Abstract
IMPORTANCE: Patients with early-onset non-small cell lung cancer (NSCLC) represent a distinct and relatively understudied population. Identifying factors associated with survival is critical to guiding early detection and treatment strategies. OBJECTIVE: To evaluate demographic, clinical, and socioeconomic factors associated with overall survival (OS) in early-onset NSCLC. DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study obtained data from the Surveillance, Epidemiology, and End Results database and included patients aged 18 to 50 years diagnosed with NSCLC between January 2010 and December 2021. Data analysis was performed between January 1 and March 31, 2025. EXPOSURE: Early-onset NSCLC. MAIN OUTCOMES AND MEASURES: The primary outcome was OS. Cox proportional hazards regression and Shapley additive explanations-enhanced machine learning models were used to improve interpretability and inform targeted interventions. Random survival forest models were used to identify factors associated with OS and assess model performance. RESULTS: Among 18 595 patients with early-onset NSCLC, 9929 (53.4%) presented with stage IV disease. Most patients (15 597 [83.9%]) were aged 40 to 50 years (mean [SD] age, 44.6 [6.0] years), and 9710 (52.2%) were female. In terms of race and ethnicity, 2338 patients (12.6%) were Hispanic; 100 (0.5%), American Indian or Alaska Native; 2229 (12.0%), non-Hispanic Asian or Pacific Islander; 2693 (14.5%), non-Hispanic Black; 11 162 (60.0%) non-Hispanic White; and 73 (0.4%), unknown. In multivariable Cox proportional hazards regression, stage IV disease (hazard ratio [HR], 17.47; 95% CI, 15.28-19.96), liver metastases (HR, 1.45; 95% CI, 1.35-1.54), low household income (HR, 1.45; 95% CI, 1.33-1.58), squamous cell carcinoma histology (HR, 1.42; 95% CI, 1.33-1.51), and male sex (HR, 1.19; 95% CI, 1.14-1.25) were independently associated with worse survival. The random survival forest model demonstrated higher discrimination (C-index 0.765; 95% CI, 0.759-0.771), comparable to the Cox proportional hazards model (C-index, 0.774; 95% CI, 0.769-0.779). Feature importance analysis identified overall stage, radiation therapy, time to treatment, and household income as the most influential factors associated with survival. Among patients with stage I disease, rural residence was associated with a 65% increased risk of death (HR, 1.65; 95% CI, 1.23-2.08) and low household income was associated with a 96% increased risk of death (HR, 1.96; 95% CI, 1.52-2.55). CONCLUSIONS AND RELEVANCE: In this cohort study of 18 595 patients with early-onset NSCLC, most patients were diagnosed at advanced stages. Modifiable factors, including household income, residence, and time to treatment, were associated with survival outcomes. These findings support the need for increased awareness, early detection strategies, and efforts to reduce socioeconomic disparities in adults with NSCLC.