Conformal radiation therapy (CRT) decreases the morbidity of prostate cancer treatment, but no published data attest to the improved ability of CRT to control disease. Therefore, we compared Prostate-Specific Antigen (PSA) response at 1 year among similarly staged patients treated by conformal techniques to those treated with conventional approaches, looking for an early indicator of tumor response. Patients with locally advanced disease were treated by pelvic field followed by prostate field conedowns; those with early stage/low grade disease received only prostate field irradiation. Between October, 1987 and November, 1991, conventional treatments used rectangular beams with or without corner blocks. Neither urethrography nor immobilization casts were used for conventionally treated patients. Between April, 1989 and December, 1992, conformal treatments have used rigid immobilization and Computed Tomography-based, beams-eye-view field design. As such, our conformal approach allowed improved targeting. Median prescribed doses (minimal doses to the Planning Target Volume) were 70 Gy (66-73 Gy) and 70.2 Gy (64.8-75 Gy) for conventionally and conformally treated patients, respectively. Median daily fraction size was 1.8 Gy for conventional treatment and 2.0 Gy for conformal therapy. Baseline PSA data were available on 170 consecutive patients treated conformally and 90 consecutive patients treated conventionally. Among those receiving only prostatic field irradiation, 12-month PSA values returned to normal in 96% and 85% of conformally and conventionally treated patients, respectively, when normalization was defined as < or = 4 ng/ml (p < 0.03) and in 76% vs. 55% of patients when PSA normalization was defined as < or = 1.5 ng/ml (p < 0.02). Among those receiving pelvic irradiation prior to prostatic conedown, PSA normalization (< or = 4 ng/ml) occurred in 82% and 61% (p < 0.01) of conformally and conventionally treated patients, respectively, and in 56% vs. 38% of patients when normalization was defined as < or = 1.5 ng/ml (p < 0.05). In a multivariate analysis, pretreatment PSA level (< or = 15 vs. > 15), and the use of conformal irradiation were statistically significant prognostic discriminants of PSA normalization at 1 year while total irradiation dose, clinical stage, and the addition of pelvic fields were not significant. As measured by PSA normalization, conformal techniques with improved targeting produced responses that were significantly better than conventional techniques among patients treated with definitive irradiation. These results, coupled with our previously documented reduction of acute and chronic sequelae, support the continued use of CRT as a more effective method of treatment for prostate cancer.