Prevalence of PD-L1 expression and <i>RAS</i> mutations and association with overall survival in metastatic colorectal cancer patients.

e15572 Background: Higher PD-L1 expression has been shown to be predictive of response to immune checkpoint inhibitors in some tumor types. However, the impact of PD-L1 expression in metastatic colorectal cancer (mCRC) remains poorly understood as well as its relationship to other biomarkers. RAS mutations (RAS mut ) are frequent (40-50%) in mCRC and are linked to resistance to anti-EGFR therapies and poor prognosis. Because RAS mut have been suggested to modulate tumor microenvironment and PD-L1 expression, we examined the prevalence of PD-L1 expression and RAS mut , their co-prevalence, and their association with overall survival (OS) in patients (pts) with mCRC. Methods: We conducted a retrospective observational study on primary tumor samples from adults aged ≥18 years with mCRC, with an initial metastatic diagnosis between January 2013 through December 2018 from Severance Hospital (Republic of Korea) and Aarhus University Hospital (Denmark). RAS mut or RAS wildtype (RAS wt ) status was extracted from clinical records. PD-L1 expression was determined by 22C3 PharmDx IHC assay in which combined positive score (CPS) was categorized as &lt; 1 and ≥1. Cox proportional hazard modeling was conducted to examine associations between the biomarkers (PD-L1 CPS ≥1 and/or RAS mut ) and OS. Results: A total of 297 pts with mCRC were included. The median age was 63 years. Most pts were men (58%), had stage IV CRC at initial diagnosis (67%), pMMR/MSS status (98%), history of liver metastasis (65%), and received systemic therapy (89%). Overall, 26% of pts had a right-sided tumor. For all pts, the median OS was 52 months; the prevalence of PD-L1 CPS ≥1 and RAS mut was 47% and 43%, respectively. The co-prevalence of PD-L1 CPS ≥1 and RAS mut was 21%. RAS mut was statistically significantly associated with worse OS compared to RAS wt (HR 1.41; 95% CI, 1.03-1.94). PD-L1 CPS ≥1 pts had numerically, but not significantly, better OS compared to PD-L1 CPS &lt; 1 pts (HR 0.78; 95% CI, 0.57-1.07). PD-L1 CPS ≥1/RAS wt pts showed a trend towards improved OS (HR 0.77; 95% CI, 0.49-1.20) compared to PD-L1 CPS &lt; 1/RAS wt pts. OS HR was 1.05 (95% CI, 0.67-1.64) for PD-L1 CPS ≥1/RAS mut pts and 1.47 (95% CI, 0.97-2.22) for PD-L1 CPS &lt; 1/RAS mut pts compared to PD-L1 CPS &lt; 1/RAS wt pts, after adjusting for age, gender, ECOG, ever liver metastasis, sidedness of tumor, received systemic therapy, dMMR/MSI, and study site. Conclusions: PD-L1 CPS ≥1 expression and RAS wt were associated with a trend towards improved OS. No OS benefit was observed in PD-L1 CPS &lt; 1/RAS wt pts when compared to PD-L1 CPS ≥1/RAS mut pts, suggesting that the benefit of PD-L1 expression may be countered by the presence of RAS mutations. PD-L1 CPS ≥1 status may therefore carry prognostic value in CRC comparable to RAS mut status. Future studies in CRC are needed to evaluate possible benefits from stratifying arms by both PD-L1 expression and RAS mut status.