Sonographic estimates of prostate volume have been applied to the diagnosis and management of prostate cancer. However, we have observed wide variation between transrectal sonographic estimates of prostate volume (TRUS-V) and the corresponding volumes of unfixed radical prostatectomy specimens (P-V). Because TRUS-V may influence the management of men with prostate disease, the relationship between TRUS-V and P-V was characterized for 100 consecutive men undergoing a radical retropubic prostatectomy (RRP). For the entire cohort, TRUS-V (35.9 +/- 16.9 mL) underestimated P-V (45.4 +/- 22.9 mL, p < .01). Mean signed percentage error between TRUS-V and P-V was -16 +/- 32%. The relationship of TRUS-V to P-V was volume dependent (p < .001). For volumes (P-V) < 30 mL, TRUS-V overestimated P-V. For volumes (P-V) > 30 mL. TRUS-V increasingly underestimated P-V. In summary, transrectal ultrasound estimates of prostate volumes differ significantly from the volume of unfixed RRP specimens. The direction and magnitude of TRUS-V error is volume dependent. Applied clinically, treatment paradigms employing prostate-volume-dependent management algorithms derived from nonuniform methods of volume estimation may not accurately reflect volume nomograms derived from uniform methods of measurement, and thus alter patient management.