Background: Colorectal cancer (CRC) is a leading cause of cancer-related mortality, with 20-25% of cases presenting with metastatic disease (mCRC). Pathogenic variants in the GNAS gene, which encodes the stimulatory alpha subunit of the G protein complex, are found in 1.5% to 4.8% of CRC cases and are more prevalent in mucinous adenocarcinomas. Presence of these mutations have been associated with a lower likelihood of regression following first-line (1L) systemic therapy. However, the specific impact of GNAS mutations on treatment (tx) outcomes and overall survival in mCRC patients (pts) remains under investigation.

Methods: We retrospectively analyzed de-identified data from 5,967 pts with stage IV mCRC treated with 1L oxaliplatin-based chemotherapy in the Tempus Database. Samples were sequenced with the Tempus xT (648-gene panel) or xF DNA assay (105 or 523 genes depending on version) and were divided into GNAS wild-type (GNASwt) and GNAS mutated (GNASmut) groups. Tumor mutational burden (TMB) and microsatellite instability (MSI) were calculated. Short variant pathogenic/likely pathogenic mutations and copy number alterations were analyzed for patients that underwent xT testing. Real-world (rw) objective response rate (rwORR) was defined as the proportion of pts with a documented complete or partial response within 90 days of tx start. Rw overall survival (rwOS) was defined as the time from 1L start to death from any cause. Hazard ratio (HR) was calculated using a Cox proportional hazards model and p-values were calculated using the Wald test.

Results: The study included 5,854 GNASwt and 113 GNASmut mCRC pts (prevalence GNASmut = 1.9%). GNASmut pts exhibited higher prevalence of mucinous adenocarcinoma (26% vs. 2.8%) and peritoneal metastases (51% vs. 22%, p < 0.001), while liver metastases were more prevalent in the GNASwt group compared to the GNASmut group (77% vs 51%, p < 0.001). KRAS, ARID1A, MSH2/3/6, and ATR were more frequently altered in the GNASmut group, while TP53 and APC mutations were more frequent in the GNASwt group. Immunophenotype markers such as high TMB and high MSI were also more often observed in the GNASwt pts compared to GNASmut pts (p < 0.001). rwORR was significantly lower in GNASmut pts compared to GNASwt (42% vs. 66%, p = 0.002). Pts with GNASmut had reduced rwOS compared to pts with GNASwt (HR = 1.31, p = 0.05).

Conclusions: mCRC pts with pathogenic GNAS variants exhibit distinct clinicogenomic features and poorer outcomes with first-line oxaliplatin-based chemotherapy compared to GNASwt pts. These findings highlight the need for alternative tx and further research on GNAS as a prognostic biomarker in mCRC.

VIEW THE PUBLICATION