This study was designed to determine the efficacy and safety of additional maintenance chemotherapy after standard induction chemotherapy/radiation therapy (XRT) in stage III non-small-cell lung cancer (NSCLC). The primary objective was to increase 1-year survival. Eligible patients (N = 220) had confirmed stage IIIA or IIIB NSCLC, and an Eastern Cooperative Oncology Group (ECOG) performance status of 0-1. Patients received induction chemotherapy (paclitaxel 200 mg/m(2) + carboplatin AUC = 6/3 weeks) for 2 cycles, followed by paclitaxel 45 mg/m(2) + carboplatin AUC = 2 weekly ×7 and concurrent daily XRT (cumulative dose = 66.6 Gy in 37 fractions) and then observation or maintenance. Before randomization, 101 patients (46%) discontinued treatment due to progressive disease (n = 34), toxicity (n = 33), patient request (n = 13), death (n = 7), or other (n = 14). The remaining 119 patients were randomly assigned to either "observation" or "maintenance" (6 cycles of paclitaxel 70 mg/m(2)/wk [3 weeks on/1 week off]); a median of 5 of 6 planned cycles was delivered in the maintenance arm. For the observation group vs. the maintenance group, the estimated 1- and 4-year survival rates were 77% vs. 66% and 38% vs. 17% (median, 26.9 months vs. 16.1 months, respectively [P = .07]); the estimated 1- and 4-year performance-free survival (PFS) were 46% vs. 24% and 25% vs. 13% (median, 10.2 months vs. 8.2 months, respectively [P = .04]). Common toxicities were neutropenia, thrombocytopenia, and fatigue. Median survival in both groups surpassed the standard, most notably the 26.9-month survival in the observation group. Maintenance chemotherapy, when added to a regimen of both induction and concurrent chemoradiotherapy, did not improve clinical outcomes, with endpoints favoring the standard arm. Copyright © 2012 Elsevier Inc. All rights reserved.