Attenuation of long-term survival benefit of biological therapy in metastatic colorectal cancer (mCRC): Real-world data validation and molecular characterization.

Abstract

e15589 Background: Incorporation of biologic drugs to chemo has improved overall survival (OS) of patients with metastatic colorectal cancer (mCRC), as shown by randomized phase 3 clinical trials. Long term survival patterns remain uncharacterized. Analyses of SEER Medicare data showed addition of biologics confer early survival benefit followed by attenuation and lower OS, suggestive of late mortality. We sought to validate our findings in a large real-world cohort to explore molecular and treatment related factors associated with long term outcomes. Methods: A total of 11,048 mCRC tumors underwent molecular profiling at Caris Life Sciences (592 gene, NextSeq; WES, WTS NovaSeq). Genetic ancestry was inferred using ancestry informative variants derived from gnomAD reference. Immune cells were estimated using Quantiseq (WTS). Patients (Pts) were stratified as chemo alone (CT) or added biologics: anti VEGF (CTV) or anti EGFR (CTE). Significance was calculated using X 2 /Fisher’s exact or Mann-Whitney U (p adjusted for multiple comparisons). Clinical outcome was from insurance claims; OS was calculated from tissue collection and real-world OS (rwOS) from therapy start to last contact/death. Cox proportional hazards model was used for Hazard Ratio and log rank for p value, estimating average treatment effects, recognizing potential non proportional and time varying hazards. Results: 4442 patients (40%) received CT, 984 (9%) CTE and 5622 (51%) CTV. Median age in CT was 65 yrs vs CTV (62) and CTE (61), p < 0.05. Left sided tumors received CTE at a higher rate (12.4%) vs right sided (3.6%), p < 0.001. African ancestry pts were more frequently treated with CTV (p = 0.015). CTE had improved OS vs CT; (27.5 vs 23.8 m, HR 0.86; 95% CI 0.79-0.94, p = 0.001), and rwOS (44.4 vs 40.6 m, HR 0.91; 95% CI 0.83-1.0, p = 0.04). Notably, OS curves crossed at ~ 50 m, with observed early benefit attenuated and reversed beyond 50 m. CTV had improved OS vs CT; (25.7 vs 23.7 m, HR 0.94; 95% CI 0.90-0.99, p = 0.013) with OS curves crossing ~ 40 m. No rwOS improvement was observed, while OS curves crossed at 30 m (OS favored CTV before cross point but favored CT after cross point). Molecular comparison showed significant differences including enriched Tregs in CT vs CTE (FC 1.24, q = 0.03) and CTV (FC 1.12, q = 0.03). Conclusions: In a large database of mCRC; CTE and CTV pts had better OS vs CT with crossing of the survival curves matching late mortality observed in the SEER data. CTV was the predominant treatment of choice, with benefit of adding anti VEGF up to but not lasting past 30 ms. For pts with survival 30-50 m, long-term use of CT with a biologic presented a key point beyond which treatment may no longer be beneficial. This work provides a foundation for investigating risks vs. benefits of selective, monitored deployment of biologics with CT over time.