Outcomes of robotic lobectomy for non-small cell lung cancer in a National Cancer Institute-Comprehensive Cancer Center vs. National Cancer Database

BACKGROUND: There continues to be a rise in the proportion of resectable non-small cell lung cancer (NSCLC) with the recent expansion of criteria for low-dose lung cancer screening. These are increasingly being treated with minimally invasive techniques. Our study aims to compare outcomes of robotic lobectomy (RL) for NSCLC at a National Cancer Institute-designated Comprehensive Cancer Center (NCI-CCC) to those of open lobectomy (OL), video-assisted thoracoscopic lobectomy (VL), or RL as reported in the National Cancer Database (NCDB). METHODS: The first 1,021 patients with NSCLC who underwent RL between 2010 and 2020 were matched with peers from the NCDB who had OL, VL, or RL. Matching was performed based on a propensity score calculated by logistic regression using multiple variables. Surgical outcomes included numbers of examined lymph nodes, performance of mediastinal lymphadenectomy, length of stay (LOS), and 30-day mortality. Kaplan-Meier curves and overall survival (OS) were analyzed using log-rank tests. RESULTS: Most common postoperative complications were persistent air leak, atrial fibrillation, and pneumonia. Median LOS was 4 days, and the 30-day mortality rate was 1% (n=10/1,021). Compared to NCDB patients who underwent OL, NCI-CCC patients had a higher mean number of retrieved lymph nodes (P=0.001), higher rate of mediastinal lymphadenectomy (P<0.001), and shorter median LOS (4 vs. 6 days; P<0.001). There was no difference in 30-day mortality (P=0.176). Kaplan-Meier analyses showed no differences in median OS (log-rank P=0.953) or 5-year OS (P=0.774). Compared to NCDB VL, NCI-CCC patients had a higher nodal yield (P<0.001), higher rates of mediastinal lymphadenectomy (P<0.001), and lower conversion rates (4.1% vs. 13.8%, P<0.001). There were no differences in 30-day mortality (P=0.379) or in median LOS (P=0.351). Kaplan-Meier analyses showed no differences in median OS (P=0.720) or 5-year OS (P=0.735). NCI-CCC patients were also matched with NCDB RL patients and had a higher nodal yield (P<0.001), higher rates of mediastinal lymphadenectomy (P<0.001), and lower conversion rates (4.1% vs. 9.5%; P <0.001). There were no differences in 30-day mortality (P=0.899) or in median LOS (P=0.252). Kaplan-Meier analyses showed no differences in median OS (P=0.484) or 5-year OS (P=0.524). CONCLUSIONS: RL for NSCLC performed in an NCI-CCC appears to have improved perioperative outcomes with comparable long-term OS compared to national benchmarks in OL and VL.