Purpose: Our center contributed 183 patients to the Asian–Oceanian Clinical Oncology Association (AOCOA) multicenter randomized trial comparing induction chemotherapy (CT) followed by radiotherapy (RT) vs. RT alone in patients with locoregionally advanced undifferentiated nasopharyngeal carcinoma (NPC). In a preliminary report no difference in terms of overall survival or relapse-free survival was found between the 2 treatment arms. To study the long-term outcome and patterns of failure after CT for NPC, we analyzed our own center data for which a uniform radiation treatment protocol was adopted and a longer follow-up time was available. Methods and Materials: Between September 1989 and August 1993, a total of 183 patients were recruited into the AOCOA randomized study from our center. Patients with newly diagnosed NPC of Ho’s T3 disease, N2–N3 disease, or with neck node size of at least 3 cm were eligible. Stratification was made according to the nodal size (≤ 3 cm, >3– 6 cm, > 6 cm). Patients were randomized to receive 2–3 cycles of CT with cisplatin 60 mg/m 2 and epirubicin 110 mg/m 2 D1 followed by RT or RT alone. Four patients were excluded from the current analysis (2 died before treatment, 2 received treatment elsewhere). The remaining 179 patients were randomized to the two treatment arms, with 92 to the CT arm and 87 to the RT arm. Two patients in the CT arm had RT only, and all patients completed radiation treatment. Overall survival (OAS), relapse-free survival (RFS), local relapse-free survival (LRFS), nodal relapse-free survival (NRFS), and distant metastases-free survival (DMFS) were analyzed using Kaplan–Meier method and significance of survival curve differences calculated using log–rank test. Analysis was performed based on the intent-to-treat. Results: The median follow-up was 70 months. At the time of analysis, 50% of patients in the CT arm and 61% in the RT arm had relapse, while 32% in the CT arm and 36% in the RT arm had died of the disease. The median RFS was 83 months in the CT arm and 37 months in the RT arm. The median OAS has not yet been reached for both arms. No significant differences were found for the various endpoints, although there was a trend suggesting better nodal control in the CT arm. The 5-year rates for the various endpoints in the CT arm vs. the RT arm were: 53% vs. 42% for RFS ( p = 0.13), 70% vs. 67% for OAS ( p = 0.68), 80% vs. 77% for LRFS ( p = 0.73), 89% vs. 80% for NRFS ( p = 0.079), and 70% vs. 68% for DMFS ( p = 0.59). There was also no significant difference in the patterns of failure between both arms: in the CT arm, 28% of failures were local only, 13% regional only, 4% locoregional, 44% distant, and 11% mixed locoregional and distant. In the RT arm, 23% of failures were local only, 13% regional only, 11% locoregional, 43% distant, and 9% mixed locoregional and distant. Conclusion: Induction chemotherapy with the regimen used in the current study did not improve the treatment outcome or alter the failure patterns in patients with locoregionally advanced NPC, although there was a trend suggesting better nodal control in the combined modality arm. Alternative strategies of combining chemotherapy and radiotherapy should be tested and employed instead.