PURPOSE: To review our experience with three-dimensional intensity-modulated radiotherapy (IMRT) in the treatment of nasopharyngeal carcinoma. METHODS AND MATERIALS: We reviewed the records of 35 patients who underwent 3D IMRT for nasopharyngeal carcinoma at the University of California-San Francisco between April 1995 and March 1998. According to the 1997 American Joint Committee on Cancer staging classification, 4 (12%) patients had Stage I disease, 6 (17%) had Stage II, 11 (32%) had Stage III, and 14 (40%) had Stage IV disease. IMRT of the primary tumor was delivered using one of the following three techniques: (1) manually cut partial transmission blocks, (2) computer-controlled autosequencing static multileaf collimator (MLC), and (3) Peacock system using a dynamic multivane intensity-modulating collimator (MIMiC). A forward 3D treatment-planning system was used for the first two methods, and an inverse treatment planning system was used for the third method. The neck was irradiated with a conventional technique using lateral opposed fields to the upper neck and an anterior field to the lower neck and supraclavicular fossae. The prescribed dose was 65-70 Gy to the gross tumor volume (GTV) and positive neck nodes, 60 Gy to the clinical target volume (CTV), and 50-60 Gy to the clinically negative neck. Eleven (32%) patients had fractionated high-dose-rate intracavitary brachytherapy boost to the primary tumor 1-2 weeks following external beam radiotherapy. Thirty-two (91%) patients also received cisplatin during, and cisplatin and 5-fluorouracil after, radiotherapy. Acute and late normal tissue effects were graded according to the Radiation Therapy Oncology Group (RTOG) radiation morbidity scoring criteria. Local-regional progression-free, distant metastasis-free survival and overall survival were estimated using the Kaplan-Meier method. RESULTS: With a median follow-up of 21.8 months (range, 5-49 months), the local-regional progression-free rate was 100%. The 4-year overall survival was 94%, and the distant metastasis-free rate was 57%. The worst acute toxicity was Grade 2 in 16 (46%) patients, Grade 3 in 18 (51%) patients and Grade 4 in 1 (3%) patient. The worst late toxicity was Grade 1 in 15 (43%), Grade 2 in 13 (37%), and Grade 3 in 5 (14%) patients. Only 1 patient had a transient Grade 4 soft-tissue necrosis. At 24 months after treatment, 50% of the evaluated patients had Grade 0, 50% had Grade 1, and none had Grade 2 xerostomia. Analysis of the dose-volume histograms (DVHs) showed that the average maximum, mean, and minimum dose delivered were 79.5 Gy, 75.8 Gy, and 56.5 Gy to the GTV, and 78.9 Gy, 71.2 Gy, and 45.4 Gy to the CTV, respectively. An average of only 3% of the GTV and 2% of the CTV received less than 95% of the prescribed dose. The average dose to 5% of the brain stem, optic chiasm, and right and left optic nerves was 48.3 Gy, 23.9 Gy, 15.0 Gy, and 14.9 Gy, respectively. The average dose to 1 cc of the cervical spinal cord was 41.7 Gy. The doses delivered were within the tolerance of these critical normal structures. The average dose to 50% of the right and left parotids, pituitary, right and left T-M joints, and ears was 43. 2 Gy, 41.0 Gy, 46.3 Gy, 60.5 Gy, 58.3 Gy, 52.0 Gy, and 52.2 Gy, respectively. CONCLUSION: 3D intensity-modulated radiotherapy provided improved target volume coverage and increased dose to the gross tumor with significant sparing of the salivary glands and other critical normal structures. Local-regional control rate with combined IMRT and chemotherapy was excellent, although distant metastasis remained unabated.