PURPOSE: A phase III randomised trial was conducted in patients with non-metastatic unresectable non-small-cell lung cancer in order to compare, in responders to induction chemotherapy, consolidation treatment by further chemotherapy to chest irradiation. PATIENTS AND METHODS: A total of 462 untreated NSCLC patients were eligible for three courses of induction chemotherapy (MIP) consisting of cisplatin (50 mg/m2), ifosfamide (3 g/m2) and mitomycin C (6 mg/m2). It was proposed that objective responders be randomised to either three further courses of MIP or to chest irradiation (60 Gy; 2 Gy per fraction given over six weeks). RESULTS: An objective response rate of 35% was achieved; 115 patients (including 52% with initial stage IIIA and 44% with initial stage IIIB) were randomised to consolidation treatment, 60 of them to further chemotherapy and 55 to chest radiotherapy. There was no significant difference in survival between the two arms, with a respective median and two-year survival of 42 weeks (95% confidence intervals (95% CI: 35-51) and 18% (95% CI: 8-28) for chemotherapy and 54 weeks (95% CI: 43-73) and 22% (95% CI: 11-33) for irradiation. There was also no statistical difference for response duration between the two arms but chest irradiation was associated with a significantly greater duration of local control than chemotherapy (median duration times: 158 vs. 31 weeks, P = 0.0007). CONCLUSIONS: For non-metastatic unresectable NSCLC treated by an induction chemotherapy regimen containing cisplatin and ifosfamide, if an objective response is obtained, consolidation treatments by further chemotherapy or by chest irradiation result in non-statistically different survival distributions, although a better local control duration is observed with radiotherapy.