Uterine tumor resembling ovarian sex cord tumor (UTROSCT) is a rare mesenchymal neoplasm, of uncertain biological potential, that was recently reported to exhibit recurrent gene fusions involving NCOA2-3. The purpose of this study was to, using a larger sample size, better characterize the histopathologic and molecular diversity of UTROSCT. Twenty-six cases of UTROSCT from 5 institutions were selected for further study. Fluorescence in situ hybridization for NCOA1, NCOA2, NCOA3, ESR1 and GREB1, and targeted RNA sequencing was performed on 17 and 8 UTROSCTs, respectively. Eight cases underwent massively parallel sequencing to detect single nucleotide variants (SNV), copy number variations, and structural variants using a targeted hybrid-capture based assay. NCOA1-3 rearrangement was identified in 81.8% (18/22) of cases. The most common fusion was ESR1-NCOA3, occurring in 40.9% (9/22). GREB1-NCOA1 (n=4), ESR1-NCOA2 (n=3), and GREB1-NCOA2 (n=1) rearrangements were also identified. No recurrent SNVs were identified and no tumor had SNVs in FOXL2, DICER1, STK11, or AKT1, which can be seen in ovarian sex cord-stromal tumors. Copy number variations were infrequent. Clinical follow-up was available for 11 cases with a mean follow-up interval of 94.4 (range, 1 to 319) months. Only one case had a recurrence 66 months after the initial diagnosis and this was the single case with a GREB1-NCOA2 fusion. This study reports the morphologic spectrum of UTROSCT and confirms the recently reported recurrent NCOA2-3 gene fusions, in addition to identifying novel rearrangements involving NCOA1 in these tumors.

Uterine cancer was first subclassified based on anatomic site, separating those tumours arising from the endometrium from cervical cancers. There was then further subclassification of endometrial cancers based on cell type, and this correlated with the Type I and Type II categories identified through the epidemiological studies of Bokhman, with endometrioid carcinoma corresponding (approximately) to Type I and serous carcinoma to Type II. These histotypes are not clearly separable in practice, however, with considerable interobserver variability in histotype diagnosis, especially for high-grade tumours. There followed studies of immunomarkers and then mutational studies of single genes, in attempts to improve subclassification. While these have revealed significant differences in protein expression and mutation profiles between endometrioid and serous carcinomas, there is also considerable overlap, so that there remain challenges in subclassification of endometrial carcinoma. Gene panel testing, using next-generation sequencing, was applied to endometrial cancers and highlighted that there are tumours that show genetic alterations intermediate between classic Type I/endometrioid and Type II/serous carcinomas. The Cancer Genome Atlas studies of endometrioid and serous carcinoma offered revolutionary insight into the subclassification of endometrial carcinoma, i.e. that there are four distinct categories of endometrial carcinoma, rather than two, based on genomic architecture. In this review, we provide an overview of immunohistochemical and molecular markers in endometrial carcinoma and comment on the important future directions in endometrial carcinoma subclassification arising from The Cancer Genome Atlas results.

POLE mutations occur in 7-12% of endometrial cancers and are identified by molecular analysis.• POLE-mutated endometrial cancers have high tumor mutation burden, tumor neoantigen production, and tumor infiltrating T cells.• This case illustrates a marked response to immune checkpoint inhibition in a POLE-mutated endometrial cancer.