Unraveling Age-Related Disparities in Lung Cancer: A SEER-Based Survival Analysis Using Traditional and Machine Learning Methods (2000–2022)

Background :Age at diagnosis is a critical determinant of prognosis in lung cancer, yet survival disparities between younger and older patients remain incompletely understood. Traditional survival models often fail to account for non-proportional hazards and complex interactions that may underlie these differences. Methods :We conducted a retrospective cohort study using the SEER 18 registries (2000–2022), including 566,502 patients with primary malignant lung cancer. Patients were stratified into young (≤ 40 years) and older (≥ 60 years) groups. We employed restricted mean survival time (RMST) analysis, landmark survival analysis, Random Survival Forests (RSF), and Joinpoint regression to evaluate absolute survival differences, conditional survival, variable importance, and temporal trends, respectively. Results :Young patients comprised 1.0% of the cohort (n = 5,929) and demonstrated a significantly longer RMST over 2 years (1.31 vs. 0.98 years; difference = 0.33 years, p  < 0.001). Landmark analysis confirmed superior conditional survival in younger patients at 2-, 5-, and 10-year intervals. RSF models revealed differing prognostic feature importance by age, with metastatic stage and surgery most predictive in both groups but chemotherapy and year of diagnosis playing a larger role in younger patients. Joinpoint analysis identified survival inflection points in 2005 for younger patients and 2010 for older patients, suggesting differential uptake of therapeutic advances. Conclusion :Significant age-related survival disparities persist in lung cancer, driven by differences in tumor biology, treatment patterns, and time-dependent survival trends. Age-stratified survival modeling using modern statistical and machine learning tools reveals insights that may inform personalized therapeutic strategies and guide future age-specific clinical trial design.