Optimal number of lymph node stations associated with improved survival among patients with non-small cell lung cancer (NSCLC).

Abstract

8571 Background: Good quality pathologic nodal evaluation is associated with improved long-term survival in NSCLC after surgical resection. However, there is no universal definition of quality. We examined the association between number of lymph node stations (NLNS) sampled and overall survival. Methods: We examined the population-based Mid-South Quality of Surgical Resection (MS-QSR) cohort, excluding patients with neoadjuvant therapy, positive margins, and secondary resections. We summarized demographic and clinical characteristics with appropriate statistics. We employed four Cox regression models: 1) including NLNS as the only predictor; 2) model 1, adjusting for age, sex, histology, clinical stage, extent of resection, and surgical technique; 3) examined the impact of pathologic staging (pStage in 3 groups: IA/IB[<4cm tumor size]; IB[>4cm]/II/IIIA[pN2-]; pN2]) by modeling interaction with pStage and NLNS; 4) model 3 plus adjustments (no clinical stage). In a cut point analysis to identify the optimal NLNS for improved survival, we dichotomized the NLNS for cut points 0 to 12. Hazard ratios (HR) were calculated (as above) and plotted for each dichotomization. Results: The 3916 eligible patients were 21% Black, 54% male, 15% on Medicaid and 49% Medicare; 84% and 85% had clinical and pathologic stage I/II, respectively; 84% had no invasive staging; 84% had a PET-CT scan. For every additional lymph node station sampled, the HR decreased by 0.94 (95% confidence interval: 0.92, 0.96). After adjustments, the HR remained significant (Table). Our pStage cohorts comprised of 2474 (64%) IA/IB[<4cm]; 1080 (28%) IB[>4cm]/II/IIIA[pN2-]; and 290 (8%) pN2+. Once stratified, HRs remained significant among those with early stage but not for pN2+ (Table). When further adjusted, only pStage IA/IB[<4cm] patients had a significant HR, 0.94 (0.91, 0.97). The cut point analysis indicated sequential survival improvement up to ≥ 7 total stations sampled, consistently across all models/adjustments. Conclusions: The number of lymph node stations examined is directly associated with survival after lung cancer resection. Examination of a total of 7 intrapulmonary, hilar and mediastinal stations seems optimal. Hazard ratios (95% confidence intervals) for modeling hazards as a function of lymph node stations sampled (model 1), further adjusted for age, sex, histology, clinical stage, extent of resection, and surgical technique (model 2), including interaction with pathologic stage with no adjustments (model 3), and interaction with pathological stage with adjustments (no clinical stage) (model 4). [Table: see text]