Stage and Access Predict Outcomes in Early Onset NSCLC
Key Clinical Summary
- In a US SEER cohort of 18,595 patients aged 18–50 years with early-onset non-small cell lung cancer (NSCLC), 53.4% presented with stage IV disease and had markedly worse overall survival.
- Modifiable factors—including time from diagnosis to treatment, household income, and rural residence—were independently associated with survival, alongside clinical features (stage and metastases).
- Machine learning and Cox models showed comparable discrimination (C-index ~0.77), highlighting stage, radiation therapy, time to treatment, and income as top drivers of risk.
Early-onset NSCLC remains understudied. A retrospective cohort analysis published in JAMA Network Open used SEER data (2010–2021) to evaluate which demographic, clinical, and socioeconomic factors are associated with overall survival (OS) among US adults 18–50 years old. The authors report substantial stage-at-diagnosis burden and survival disparities linked to modifiable social determinants and treatment timing.
Study Findings
Among 18 595 patients, most were 40 to 50 years old (83.9%) and female (52.2%); 60.0% were non-Hispanic White, 14.5% non-Hispanic Black, 12.6% Hispanic, and 12.0% non-Hispanic Asian or Pacific Islander. More than half (9,929; 53.4%) had stage IV disease at diagnosis; adenocarcinoma was the most common histology (59.0%).
In multivariable Cox analysis, stage IV disease was the strongest factor associated with mortality (HR, 17.47; 95% CI, 15.28–19.96). Additional factors linked to worse survival included liver metastases (HR, 1.45; 95% CI, 1.35–1.54), low household income (HR, 1.45; 95% CI, 1.33–1.58), squamous histology (HR, 1.42; 95% CI, 1.33–1.51), and male sex (HR, 1.19; 95% CI, 1.14–1.25). Chemotherapy (HR, 0.68; 95% CI, 0.64–0.73) and surgery (HR, 0.80; 95% CI, 0.74–0.86) were associated with lower mortality.
Model performance was strong (Cox C-index 0.774; RSF 0.765). Feature importance analyses identified overall stage, radiation therapy, time to treatment, and household income as the most influential factors. Disparities were most pronounced in early-stage disease: for stage I, rural residence (HR, 1.65; 95% CI, 1.23–2.08) and low household income (HR, 1.96; 95% CI, 1.52–2.55) were associated with higher mortality.
Timing to treatment mattered by stage. In stage I disease, initiating treatment within 2 weeks was associated with better survival; starting after >4–6 weeks was worse (HR, 1.53; 95% CI, 1.10–2.12). Conversely, in stage IV disease, treatment within 2 weeks correlated with poorer outcomes, while initiation after >8 weeks showed improved survival (HR, 0.69; 95% CI, 0.63–0.76).
Clinical Implications
For managed care teams in the United States, these findings underscore dual priorities: earlier detection and mitigation of access barriers. The large proportion of stage IV presentations among younger adults limits curative options, suggesting potential value in risk-adapted awareness and diagnostic pathways outside traditional screening criteria. At the plan and system level, reducing socioeconomic friction—travel burden for rural members, financial toxicity, and scheduling delays—may be especially impactful in stage I–II disease, where survival penalties from low income and rural residence were greatest. Operationally, streamlining preauthorization and care coordination to minimize time-to-treatment in early stages, while allowing adequate molecular workup for advanced stages, aligns with the stage-specific associations observed.
Conclusion
In US adults younger than 50 years with NSCLC, survival is most strongly shaped by stage but also by modifiable socioeconomic and care-delivery factors. Health plans and systems can act now by accelerating early-stage treatment, addressing rural and income-related barriers, and tailoring timelines for advanced disease to enable comprehensive workups.
Reference
Kar I, Vhora F, Bou Zerdan M, et al. Survival determinants and sociodemographic disparities in early-onset non-small cell lung cancer. JAMA Netw Open. 2025;8(10):e2537307. Published 2025 Oct 1. doi:10.1001/jamanetworkopen.2025.37307
