Radical prostatectomy is the most effective form of therapy for prostate cancer when the tumour is organor specimen-con®ned at the time of surgery. Many preoperative clinical variables have been proposed to predict pathological stage and to identify patients at risk for failure. Of the histological variables, the Gleason histological score has been studied most extensively and has been shown to be reliable prognostic value. Some authors, however, have pointed out that the subjective histological grading required for the Gleason score is characterised by low reproducibility. Furthermore, the Gleason score depends solely on the architecture of prostate cancer tissue and does not consider features related to the size, shape or arrangement of the nucleus or proliferation. Estimates of volume-weighted mean nuclear volume (MNV) are the only means by which unbiased estimates of three-dimensional parameters can be obtained from a single two-dimensional section, with stereological methods. We have reported that MNV calculated by stereology is a powerful predictor of the prognosis of prostate cancer Ð as is cancer cell architecture, on which Gleason score depends. The present study was conducted to elucidate the role of MNV in predicting tumour biology for patients treated with radical prostatectomy.