Distinguishing primary mucinous ovarian tumors from metastases of non-gynecologic mucinous cancers: Can we leverage next-generation sequencing?

person Muhammad Danyal Ahsan Weill Cornell Medicine - Qatar, Doha, Qatar info_outline Muhammad Danyal Ahsan, Murtaza Qazi, Ryan M Kahn, Emily M Webster, Sarah R. Levi, Benedict Harvey, Luiza Perez, Jesse T Brewer, Laura Keenahan, Gaurav Thareja, Ravi N. Sharaf, Evelyn Cantillo, Eloise Chapman-Davis, Melissa Kristen Frey, Dennis S Chi, Kevin Holcomb

Authors person Muhammad Danyal Ahsan Weill Cornell Medicine - Qatar, Doha, Qatar info_outline Muhammad Danyal Ahsan, Murtaza Qazi, Ryan M Kahn, Emily M Webster, Sarah R. Levi, Benedict Harvey, Luiza Perez, Jesse T Brewer, Laura Keenahan, Gaurav Thareja, Ravi N. Sharaf, Evelyn Cantillo, Eloise Chapman-Davis, Melissa Kristen Frey, Dennis S Chi, Kevin Holcomb Organizations Weill Cornell Medicine - Qatar, Doha, Qatar, Weill Cornell Medicine, New York, NY, Memorial Sloan Kettering Cancer Center, New York, NY Abstract Disclosures Research Funding No funding received None. Background: Primary mucinous ovarian cancers (MOC) are histopathologically challenging to differentiate from ovarian metastases of non-gynecologic origin, with this distinction being critical for appropriate management and prognosis. We compared the somatic gene variant landscape of MOC to that of non-gynecologic mucinous tumors. Methods: Data were extracted from the American Association for Cancer Research’s (AACR) Project Genomics Evidence Neoplasia Information Exchange (GENIE) database version 13.0 via cBioPortal. This publicly available, multi-institutional database provides next generation sequencing (NGS) genomic profiles of tumors. We queried this database for samples of MOC, mucinous colorectal cancer (MCRC), mucinous appendiceal cancer (MAC), mucinous breast cancer (MBC) and gastric type mucinous cancer (GMC). Frequencies of somatic gene variants including mutations, copy number alterations and structural variants were compared using Chi-squared or Fischer’s exact tests, using the Benjamini-Hochberg method to control for multiple hypothesis testing with q-values reported. Results: A total of 883 tumors were included for analysis: 358 MCRC, 268 MAC, 157 MOC, 59 MBC and 41 GMC samples. Compared to MAC, MOC samples had higher variant frequencies of CDKN2A (33.3% vs 0.4%, q<0.001), CDKN2B (24.0% vs 0.0%, q<0.001), TP53 (64.3% vs 23.5%, q<0.001), ERBB2 (14.3% vs 1.1%, q<0.001) and CDK12 (13.2% vs 0.0%, q<0.001), whereas GNAS variants were more common in MAC (45.5% vs 5.7%, q<0.001). Compared to MCRC, MOC samples had higher variant frequencies of CDKN2A (33.3% vs 2.3%, q<0.001), CDKN2B (24.0% vs 3.7%, q<0.001), TP53 (64.3% vs 42.9%, q<0.001), KRAS (69.4% vs 51.1%, q<0.001) and ERBB2 (14.3% vs 5.3%, q<0.001), whereas MCRC had higher variant frequencies of 57 genes, with the largest differentials among APC (48.8% vs 2.6%, q<0.001), SMAD4 (25.2% vs 5.8%, q<0.001) and TCF7L2 (19.0% vs 0.0%, q<0.001). Samples of MOC had significantly higher rates of KRAS variants compared to GMC (69.4% vs 31.7%, q<0.001) and lower rates of STK11 variants (1.9% vs 22.0%, q<0.001). Compared to MBC, MOC samples had higher variant rates of CDKN2A (33.3% vs 3.4%, q<0.001), TP53 (64.3% vs 10.2%, q<0.001) and KRAS (69.4% vs 0.0%, q<0.001), whereas MBC samples had higher variant frequencies of 11 genes, with the largest differentials among GATA3 (32.1% vs 0.8%, q<0.001), FGF3 (30.4% vs 2.4%, q<0.05) and CCND1 (28.1% vs 1.6%, q<0.001). Conclusions: NGS demonstrates that MOCs carry a distinct genetic signature compared to mucinous tumors of non-gynecologic origin, most commonly with significantly higher variant frequencies of CDKN2A , CDKN2B and lower variant frequencies of GNAS , APC , STK11 and GATA3 . This provides rationale for prospective studies evaluating genetic signatures as an adjunct to histopathology in the diagnosis of primary MOC.