Authors:
Cohen, AB, Neri, B, Adamson, BJS, Scanlon, CM, Gross, CP, Meropol, NJ, Miksad, RA
Background: Risk models for NSCLC have identified age as a prognostic factor, but the relevance of molecular biomarkers to age-related outcomes is undefined. We explored the prevalence of biomarkers by age group and the impact of biomarker status on agerelated differences in OS among aNSCLC patients (pts) in a real-world population. Methods: Pts with stage IIIB, IV or recurrent metastatic NSCLC (diagnosed 1/1/11 to 11/30/18) were identified in a US-based, EHR-derived, deidentified database (Flatiron Health). Cox proportional hazard regressions evaluated OS by age at aNSCLC diagnosis (dx) (< vs 50 years), controlling for gender, histology, race, stage, dx year, smoking status, visit site, region and PDL1 status. Sequential and subgroup regressions modeled ALKþ, EGFRþ and ROS1þ biomarker status as mediators of the age effect on OS. Results: Our sample included 41,024 pts (median age 69 years); median OS was 11 months. Pts <50 years lived longer than pts 50 years (HR: 0.83; p < 0.001). Molecular testing and positive results for ALK, EGFR and ROS1 were higher in age <50 (Table); this finding was maintained for non-squamous pts (p 0.001). Within ALKþ, EGFRþ or ROS1þ subgroups, age was not associated with OS. Sequential variable analysis found 35% of the association between age and OS may be explained by ALK, EGFR and ROS1 biomarker status. Conclusions: Young aNSCLC pts have longer OS than older pts; this benefit is significantly mediated through biomarker status. Pts with mutations had similar OS regardless of age group. As older pts are tested less often and have a lower prevalence of mutations, future work should explore whether increased testing in older pts may mitigate the age-related OS disparity.
Sources:
ESMO Annual Congress