Prostate volume estimation using the ellipsoid formula consistently underestimates actual gland size.

PURPOSE

Historically estimating prostate volume by transrectal ultrasound underestimates actual prostate weight. We quantified and determined whether trends could be identified to predict or understand the underestimation.

MATERIALS AND METHODS

The characteristics transrectal ultrasound volume, pathological prostate weight and dimensions in 181 patients were entered into an electronic spread sheet. Pathological and transrectal ultrasound volume was estimated using the standard ellipsoid formula, width x height x length x pi/6. In 87 of the 181 cases transrectal ultrasound dimensions were compared to pathological dimensions.

RESULTS

Using pathologically determined dimensions the ellipsoid formula accurately (+/-10%) predicted weight in 26.5% of the cases vs 13.3% by ultrasound. Transrectal ultrasound underestimated it by greater than 30% in 55% of cases and overestimated (greater than 10%) it in only 6.4%. Small vs large gland weight did not predict less underestimation. For prostate weight less than 30, 30 to 60 and more than 60 gm transrectal ultrasound underestimated by greater than 20% in 22.2%, 24.7% and 25.7% of cases, respectively. Paired analysis of transrectal ultrasound measurements and pathological dimensions revealed that transrectal ultrasound length was accurate (4.4 vs 4.3 cm) and had a good correlation with prostate weight. Height was underestimated (3.2 vs 3.7 cm) but it correlated with weight. Width was inaccurate (4.8 vs 5.2 cm) and it correlated poorly with weight. Age, prostate specific antigen, stage, Gleason score, American Urological Association symptom score and body mass index were not predictive of the underestimation. Displaced water volume in cc per prostate weight in gm showed a correlation of 0.997.

CONCLUSIONS

The primary underlying factor for inconsistency with volume estimation of prostate weight appears to be the ellipsoid formula since pathologically determined dimensions still had a 75% error. Independent of gland size the transrectal ultrasound correlation underestimated weight 80% of the time by greater than 30% in 55% of patients. Contrary to previous reports, transrectal ultrasound width and not length is the least reliable factor.