A nomogram to predict seminal vesicle invasion by the extent and location of cancer in systematic biopsy results.

PURPOSE We determined whether systematic biopsy results increases the accuracy of standard clinical information in predicting seminal vesicle invasion (SVI). MATERIALS AND METHODS We analyzed a retrospective cohort of 763 patients with clinical stages T1c-T3 prostate cancer who were diagnosed by systematic biopsy and treated with radical prostatectomy. We recorded the location of each biopsy core and measured the length of cancer and total length of each core. Using logistic regression analysis we constructed and internally validated a nomogram to predict SVI. RESULTS A total of 60 patients (7.9%) had SVI. Cancer was present in a biopsy core from the base in 437 patients, of whom 12.8% had SVI compared with only 1.2% of the 326 without cancer at the base. None of the 275 patients with prostate specific antigen (PSA) 10 ng/ml or less and no cancer at the base had SVI. On multivariate analysis serum PSA (p <0.0005), primary Gleason grade (p = 0.028) and percent cancer at the base (p <0.005) were the only significant predictors of SVI. The predictive accuracy of a standard model that included only stage, grade and PSA was maximally enhanced by including the percent cancer at the base (p = 0.0013). A nomogram that incorporated this variable produced probabilities of SVI that differed from the standard model by +/- 10% in 68% of the cases. CONCLUSIONS The presence and amount of cancer in systematic needle biopsy cores from the base of the prostate strongly predicts the presence of SVI. Systematic biopsy results enhance the accuracy of nomograms to predict SVI.