Stereotactic ablative radiotherapy (SABR) is first-line treatment for patients with early-stage non-small cell lung cancer (NSCLC) who cannot undergo surgery. However, up to 1 in 6 such patients will develop isolated local recurrence (iLR) or isolated regional recurrence (iRR). Little is known about outcomes when disease recurs after SABR, or about optimal management strategies for such recurrences.

To characterize long-term outcomes for patients with iLR or iRR after SABR for early-stage NSCLC with the aim of informing treatment decision making for these patients with potentially curable disease.

In this cohort study, a retrospective review was conducted of 912 patients prospectively enrolled in an institutional database at a tertiary cancer center from January 1, 2004, through December 31, 2014.

Overall survival, progression-free survival, recurrence patterns, demographics, salvage techniques, patterns of salvage failure, and toxic effects.

Of the 912 patients in the study (456 women and 456 men; median age, 72 years [range, 46-91 years]), 756 (82.9%) had T1 tumors at initial diagnosis; 502 tumors (55.0%) were adenocarcinomas and 309 tumors (33.9%) were squamous cell carcinomas. Of 912 patients with early-stage I to II NSCLC who received definitive SABR (50 Gy in 4 fractions or 70 Gy in 10 fractions), 102 developed isolated recurrence (49 with iLR and 53 with iRR), and 658 had no recurrence. Median times to recurrence after SABR were 14.5 months (range, 1.5-60.8 months) for iLR and 9.0 months (range, 1.9-70.7 months) for iRR; 39 of 49 patients (79.6%) with iLR and 48 of 53 patients (90.6%) with iRR underwent salvage with reirradiation, surgery, thermal ablation, or chemotherapy. Median follow-up times for patients with iLR or iRR were 57.2 months (interquartile range, 37.7-87.6 months) from initial SABR and 38.5 months (interquartile range, 19.9-69.3 months) from recurrence. Rates of overall survival at 5 years from initial SABR were no different between patients with iLR and salvage treatment (57.9%) and patients with no recurrence (54.9%; hazard ratio, 0.89; 95% CI, 0.56-1.43; P = .65) but were lower for patients with iRR and salvage treatment (31.1%; hazard ratio, 1.43; 95% CI, 1.00-2.34; P = .049). Patients receiving salvage treatment had longer overall survival than patients who did not (median, 37 vs 7 months after recurrence; hazard ratio, 0.40; 95% CI, 0.09-0.66; P = .006). Twenty-four of 87 patients (27.6%) who received salvage treatment for iLR or iRR subsequently developed distant metastases. No patient experienced grade 5 toxic effects after salvage treatment.

Life expectancy after salvage treatment for iLR was similar to that for patients without recurrence, but survival after salvage treatment for iRR was similar to that of patients with stage III NSCLC. Patients who received salvage treatment had significantly improved survival. Because salvage treatment for iLR or iRR was based on a consistent multidisciplinary approach, this may help in clinical decision making.