Although prostate specific antigen (PSA) density appears to be an important discriminator between benign and malignant prostatic disease, conflicting data exist concerning its prognostic value. The present study was undertaken to confirm whether PSA density represents a new prognostic indicator of disease-free survival for prostate cancer treated with conformal radio-therapy. Between April 1989 and December 1992, 186 patients with organ confined prostate cancer were treated with definitive irradiation according to previously published conformal guidelines. The PSA density was defined as the ratio of the pretreatment serum PSA (ng./ml.) to the prostate volume (ml.) as determined from treatment planning computerized tomography. The median PSA density was 0.15 with a range of 0.02 to 2.12. A statistically significant advantage in actuarial freedom from biochemical relapse was noted for patients with pretreatment PSA levels less than 15 ng./ml. when compared to those with higher pretreatment PSA levels (3-year freedom from biochemical relapse 85% versus 28%, p < 0.001). Also, patients with PSA density of 0.15 or less had statistically superior freedom from biochemical relapse compared to their counterparts with higher PSA density (3-year freedom from biochemical relapse 88% versus 28%, p < 0.001). In a multivariate analysis only the baseline PSA (p < 0.002) and the Gleason score (p < 0.002) emerged as significant predictors of prolonged freedom from biochemical relapse. The PSA density had no impact on freedom from biochemical relapse whether it was entered into this multivariate model as a continuous or a dichotomous variable. In our data base baseline PSA levels remain the most powerful independent discriminant of response to conformal irradiation. PSA density is only a surrogate for baseline PSA levels and does not refine the ability to predict prolongation of freedom from biochemical relapse following conformal radiotherapy.