Is tumor volume an independent prognostic factor in clinically localized prostate cancer?

Purpose: There continues to be debate regarding the prognostic significance of tumor volume (TV) in radical prostatectomy (RP) specimens. We assessed the prognostic significance of TV in a large series of patients followed for a long time to discover whether the effect of TV has changed with earlier detection of smaller tumors. Materials and Methods: TV was measured planimetrically in 1,302 consecutive RP specimens with clinical stage T1-3 prostate cancer from 1983 to 2000. We correlated TV with standard clinical and pathological features, and determined the prostate specific antigen nonprogression rate. Median followup was 46 months (range 1 to 202). Results: TV was weakly associated with other clinical and pathological features. Median TV decreased significantly over time (2.16 cm(3) before 1995 vs 1.25 cm(3) after 1995, p <0.001) and this decrease was also found within each clinical stage. In univariate analysis TV correlated strongly with the probability of progression. However, in multivariate analysis TV was not a significant independent predictor of prognosis, either in the whole cohort of patients or in those with peripheral zone cancer only. Even in univariate analysis TV had no effect on prognosis for patients in whom cancer was either confined to the prostate or was Gleason score 2 through 6. Conclusions: TV provides no independent prognostic information when considered in multivariate analysis with Gleason score and pathological stage. Measurement of TV before treatment is less likely to characterize prostate cancer accurately than assessment of tumor grade and extent. There seems to be little reason to measure TV routinely in RP specimens.