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

Background: Molecular profiling and programmed death-ligand 1 (PD-L1) status guide first-line treatment selection in metastatic non-small cell lung cancer, yet real-world testing, treatment delivery, and documentation vary across health systems. We described molecular testing documentation, first-line treatment patterns, safety capture, and survival outcomes in a Colombian multicenter registry cohort. Methods: We conducted a retrospective registry-based cohort study using the RECAPC central repository. Data extraction was finalized on October 10, 2025 (registry cutoff). The analytic cohort included patients with metastatic-at-diagnosis disease. Driver status was summarized using a prespecified mutually exclusive cascade (EGFR/ALK altered; Other actionable altered; no driver detected; not tested/unknown). First-line treatment was classified into five categories. Overall survival was estimated with Kaplan-Meier methods and modeled using Cox regression. Safety variables were obtained as binary indicators by treatment line. Results: The cohort included 585 patients; mean age was 72.0 years and 52.0% were female. Driver groups were EGFR/ALK altered (32.8%), Other actionable altered (4.6%), no driver detected (15.0%), and not tested/unknown (47.5%). First-line therapy was chemotherapy alone (34.4%), chemo-immunotherapy (21.0%), immunotherapy alone (3.9%), targeted therapy (26.7%), and not reported (14.0%). PD-L1 expression was <1% in 149 patients (25.5%), 1-49% in 119 (20.3%), and ≥50% in 74 (12.6%); testing was not performed in 156 (26.7%), and data were missing in 87 (14.9%). Among 342 patients with an interpretable PD-L1 result, 193 (56.4%) had PD-L1 ≥1% and 74 (21.6%) had PD-L1 ≥50%. Overall survival was evaluable in 490 patients with 229 deaths; median follow-up was 34.99 months. Survival differed by driver group (log-rank P < 0.001). In adjusted analyses, not tested/unknown remained associated with higher mortality versus EGFR/ALK altered (hazard ratio 2.26; 95% CI, 1.62-3.16; P < 0.001). Age, expressed per 10-year increase, was also associated with mortality (hazard ratio 1.28; 95% CI, 1.10-1.48; P < 0.001), as was ECOG performance status 2 (hazard ratio 1.76; 95% CI, 1.10-2.81; P = 0.02). No PD-L1 category was independently associated with overall survival relative to <1%. A single cutaneous adverse event was recorded in one patient receiving targeted therapy. Conclusions: In this Colombian registry cohort, incomplete molecular characterization and missing first-line regimen documentation were common. In this setting, missing regimen data likely reflect both incomplete capture and discontinuities across care pathways. Findings should be interpreted as observational associations rather than causal effects.