Molecular testing, first-line treatment patterns, and survival in metastatic non–small cell lung cancer in Colombia: The RECAPC multicenter registry.

Abstract

e20670 Background: Real-world Latin American NSCLC cohorts that jointly capture molecular testing, insurance coverage, treatment exposure, and time-to-event outcomes remain limited. We summarized the Colombian RECAPC registry using prespecified, reproducible endpoint definitions. Methods: Registry-based multicenter observational cohort of adults with NSCLC treated across 11 Colombian centers (5 cities) from December 15, 2000, through August 15, 2025; data cutoff October 6, 2025. Primary outcome was overall survival (OS). Progression-free survival (PFS) used patient-level progression dates; when unavailable, a prespecified real-world endpoint based on treatment-line dates (time to next treatment or discontinuation) was applied. OS and PFS were estimated with Kaplan–Meier methods; median follow-up was estimated with reverse Kaplan–Meier. Cox regression evaluated associations with OS and PFS. Results: Among 879 patients, 53% were women and mean age was 71 years (SD, 12); 585 (67%) had metastatic disease. In the full cohort, testing rates were 69% for EGFR, 57% for ALK, and 38% for ROS1; alterations were identified in 180 (20%), 72 (8.2%), and 19 (2.2%) patients, and other actionable variants in 58 (6.6%). Among metastatic patients (N = 585), 244 (42%) had a documented actionable alteration, 235 (40%) were no-driver, and 106 (18%) had unknown driver status. Median follow-up was 28 months; median OS was 22 months (95% CI, 19–28) and median PFS was 14 months (95% CI, 12–17). In patients with a molecular result documented in the medical record (OS-evaluable subset, n = 302), median OS was 36 months in driver-altered tumors versus 17 months in no-driver tumors (log-rank P < 0.001). First-line therapy (mutually exclusive, based on the initial-regimen field) was chemotherapy alone in 229 (39.1%), chemotherapy plus immunotherapy in 117 (20.0%), immunotherapy alone in 37 (6.3%), targeted therapy in 152 (26.0%), and no systemic therapy in 50 (8.5%). Median PFS was 19 months for EGFR (n = 113; P = 0.021 vs no-driver), 26 months for ALK (n = 54; P < 0.001), 19 months for ROS1 (n = 13; P = 0.28), 24 months for other actionable variants (n = 41; P = 0.034), and 12 months for no-driver tumors (n = 161). In multivariable models, ECOG ≥2 (HR, 3.23; 95% CI, 1.97–5.30) and age ≥75 years (HR, 2.02; 95% CI, 1.15–3.56) were associated with shorter OS; for PFS, ECOG ≥2 (HR, 1.68; 95% CI, 1.20–2.36) and > 3 metastatic sites (HR, 2.07; 95% CI, 1.29–3.33) were associated with shorter PFS, while pembrolizumab monotherapy was associated with longer PFS (HR, 0.60; 95% CI, 0.43–0.84). Conclusions: RECAPC provides national real-world metastatic NSCLC data integrating molecular epidemiology, payer type, treatment patterns, and reproducible endpoints.