CREB fusion–associated epithelioid mesenchymal neoplasms of the female adnexa: three cases documenting a novel location of an emerging entity and further highlighting an ambiguous misleading immunophenotype

The EWSR1/FUS-CREB fusion family of neoplasms have been ever expanding to encompass a variety of histologically, immunophenotypically, biologically, and topographically distinct entities. This was the consequence of the ever-increasing wide use of next generation sequencing tools in routine surgical pathology. With accumulating knowledge, it became evident that the mere presence of a specific gene fusion does not define an entity without consideration of other anatomic and immunophenotypic context. This is more obvious for clear cell sarcoma (definitionally with a melanocytic cell phenotype) and malignant gastrointestinal neuroectodermal tumor (coexpression of S100 and SOX10 but not specific melanocytic markers). Other entities in this genetic tumor spectrum such as AFH possess defining morphological (e.g., syncytial epithelioid or spindle cells in fascicular growth associated with variable aneurysmal/angiomatoid changes and lymphoid cuffs) and immunophenotypic (variable expression of cytoplasmic CD99, desmin, EMA, and ALK) features. Another subset is defined principally by prominent myxoid pattern mimicking extraskeletal myxoid chondrosarcoma and occurring at specific sites (lung, intracranial) but lacking specific immunophenotype. The last and most recently proposed member in this tumor category are intra-abdominal epithelioid neoplasms (sarcomas) that except for their frequent keratin expression, intra-abdominal location and neuroendocrine-like morphological and/or immunophenotypic features are not readily diagnosable by morphology and immunophenotyping alone. The latter group has been increasingly recognized in different abdominal sites.

CREB-associated sarcomas have not been defined in the gynecological tract, but they might have been under-recognized or diagnosed as other entities given their highly ambiguous and misleading immunophenotype as illustrated by our current cases. To our knowledge, only a single case of an AFH presenting as an ovarian mass has been reported to date. The tumor affected a 22-year-old woman, measured 6 cm in size and was resected 2 years after being incidentally discovered on ultrasound. The patient remained well, 8 months later. Molecular testing revealed an EWSR1::CREB1 fusion [9]. Our study represents the first one to delineate these tumors at a novel anatomic site and at same time reporting features that are not AFH-typical in line with the emerging notion that these tumors might represent a distinctive family of neoplasms having the EWSR1/FUS::CREB fusions in common, but with variable morphological and immunophenotypic characteristics.

Two of our cases were associated with constitutional inflammatory symptoms such as fever, anemia, weight loss, and elevated serum CRP, all regressed immediately after tumor removal. Similar paraneoplastic phenomena are well known in AFH of soft tissue and are mostly mediated by IL-6 excess produced by the neoplastic cells. All of our cases and the one reported by Chen et al occurred in adult patients, 22 to 42 years old. The first ovarian case reported by Chen et al. [9] presented the morphology of AFH, with a characteristic pericapsular rim/ cuff of lymphoid tissue. Two of our cases were also surrounded by a fibrous capsule containing inflammatory cells but without aggregation of lymphocytes. However, our case 3 showed a malignant behavior with tumor extension into the myometrium and lymph node metastasis. Our cases showed variable morphological overlap with Chen et al.’s, being composed of islands of epithelioid cells separated by a sclerotic stroma infiltrated by chronic inflammatory cells. Bizarre and pleomorphic nuclei were seen without numerous mitoses. Based on the morphology and the immunoprofile with desmin and EMA positivity and the finding of an EWSR1::CREB1 fusion, Chen et al. classified their case as an extrasomatic AFH [9].

AFH is a rare soft tissue neoplasm (0.3% of all soft tissue tumors) of low malignant potential (with 15% local recurrence and less than 2–5% metastases), typically occurring in the superficial soft tissues (deep dermis, or subcutis) of the extremities in children and young adults. It presents as slowly growing, superficial painless mass often simulating a hematoma [12]. Occasionally, systemic symptoms such as anemia, pyrexia, malaise, and weight loss, suggesting tumoral cytokine production, have been reported [12]. AFH lacks a specific immunophenotype, with desmin, EMA, ALK, CD68, and CD99 positivity being demonstrable in most cases [13]. In the appropriate morphological and clinical context, these markers proved valuable in recognizing AFH. In difficult cases, demonstration of the characteristic gene fusion involving EWSR1 or FUS and one of the CREB family genes (ATF1 or CREB1) can be very helpful [13, 14].

Rare cases of AFH outside somatic soft tissue have been described involving the lung, mediastinum, bone, brain, vulva, retroperitoneum, and ovary [9]. As compared to their soft tissue counterparts, extrasomatic AFH occur in older patients (median age 35 versus 20-year-old), with a higher frequency of systemic symptoms (30% versus 10–15%), larger tumors, higher rate of recurrence (33% versus 11%), and a higher frequency of ATF1 as fusion partner (63% versus 19%) [9]. The morphology of AFH is identical in soft tissue and extrasomatic site, with a characteristic peritumoral cuff of lymphoplasmacytic infiltrate, occasionally simulating a lymph node, a multinodular growth pattern, dendritic-like, and epithelioid tumor cells with eosinophilic cytoplasm and abundant admixed plasma cells [12]. Some cases present with a syncytial or histiocytoid aspect often admixed with abundant hemosiderin. We could argue that our two first cases, well-circumscribed, and showing EWSR1::ATF1 fusion, could be closer to an AFH than to the CREM fusion family of malignant intra-abdominal epithelioid neoplasms. However, our case 3 demonstrated malignant features such as myometrial extension and lymph node metastasis, without any desmin positivity and had CREM as partner for EWSR1. Indeed, the majority of intra-abdominal malignant epithelioid neoplasms do not express desmin and involve CREM as partner for ESWR1/FUS, while AFHs show diffuse desmin positivity with ESWR1/FUS::CREB1/ATF1 fusions, like our two first cases.

