Salvage surgery for non-small cell lung cancer after tyrosine kinase inhibitor treatment

Objectives The prognostic impact of surgical intervention for recurrent or residual non-small cell lung cancer (NSCLC) harboring epidermal growth factor receptor (EGFR) mutation or anaplastic lymphoma kinase (ALK) rearrangement after tyrosine-kinase inhibitor (TKI) treatment remains unclear. We aimed to describe the characteristics and outcomes of patients undergoing salvage surgery in this setting. Methods We retrospectively collected and analyzed nationwide Japanese data on perioperative and postoperative outcomes of patients who underwent salvage surgery after EGFR or ALK-TKI during 2010−2015. The primary endpoint was a 3-year overall survival (OS) rate and secondary endpoints were the rate of adverse events, perioperative mortality rate, 3-year recurrence-free survival (RFS) rate, and median survival time after salvage lung resection. Univariate and multivariate analyses were performed to identify independent prognostic factors of OS and RFS. Results Thirty-six patients were included (EGFR-TKI: 33, ALK-TKI: 3). The 3-year OS and RFS after the surgery were 75.1 % (95 % confidence interval [CI] 55.9–86.9 %) and 22.2 % (95 % CI 8.6–39.7 %), respectively. Of clinicopathological factors, the progression of disease while on TKI and preoperative carcinoembryonic antigen (CEA) levels (≥5 ng/mL) were shown to be worse independent prognosticators of OS (hazard ratio [HR] 9.38, 95 % CI 1.57–55.88, P = .014; HR 4.84, 95 % CI 1.62–14.46, P = .005, respectively). Older age at initial treatment (≥70 years) and advanced pathological T stage (T2-T4) were the worse prognosticators for RFS (HR 12.58, 95 % CI 2.51–62.97, P = .002; HR 3.06, 95 % CI 1.04–9.03, P = .043, respectively). Grade 3 adverse events occurred in 5.6 % (2/36) patients, but no deaths were reported within 90 days after surgery. Conclusion Our study showed that salvage surgery after TKI treatment was safe and feasible and may contribute to prolonged OS time by reducing the local tumor burden.