Estimation of prostate cancer volume is a potentially important application of transrectal ultrasound imaging. Preoperative volume estimation was performed in 110 patients undergoing transrectal prostatic ultrasonography and subsequent radical prostatectomy. Several different methods for calculation of cancer volume were compared. Most of the sonographic methods that have proved to be useful for estimation of prostate gland volume were inaccurate in estimation of the cancer volume in the prostatectomy specimen. All methods underestimated the cancer volume in the majority of the cases. Step-section planimetry provided the highest correlation with cancer volume (r = 0.84) but the wide range of error makes clinical application impractical. With the current technology, ultrasound imaging alone does not appear to be a useful means to predict cancer volume.