Calculating prostate cancer volume preoperatively: the D'Amico equation and some other observations

The primary morphological determinants of cancer progression in the prostate are tumor volume and the percentage of Gleason grades 4 and/or 5 disease. To date the best estimate of cancer volume before therapy has been serum prostate specific antigen (PSA) with Pearson's correlation coefficient value of approximately 0.5. Recently it was reported that prostate cancer volume calculated according to the formula, cancer specific serum PSA/amount of PSA leaking into the serum per cm.3 of cancer, highly correlates with actual cancer volume (R = 0.98). Because there is a definite need for greater accuracy in estimating cancer volume before therapy, we attempt to confirm this proposed equation in our radical prostatectomy series.

We applied this equation to the initial 318 men with peripheral zone cancer treated only with radical prostatectomy at our institution who were followed for a mean and median of greater than 5 years. Calculated prostate cancer volume was determined according to the aforementioned equation with minor modifications, and correlated with the actual cancer volume measured in radical prostatectomy specimens. Pearson's correlation coefficient and the coefficient of determination were calculated using a linear regression model. Calculated prostate cancer volume was also previously used to predict pathological stage pT3. We compared calculated prostate cancer volume, clinical stage, Gleason grade and preoperative serum PSA in logistic univariate and multivariate regressions to predict stage pT3 disease.

Overall correlations for calculated prostate cancer volume were R = 0.537 and R2 = 0.289 (p <0.0001), which are much less than those previously noted (R = 0.98 and R2 = 0.96). As in the original report, we also divided our 318 cases into the 4 cancer volume subgroups of 0.5 cm.3 or less, 0.5 to 2.0, 2.0 to 4.0 and greater than 4.0 cm.3 (R = 0.251, 0.288, 0.382 and 0.462, and R2 = 0.063, 0.083, 0.146 and 0.213 alone respectively). There was an increasing trend for better R and R2 values with increasing prostate cancer volume. Calculated prostate cancer volume was less than 0.5 cm.3 in 156 of our 318 patients (49%) and less than 0 in 37 (11.6%). In these cases serum PSA alone strongly correlated with calculated prostate cancer volume (R = 0.877 and R2 = 0.77). Univariate analysis demonstrated statistical significance for prediction of stage pT3 disease, Gleason grade, clinical stage, serum PSA and calculated prostate cancer volume but multivariate analysis revealed statistical significance only for Gleason grade (p <0.0001) and clinical stage (p <0.0036). Values for PSA and calculated prostate cancer volume were not significant (p = 0.0640 and 0.7920, respectively).

Calculated prostate cancer volume did not predict cancer volume in our 318 patients who underwent radical prostatectomy. While we are uncertain how to interpret the excellent correlation of calculated prostate cancer volume with PSA, we believe that this correlation strongly suggests that most predictive information of calculated prostate cancer volume is related to serum PSA. Importantly in our 318 patients serum PSA was a much stronger predictor of cancer volume than calculated prostate cancer volume. As expected, Gleason grade and clinical stage are excellent predictors of stage pT3 disease but not of serum PSA or calculated prostate cancer volume.