Clinical and Genomic Determinants of Survival in EGFR-Mutant Non-Small Cell Lung Cancer: An Integrated Analysis of UK Biobank and cBioPortal Cohorts

Background Prognostic heterogeneity in EGFR-mutant non-small cell lung cancer (NSCLC) remains incompletely understood. Although differences between EGFR mutation subtypes have been widely reported, the extent to which co-occurring genomic alterations contribute to survival variability remains unclear. Methods We conducted an integrated analysis combining population-based clinical data from the UK Biobank (UKB; n = 818) and genomic data from cBioPortal (n = 1,126). Kaplan–Meier analysis, Cox proportional hazards models, inverse probability of treatment weighting (IPTW), Fine–Gray competing-risk models, and restricted mean survival time (RMST) were used to evaluate survival outcomes. A prognostic nomogram was constructed and internally validated. Results In the UKB cohort, age at diagnosis was the only independent clinical predictor of overall survival. Treatment status, smoking status, and body mass index were not independently associated with survival after multivariable adjustment. IPTW, Fine–Gray, and RMST analyses consistently showed no clinically meaningful survival advantage associated with treatment status. In the cBioPortal cohort, the apparent survival difference between EGFR L858R and exon 19 deletions was attenuated after adjustment. TP53 was the most frequent co-mutation and was consistently associated with worse survival, including within subtype-specific analyses. Multivariable analysis identified TP53, PIK3CA, and CDKN2A as independent adverse prognostic factors, whereas EGFR L858R was associated with improved survival. The nomogram showed acceptable calibration but limited discriminatory performance. Conclusions Survival outcomes in EGFR-mutant NSCLC are jointly influenced by clinical characteristics and co-occurring genomic alterations. Age represents a key clinical determinant, while TP53-centered co-mutation patterns provide additional prognostic stratification. These findings may help improve risk stratification and support more individualized clinical management.