Summary Background Two phase II trials in patients with previously-treated advanced non-small-cell lung cancer suggested that gefitinib was efficacious and less toxic than was chemotherapy. We compared gefitinib with docetaxel in patients with locally advanced or metastatic non-small-cell lung cancer who had been pretreated with platinum-based chemotherapy. Methods We undertook an open-label phase III study with recruitment between March 1, 2004, and Feb 17, 2006, at 149 centres in 24 countries. 1466 patients with pretreated (≥one platinum-based regimen) advanced non-small-cell lung cancer were randomly assigned with dynamic balancing to receive gefitinib (250 mg per day orally; n=733) or docetaxel (75 mg/m2 intravenously in 1-h infusion every 3 weeks; n=733). The primary objective was to compare overall survival between the groups with co-primary analyses to assess non-inferiority in the overall per-protocol population and superiority in patients with high epidermal growth factor receptor (EGFR)-gene-copy number in the intention-to-treat population. This study is registered with ClinicalTrials.gov , number NCT00076388. Findings 1433 patients were analysed per protocol (723 in gefitinib group and 710 in docetaxel group). Non-inferiority of gefitinib compared with docetaxel was confirmed for overall survival (593 vs 576 events; hazard ratio HR 1·020, 96% CI 0·905–1·150, meeting the predefined non-inferiority criterion; median survival 7·6 vs 8·0 months). Superiority of gefitinib in patients with high EGFR-gene-copy number (85 vs 89 patients) was not proven (72 vs 71 events; HR 1·09, 95% CI 0·78–1·51; p=0·62; median survival 8·4 vs 7·5 months). In the gefitinib group, the most common adverse events were rash or acne (360 49% vs 73 10%) and diarrhoea (255 35% vs 177 25%); whereas in the docetaxel group, neutropenia (35 5% vs 514 74%), asthenic disorders (182 25% vs 334 47%), and alopecia (23 3% vs 254 36%) were most common. Interpretation INTEREST established non-inferior survival of gefitinib compared with docetaxel, suggesting that gefitinib is a valid treatment for pretreated patients with advanced non-small-cell lung cancer. Funding AstraZeneca.