OBJECTIVE Pelvic irradiation in addition to prostate irradiation may improve outcome in locally advanced prostate cancer, but is associated with dose-limiting bowel toxicity. We report the preliminary results of a dose escalation study using intensity-modulated radiotherapy. METHODS Eligible patients had high-risk (T3, Gleason > or =8 or prostate-specific antigen > or =20 ng/ml) or lymph node-positive disease. Intensity-modulated radiotherapy was inverse planned giving 70 Gy/35 fractions to the prostate and 50 Gy/55 Gy/60 Gy in sequential cohorts to the pelvis with a 5 Gy boost to positive lymph nodes. Acute and late toxicity were recorded with Radiation Therapy Oncology Group (RTOG) and Late Effects Normal Tissue - Subjective Objective Management LENT-SOM scales. Neoadjuvant androgen suppression was given for 3 years. This report concerns the 50 and 55 Gy cohorts. RESULTS Seventy-nine men were recruited (25 to 50 Gy/54 to 55 Gy) with a median follow-up of 2 years. Patients were divided into two groups according to the total bowel volume outlined (median 450 cm(3)). Acute RTOG (> or =2) bowel toxicity was 40 and 50% for the 50 and 55 Gy groups and 38 and 51% for bowel volume <450 cm(3) and > or =450 cm(3), respectively, suggesting both volume and dose relationships for acute effects. Late RTOG diarrhoea > or =grade 2 was only seen with bowel volume > or =450 cm(3), but no dose effect was apparent (12%/50 Gy and 10%/55 Gy). LENT-SOM bowel > or =grade 2 toxicity occurred in 22%/50 Gy and 15%/55 Gy. Only one patient had grade 3 toxicity. A dose volume histogram analysis showed increased late RTOG diarrhoea > or =grade 2 with larger bowel volume irradiated, significant for BV40 >124 cm(3) (P=0.04), BV45 >71 cm(3) (P=0.03) and BV60 >2 cm(3) (P=0.01). CONCLUSIONS Acute and late bowel toxicity was acceptably low using a pelvic dose of up to 55 Gy over 7 weeks. Both relate to total pelvic bowel volume and dose volume constraints have been defined.