OBJECTIVE To investigate whether tumor volume, an important prognostic factor in prostate cancer, could be estimated from the amount of cancer in multiple core biopsies. METHODS In 80 men, transrectal ultrasound-guided biopsies were taken from focal lesions detected by ultrasound and 8 to 10 standardized positions, including sextant biopsies (apex, midmedial, base) and midlateral and transition zone biopsies. The cancer length in the biopsies was measured. After radical prostatectomy, the prostates were totally embedded, whole-mounted, and tumor volume was measured planimetrically. RESULTS The tumor volume correlated significantly with the total cancer length of all biopsies (r = 0.56) and of the sextant biopsies (r = 0.39). It was found that midlateral and transition zone biopsies provided independent information when included in a multiple regression model with tumor volume as the dependent variable and the sextant biopsies as explanatory variables. All men (n = 6) with less than 3 mm cancer length in only one positive biopsy and a Gleason score less than 7 had a tumor volume less than 1 mL. Nine of 10 men with less than 7 mm of cancer in one positive biopsy and Gleason score less than 7 had tumors smaller than 1 mL. Sextant biopsies did not reliably predict cancer volumes less than 1 mL. CONCLUSIONS The cancer yield of 8 to 10 biopsies correlated better with the volume of prostate cancer than sextant biopsies. This extended biopsy protocol could be used to predict cancers of less than 1 mL in volume.