Retrospective analysis of the impact of age on overall survival in patients with non-small cell lung cancer.
e18018 Background: Clinical trials have failed to demonstrate that age is a significant prognostic indicator among patients treated for non-small cell lung cancer (NSCLC). Clinical trials do not necessarily represent real-world experience, however. We sought to analyze the impact of age on survival in patients in the National Comprehensive Cancer Network (NCCN) NSCLC Outcomes Database. Methods: We performed a retrospective analysis of 6,834 NSCLC patients from the NCCN NSCLC Database representing 8 NCCN institutions. Of this population, 4,943 patients were eligible for our analysis. Exclusion criteria included the following: alive patients with < 180 days of follow-up, patients with incomplete staging, and patients with a prior cancer diagnosis. The study population was separated into five age quintiles with equal number of patients in each group. Variables included institution, smoking status, gender, race, Charlson comorbidity score, ECOG performance status (PS), histology, stage, and receipt of resection, drug and radiation therapy. Multivariable Cox model was performed for the effect of age on survival after adjusting for the above variables. Model assumptions were evaluated via graphs and residual tests. Results: Across the five quintiles (< 54, 54-60, 61-66, 67-72 and ≥ 73) there was a trend towards lower stage and higher Charlson score with increasing quintile. In addition, there was an increased proportion of patients with squamous cancer in the older age group. In the adjusted Cox model, there was a statistically significant longer survival in each of four younger quintiles compared to the reference group of ≥ 73 years of age (p=0.01). The adjusted hazard ratio of death for patients < 54 was .82 (95% CI = .72 to .94), for patients 54-60 was .86 (95% CI = .76 to .97), for patients 61-66 was .84 (95% CI = .74 to .95), and for patients 67-72 was .84 (95% CI = .74 to .95). There were no statistically significant pairwise interactions among age, smoking status and stage. Conclusions: Even after adjusting for institution, comorbidity scores, smoking status, race, gender, ECOG PS, histology, stage and treatment, NSCLC patients who were ≥ 73 years of age had a worse survival when compared to younger age groups.