To ascertain which variables of bladder urothelial carcinoma (UC) might be useful in predicting either UC involving the prostate (UCP) or incidental prostate adenocarcinoma in radical cystoprostatectomy specimens.

The bladder and whole-mount prostate sections of 248 radical cystoprostatectomy specimens were reviewed. Stepwise discriminant analysis was used to predict UCP or incidental prostate adenocarcinoma.

UCP was present in 94 patients (37.9%). UC originated from the prostatic urethra and periurethral ducts in 78 (31.45%), and isolated direct extension of UC from the bladder was present in 16 patients (6.45%). The periurethral ducts coexisted with direct extension of bladder UC in 11 patients (4.4%). Prostate adenocarcinoma was identified in 123 patients (49.6%). Carcinoma in situ and high-grade urothelial papillary carcinoma were seen in 8 (3.2%) and 5 (2.0%) patients, respectively. In 57 (23%), 64 (25.8%), and 87 (35.1%) patients, UC had invaded the subepithelial connective tissue, muscularis propria, and perivesical tissue, respectively. UC was multifocal in 53 patients (21.4%). The tumor was in the trigone and bladder neck in 160 patients (64.5%). Of the 248 patients, 98 (39.5%) had a history of recurrence. Stepwise discriminant analysis selected 3 variables of bladder UC (previous recurrence and location and number of foci) and correctly predicted the group in 72.2% of patients without and with UCP. Discriminant analysis selected 2 variables of bladder UC (focality and previous recurrence) and correctly predicted the group in 57.7% of patients without and with prostate adenocarcinoma.

Our approach can identify bladder UC variables that could guide urologists in the selection of the most appropriate surgical procedure.