or the other remaining soft tissue and epithelial tumors. Double immunohistochemistry studies showed coexpression of myogenin and PLAP in fetal skeletal muscle cells, and western blot analysis showed a 70-kDa band in samples derived from grossly normal placenta, benign myometrium, and a uterine leiomyoma. PLAP immunoreactivity is detected in soft tissue tumors with known myogenic differentiation. PLAP immunoreactivity seems to relate to the degree of myogenic differentiation in soft tissue tumors and is more frequently expressed in cells with skeletal muscle differentiation and least in those with myofibroblastic features. The biologic function of PLAP in muscle and tumors with myogenic differentiation is unknown and merits further investigation. In addition to its role as a germ cell marker, PLAP may also be used as a myogenic marker in the diagnosis of soft tissue tumors.

Gastrointestinal stromal tumors are a heterogeneous group of neoplasms that have clinical and histologic features that vary depending on their location within the gastrointestinal tract. Prediction of clinical behavior in this group of tumors is notoriously difficult, and the same criteria for malignancy do not necessarily apply to stromal tumors from different sites within the gastrointestinal tract. Using known clinical behavior with long-term follow-up, we attempted to determine which features, if any, are associated with clinical behavior in stromal tumors arising in the stomach, the most common site for such tumors. Seventy-seven gastric stromal tumors were studied and classified as "adverse outcome (AO) tumors" (malignant) or "nonadverse outcome tumors" (benign) based on their known clinical outcome. AO was defined as metastasis and/or death due to tumor. Patients with a non-AO had at least 5 years of tumor/metastasis-free follow-up. Thirty-seven patients had an AO (follow-up [metastasis at presentation] 0-73 months; median 6 months), and 40 patients had a non-AO (follow-up 60-264 months; median 84 months). All cases were reviewed by two authors (J.R.G., H.D.A.), who were blinded to clinical outcome and gross features, and classified as histologically benign or not benign using preset, defined histologic criteria based upon the authors' prior experience with a large number of these tumors. If the tumor did not fit with either the characteristic cellular spindle cell or benign epithelioid cell patterns, the tumor was classified as not benign. Clinical outcome was then correlated with the histologic designation to determine if these preset criteria were valid. The authors were able to accurately classify the tumors as benign or not benign with a sensitivity of 100% and a specificity of 92%. In addition, for all cases individual morphologic and clinical features were examined. Features associated with an AO included tumor size >/=7 cm, high cellularity, mucosal invasion, high nuclear grade, mitotic counts >/=5/50 high power fields, mixed cell type, and the presence of a myxoid background and/or absence of stromal hyalinization. By recognizing several well-defined patterns of benign gastric stromal tumors and the myriad of individual features shown to correlate with an AO, one can better predict the clinical behavior of gastric stromal tumors.

Pseudoachalasia is an esophageal motor disorder usually associated with malignancy that has clinical, radiographic, and manometric findings that are often indistinguishable from primary achalasia. There are few reports examining the histologic features of the associated neoplasms and their relationship with the esophageal myenteric plexus. We studied the clinical and pathologic features of 13 cases of pseudoachalasia seen at our institution between 1979 and 1999. Detailed clinical and radiographic data were obtained from medical records. In all cases available histologic material was reviewed to confirm the presence and type of associated neoplasm. When possible, the relationship of the neoplasm to the esophageal myenteric plexus was examined. In selected cases immunohistochemical stains were performed to further evaluate this relationship. All patients had clinical, radiographic, and manometric features similar to primary achalasia. The cohort included seven men and six women, age range 24-79 years (median 61 years). Associated neoplasms included esophageal adenocarcinoma arising in Barrett's esophagus (n = 1), adenocarcinoma of the esophagogastric junction (n = 7), metastatic renal cell carcinoma to the esophagogastric junction (n = 1), breast adenocarcinoma (n = 1), pulmonary small cell carcinoma (n = 1), pleural malignant mesothelioma (n = 1), and mediastinal fibrosis (n = 1). The mechanism of pseudoachalasia was consistent with neoplastic infiltration of the esophageal myenteric plexus in 11 cases. Neoplastic cells surrounded myenteric ganglion cells, which appeared normal in number and morphology. In the patient with pulmonary small cell carcinoma, there was no evidence of neoplastic infiltration of the esophagogastric junction, and anti-ANNA-1 antibody was detected, suggesting a paraneoplastic syndrome. Tissue obtained at the time of esophagomyotomy revealed lymphocytic myenteric inflammation and marked depletion of ganglion cells identical to that seen in primary achalasia. The mechanism pseudoachalasia in the patient with breast adenocarcinoma is uncertain, as there was no evidence of direct involvement of the esophagogastric junction. In summary, we describe 13 cases of pseudoachalasia resulting in a clinical syndrome indistinguishable from primary achalasia. The most common mechanism is direct involvement of the esophageal myenteric plexus by neoplastic cells. Rarely, a distant neoplasm may cause this syndrome as a paraneoplastic process.