INTRODUCTION: Existing guidelines for surveillance after non-small-cell lung cancer (NSCLC) treatment are inconsistent and have relatively sparse supporting literature. This study characterizes detection rates of metachronous and recurrent disease during surveillance with computed tomography scans after definitive treatment of early stage NSCLC. MATERIALS AND METHODS: The incidence of metachronous and recurrent disease in patients who previously underwent complete resection via lobectomy for stage IA NSCLC at a single center from 1996 to 2010 were evaluated. A subgroup analysis was used to compare survival of patients whose initial surveillance scan was 6 ± 3 months (early) versus 12 ± 3 months (late) after lobectomy. RESULTS: Of 294 eligible patients, 49 (17%) developed recurrent disease (14 local only, 35 distant), and 45 (15%) developed new NSCLC. Recurrent disease was found at a mean of 22 ± 19 months, and new primaries were found at a mean of 52 ± 31 months after lobectomy (P < .01). Five-year survival after diagnosis of recurrent disease was significantly lower than after diagnosis of second primaries (2.3% vs. 57.5%; P < .001). In the subgroup analysis of 187 patients, both disease detection on the initial scan (2% [2/94] vs. 4% [4/93]; P = .44) and 5-year survival (early, 80.8% vs. late, 86.7%; P = .61) were not significantly different between the early (n = 94) and the late (n = 93) groups. CONCLUSION: Surveillance after lobectomy for stage IA NSCLC is useful for identifying both new primary as well as recurrent disease, but waiting to start surveillance until 12 ± 3 months after surgery is unlikely to miss clinically important findings.