e15205 Background: Mucinous carcinoma (MC) is a rare, aggressive histological subtype of adenocarcinoma that is associated with poor prognosis and peritoneal spread. Molecular features of MC across gastrointestinal (GI) primary sites remain incompletely defined. Utilizing the American Association for Cancer Research (AACR) Project Genomics Evidence Neoplasia Information Exchange (GENIE) v18.0 and OncoKB annotator, we performed a comprehensive analysis of GI MCs. Methods: The AACR GENIE v18.0 database was used to select GI MC and GI non-MC samples by OncoTree code. Samples were analyzed for clinicogenomic variables including age at sequencing, sex, and oncogenic molecular alterations by OncoKB classification (somatic mutations, structural variants, copy number alterations), which were further annotated by therapeutic level of evidence (Levels 1-4). Samples were considered to have a potentially actionable alteration if there was ≥1 Level 1-3B alteration. Two-way models regression evaluated main effects of primary site and mucinous status and their interaction. Benjamini-Hochberg correction controlled false discovery rate, with q < 0.05 considered significant. Results: 786 patients with GI MC were analyzed across primary sites (colorectal (CRC) n = 423; appendix n = 346; stomach n = 17) with corresponding non-MC comparators (CRC n = 20,209; appendix n = 576; stomach n = 2082). Age at sequencing and sex varied by primary site (q-site < 0.001), but no difference was observed by mucinous status. There was higher prevalence of potentially actionable alterations in MC compared to their non-MC variants across primary tumor sites (q-mucinous < 0.001). MC exhibited distinctive molecular patterns with enrichment in KRAS and GNAS (q-mucinous < 0.001), and there were significant site-by-mucinous interactions for KRAS , TP53 , SMAD4 , PIK3CA (q-interaction < 0.05), supporting site-specific drivers in MC biology. Conclusions: To our knowledge, this study represents the largest molecular analysis of MC across the GI tract, revealing distinctive molecular alteration patterns. These findings underscore site-specific genomic patterns and motivate future studies to prioritize therapeutic strategies. Clinical Variable/ Alteration MC CRC Non-MC CRC MC appendix Non-MC appendix MC stomach Non-MC stomach q-value (mucinous) q-value (primary site) q-value (interaction) Age, median (IQR) 61 (49-70) 59 (49-69) 58 (49-67) 58 (50-66) 62 (57-67) 63 (53-72) 0.42 <0.001 0.56 Female (%) 46.3 44.5 53.8 54.9 47.1 40.2 0.82 <0.001 0.82 Potentially actionable alteration (%) 47 40.2 45.7 34.4 52.9 40.6 <0.001 0.025 0.56 KRAS (%) 52.2 40.0 70.5 36.3 5.9 12.8 <0.001 <0.001 <0.001 TP53 (%) 40.9 63.6 22.0 29.5 41.2 48.8 <0.001 <0.001 <0.001 APC (%) 45.2 58.3 5.8 9.5 17.6 <jats:td colspan="1"
Wu et al. (Thu,) studied this question.