Comprehensive molecular assessment of mismatch repair deficiency in Lynch-associated ovarian cancers using next-generation sequencing (NGS) panel.

Abstract

1523 Background: Abnormalities in mismatch repair (MMR) gene may be the result of pathogenic germline (Lynch syndrome) and somatic mutations as well as epigenetic events. Abnormalities in MMR have been described in non-serous/non-mucinous ovarian cancer (OC) but few studies have examined the causes of these MMR defects (MMRd). To address this, we have completed targeted mutational and methylation sequencing on MMRd OC cases. Methods: Women with newly diagnosed non-serous/mucinous OC (N = 215) were prospectively recruited from three cancer centers in Ontario, Canada between 2015-18. Tumors were reflexively assessed for MMR protein expression by immunohistochemistry. Tumor DNA was extracted from macrodissected MMRd cases and MMR-intact (MMRi) controls following pathology review. Matched tumor-normal samples were run on a custom NGS panel to identify germline and somatic mutations, copy number variants, rearrangements and promoter methylation in MMR and associated genes. Results: Of the 215 women enrolled in our study, 185 (86%) had OC alone and 30 (14%) had synchronous OC and endometrial cancer. Twenty-eight (13%) cases were MMRd, 11 of which were synchronous. The MMRd cohort had median age of 52.5 years, with mostly stage I (N = 14; 50%), grade 1 or 2 disease (N = 18; 64%) with endometrioid histotype (N = 18; 64%). One patient had recurrence after median follow-up of 33.6 months (13.2-93.6). There was no significant difference in overall/progression-free survival between the MMRd and MMRi patients. Using the NGS panel, Lynch syndrome (LS) was detected in 39% of MMRd cases (11/28; 7 OC and 4 synchronous): 7 MSH6, 2 MLH1, 1 PMS2, and 1 MSH2. Clinical germline sequencing was performed on all cases and verified panel findings. An explanation for the observed MMR phenotype was available for 18/20 deficient cases, including 9/10 MLH1−/PMS2− (7 somatic methylation, 1 bi-allelic somatic deletion, 1 germline mutation), 0/1 PMS2−, 6/7 MSH6− (6 germline mutations) and 2/2 MSH2−/MSH6− (1 germline mutation, 1 bi-allelic somatic mutation). Concordance between clinical and research panel sequencing results was 90%. None of the germline mutations were missed by the panel. Conclusions: Use of our custom NGS panel allows for the streamlined assessment of hereditary and somatic causes of MMR deficiency in OC and may be an attractive screening strategy for LS in this population.