Utilization and Survival Outcomes of Surgical Resection for Clinical Stage IIIB Non-Small Cell Lung Cancer

R. C. Jacobs¹, J. E. Williams², A. B. Chang³, S. N. Bharadwaj³, K. C. Lung³, D. M. Avella Patino³, A. Bharat³, D. D. Odell², S. S.. Kim^3
1Northwestern University Feinberg School of Medicine, Boston, Massachusetts ²University of Michigan, Ann Arbor, Michigan ³Northwestern University, Chicago, Illinois

Purpose: To 1) evaluate predictors of utilization of surgical resection for patients diagnosed with clinical stage IIIB non-small cell lung cancer (NSCLC), 2) analyze survival outcomes of treatment strategies utilizing surgical resection compared to no surgical resection, and 3) evaluate predictors of survival for patients with clinical stage IIIB NSCLC.

Methods: Patients diagnosed with clinical stage IIIB (cT3-T4N2, cT1-T2N3) NSCLC were identified within the National Cancer Database NSCLC Participant User File (PUF) from 2010-2021. Treatment strategies were stratified into those involving surgical resection versus no surgical resection. Multivariable logistic regression was used to identify demographic and clinical predictors of utilization of surgical resection compared to no surgical resection. Patients were propensity score-matched into cohorts undergoing surgical resection versus no surgical resection using 17 covariates. A multivariable Cox regression model and Kaplan-Meier curve with log-rank test were estimated to evaluate the association of overall survival by propensity score-matched cohorts of patients undergoing surgical resection versus no surgical resection. From the propensity score-matched cohort, an exploratory subgroup analysis using multivariable Cox regression modeling was estimated to compare survival outcomes for patients undergoing surgical resection, stratified by use of neoadjuvant chemotherapy, neoadjuvant chemoradiation, and neoadjuvant chemotherapy-immunotherapy (chemo-IO).

Results: There were 42,316 patients with clinical stage IIIB NSCLC included in unadjusted analysis, with 5,140/42,316 (12.2%) of patients undergoing surgical resection and 37,176/42,316 (87.9%) of patients undergoing no surgical resection. Predictors of utilization of surgical resection include private insurance compared to Medicare (aOR 1.28 95% CI [1.16-1.42]), highest quartile compared to lowest quartile income (aOR 1.26 95% CI 1.10-1.44]), and highest quartile compared to lowest quartile hospital-level annual surgical volume (aOR 6.57 95% CI [5.50-7.85]). There were 7,423 patients in the propensity score-matched model. On multivariable propensity score-matched Cox regression, patients undergoing surgical resection had a lower risk of death compared to those undergoing no surgical resection (Figure). Predictors of overall survival included cT3N2 5-7 cm or cT3N2 with invasion compared to cT4N2 with invasion (aHR 0.74 95% CI [0.65-0.84] and aHR 0.84 95% CI [0.78-0.91], respectively). No difference in overall survival was seen for patients with cT4N2 >7 cm compared to those with cT4N2 with invasion (aHR 1.08, 95% CI [0.97-1.21]). On subgroup analysis in the propensity score-matched cohort, no survival differences were seen between patients undergoing surgical resection with neoadjuvant chemotherapy, neoadjuvant chemoradiation, or neoadjuvant chemo-IO (Table).

Conclusion: Most patients with clinical stage IIIB NSCLC did not undergo surgical resection, and predictors of utilization of surgical resection included private insurance, higher income, and higher hospital-level surgical volume status. Patients undergoing surgical resection had greater overall survival compared to those undergoing no surgical resection. Predictors of overall survival include cT3N2 compared to cT4N2, and no differences in overall survival between clinical T stage by size or invasion were observed. Further investigation is needed to understand how to best select patients for surgical resection and improve access to surgical care in patients with resectable clinical stage IIIB NSCLC.

Identify the source of the funding for this research project: National Cancer Institute training grant - 5R38CA245095