The natural history of stage A2 prostate cancer is unknown. Previous studies from this institution have shown that, without treatment, a third of the men with clinically localized stage A2 prostatic adenocarcinoma will have disease progression within 4 years. Presently, most patients who present with stage A2 prostate cancer receive surgical or radiation therapy. The degree of differentiation of the tumor (Gleason score) presently is used to predict the prognosis among patients with clinically localized prostate cancer. The Gleason score does well to predict the prognosis for patients with scores of 2 to 4 and 8 to 10. Unfortunately, the majority of patients fall within the range of Gleason scores of 5 to 7. Better methods are needed to predict which patients diagnosed with stage A2 prostate cancer have a high probability of disease progression. Several studies have reported that morphometrically determined nuclear shape descriptors provided accurate separation of these patients that was superior to Gleason grading methods. To evaluate critically the usefulness of nuclear morphometry for prediction of prognosis we developed a system. The Hopkins Morphometry System, that calculated and compared 15 different shape descriptors that were analyzed by 17 different statistical tests. We tested this system on 18 untreated patients with stage A2 prostate cancer with an average followup of 10.5 years (range 5 to 18 years). For each patient 17 statistical analyses of the 15 shape descriptors (255 total) were evaluated and 50 analyses (50 of 255, 19.6%), including average nuclear roundness factor, provided significant separation (p less than 0.01) of the patients on the basis of outcome, whereas the Gleason score (p equals 0.076) did not. The best separation (p less than 0.01) was provided by the lower quartile analysis of the ellipticity shape descriptor.