The location of our cases 1 and 3 at the serosal surface of the adnexa and the morphological findings reveal certain overlap with mesothelioma in young adults [5] that also occasionally harbors ESWR1/FUS::CREB-related fusions. Although epithelioid peritoneal mesotheliomas of young adults without a history of asbestos exposure may retain BAP1 expression and may lack calretinin positivity, they show at least focally some degree of frankly papillary or glandular patterns [5], features absent in our cases. In addition, our cases 1 and 3 lacked cytokeratin positivity which is required for the diagnosis of mesothelioma [5]. Overall, the morphology of the three tumors is not compatible with epithelioid mesothelioma.

In the female genital tract, an epithelioid neoplasm with numerous lymphocytes and plasma cells, such as that seen in our cases, suggests several differential diagnoses. A gestational trophoblastic tumor can be ruled out by the absence of diffuse keratin expression, while the absence of any immunoreactivity with H-caldesmon and hormone receptors is not in favor of an epithelioid smooth muscle tumor. Since 2/3 of AFH and epithelioid EWSR1::CREB-rearranged malignant neoplasms may express ALK protein, an epithelioid inflammatory myofibroblastic tumor (IMT) is a major differential diagnosis. However, the diffuse cytoplasmic ALK pattern is different from the distinct nuclear membrane ALK seen in majority of epithelioid IMTs. Moreover, lack of ALK and ROS1 gene fusion by NGS excludes this possibility. Given the focal positivity of WT1, an adnexal sex cord stromal tumor is also in the differential diagnosis, especially since inhibin immunoreactivity has been described in EWSR1/FUS::CREB-rearranged tumors [6]. The expression of EMA in these tumors and our cases, along with the absence of hormone receptor positivity and of sex cord markers all argue against this diagnosis.

Parallel to our study, a series of male gonadal stromal tumors harboring EWSR1::ATF1 gene fusions has been published online using the terminology “inflammatory and nested testicular sex cord tumor with aggressive clinical behavior and EWSR1::ATF1 gene fusions” [15]. The morphology of these cases is quite similar to ours, including the presence of inflammatory cells and hyaline globules as well as the epithelioid aspect of the tumor cells, although nesting is more prominent in the testicular variant. Because of inhibin and SF-1 positivity and the nesting, these tumors have been designated as sex cord tumors. However, the positivity of EMA reported in all 9 testicular cases is very unusual for a sex cord tumor. Beside, inhibin positivity has been reported in EWSR1/FUS::CREB-rearranged tumors [6]. These testicular tumors might correspond to the gonadal male counterpart of the cases we report herein. However, all of our cases lacked sex cord marker positivity.

The most appropriate terminology for the cases we are reporting herein remains to be defined. We conclude that our cases do not represent variants of sex cord stromal tumors or any other defined mesenchymal/ stromal category of the female genital tract neoplasms. Instead, these tumors overlap significantly with the recently reported CREM fusion–associated intra-abdominal neoplasms, both morphologically and immunophenotypically, including occasional ambiguous (polyphenotypic) immunoprofile with expression of epithelial markers, MUC4, ALK, and neuroendocrine markers in subsets. In addition, frequent desmin reactivity and certain morphological features overlap with AFH of soft tissue. We propose the descriptive term “CREB fusion-associated epithelioid mesenchymal neoplasms of the female adnexa” for these tumors to enable better recognition by gynecological pathologists, and hence further characterization in the future. While the low number of reported cases does not allow for conclusive prognostic statement, it seems that these tumors of the adnexa likely are indolent but may be associated by locoregional aggressive behavior. As of now, none of the 4 reported cases developed distant metastases.

In conclusion, we described three cases of epithelioid neoplasms involving the female adnexa with ESWR1::ATF1/CREM fusions. These tumors belong to the spectrum of mesenchymal lesions harboring ESWR1/FUS::CREB family fusions involving the peritoneal cavity and intra-abdominal organs of young adults and including AFH-like and unclassified malignant epithelioid neoplasms. While AFH-like intra-abdominal neoplasms carry a low malignant potential, the CREB-positive intra-abdominal unclassified epithelioid neoplasms have a very aggressive behavior. As these tumors might feature a misleading bland or cystic appearance on imaging which combined with the young age may argue for a benign lesion, a high suspicion index and in particular presence of constitutional symptoms should alert to this differential. Same applies to their ambiguous and potentially misleading immunophenotype with a wide differential including sex cord tumors, mesothelial neoplasms, neuroendocrine tumors, IMT, and others. Accordingly, gynecopathologists should be aware of these entities’ in the differential diagnosis of an adnexal mass with epithelioid cell morphology.

留言 (0)

沒有登入
gif