More accurate noninvasive estimation of prostate size is important in therapeutic trials for benign prostatic hyperplasia. The accuracy of MRI and transrectal ultrasound (TRUS) in assessing prostate weight was evaluated in 48 patients who underwent radical prostatectomy for stage A or B cancer. The volume derived from the wet weight of the freshly excised specimen was used as a reference. We compared that volume with volume estimates derived from the three-axis linear dimension measurement by MRI and TRUS using a tissue density of 1.05 g/cc and the standard formula for an ellipsoid object. Prostate and seminal vesicle volumes were also computed by contouring T2-weighted 5 mm thick contiguous MR images using a semiautomatic edge detection program and pixel summation. Three-axis volume MRI method versus volume from wet weight has slightly less scatter than TRUS three-axis method (r = 0.85 vs r = 0.81). Contoured MR volume method has the least scatter r = 0.93, statistically better than the linear axis method. Contoured MRI volumetric analysis appears superior to linear MRI or TRUS methods in estimating true prostate volume.