Comparative Effectiveness of 5 Local-Regional Control Strategies for IIIA (N2) Non-small Cell Lung Cancer Using SEER Data: Outcomes after Treatment of 20,468 Patients

Xiaozheng Kang,Yanru Wang,Qingyi Wei,Thomas D'Amico,Keneng Chen
DOI: https://doi.org/10.1016/j.chest.2016.02.287
2016-01-01
CHEST Journal
Abstract:SESSION TITLE: Lung Cancer: Diagnosis and Prognosis SESSION TYPE: Original Investigation Slide PRESENTED ON: Saturday, April 16, 2016 at 09:45 AM - 11:15 AM PURPOSE: The rationale of systemic therapy for locally advanced lung cancer is well understood. We evaluated various local-regional control (LRC) strategies for IIIA (N2) non-small cell lung cancer (NSCLC) by assessing the association between LRC and cancer-specific survival (CSS) in a retrospective population-based analysis from the Surveillance, Epidemiology, and End Results (SEER) database (1988 - 2012). METHODS: Eligible patients had IIIA (N2) NSCLC and underwent local therapy. Survival estimates were obtained using the Kaplan-Meier method in both univariate and multivariable Cox proportional hazards models and log-rank tests. Primary outcome, CSS, was defined as the time between a landmark time of 6 months from the date of diagnosis and cancer-specific death, with living patients censored on the date of last follow-up (31-Dec-2012). In a second exploratory analysis, we used propensity-score matching (PSM) to compare outcomes between different LRC modalities. RESULTS: The analysis included 20,468 patients with IIIA (N2) NSCLC divided into six subgroups: no local treatment (NLT), radiotherapy (RT) only, surgery (S) only and the combined modalities (RT/S, S/RT and RT/S/RT). In the multivariable analysis, RT/S was associated with the significantly better survival (hazard ratio [HR] = 0.39, 95% confidence interval [CI] 0.35 - 0.43, P < 0.001, compared with NLT). In the PSM analysis, survival after RT/S was similar to that after RT/S/RT and that after S only (RT/S vs. S only HR = 0.97, 95% CI 0.94 - 1.01, P = 0.094; RT/S vs. RT/S/RT HR = 0.88, 95% CI 0.75 - 1.04, P = 0.135; RT/S/RT vs. S only HR = 1.13, 95%CI 0.86 - 1.34, P = 0.151). CONCLUSIONS: In this large SEER dataset, the inclusion of surgical resection with or without induction or adjuvant RT offers better outcomes than RT without surgery in select patients with IIIA (N2) NSCLC, and thus the best LRC. Additional evaluation of these therapies in randomized trials is needed. CLINICAL IMPLICATIONS: These findings provide a landscape for the potential role of LRC in IIIA (N2) NSCLC. DISCLOSURE: The following authors have nothing to disclose: Xiaozheng Kang, Yanru Wang, Qingyi Wei, Thomas D'Amico, Keneng Chen No Product/Research Disclosure Information
What problem does this paper attempt to address?