Comprehensive genomic analysis of mucinous ovarian cancer reveals unique therapeutic vulnerabilities.

Abstract

5571 Background: Mucinous ovarian carcinoma (MOC) is a rare subtype of epithelial ovarian cancer that responds poorly to ovarian chemotherapies and has an unknown etiology. It is diagnostically challenging and can be confused with metastases from gastro-intestinal tract primaries. The GAMuT study is a multi-national effort to understand molecular drivers and cell of origin of this rare tumour, including identification of a genetic progression model and novel therapeutic options. Methods: We performed RNAseq (n = 67), exome sequencing (n = 61), SNP arrays (n = 67) and whole genome sequencing (n = 5) on MOC and precursor lesions. A subset of ~500 genes was further evaluated by targeted sequencing, including 129 MOC, 23 borderline mucinous tumours (non-invasive) and 23 extra-ovarian mucinous metastases. Immunohistochemistry data was collected for CK7, CK20, ER, PAX8, p53 and HER2 (n = 162-256). Extensive pathology review was performed and associated clinical data obtained. Results: Comparison with TCGA and other data sets showed that MOC are distinct from mucinous tumours from other organs, including colorectal, appendiceal and gastric cancers. Our data supports a clear genetic progression model from benign and borderline precursors to both low- and high-grade MOC. TP53 mutation, ERBB2 amplification and increasing copy number changes were key events associated with progression to invasive disease, including a novel amplicon on 9p13. Copy number aberration burden was significantly associated with poor survival. We identified several recurrent mutational events suggesting utility of an existing targeted therapy, including ERBB2 amplification (26%), ERBB3 mutation (4%) and BRAF mutation (9%). MOC could be included in clinical trials for novel agents targeting TP53 missense mutation (46%), RNF43 mutation (12%), PIK3CA mutation (8%) and KRAS/ NRAS mutations (66%). Other frequent events included CDKN2A inactivation (57%), ARID1A mutation (9%) and TP53 inactivating mutations (15%). Conclusions: MOC of any grade can derive from a primary ovarian tumour precursor, and is distinct from extra-ovarian metastases. MOC is genetically diverse and advanced disease should be assessed for targetable mutations which may provide novel therapeutic options.