OBJECTIVETo develop a nomogram for predicting the probability of prostate cancer at transrectal ultrasound-guided repeat prostate biopsy in Chinese men.

METHODSWe performed repeat biopsy for 170 patients with benign prostate diseases diagnosed on the first biopsy, and analyzed the correlation of positive repeat biopsy with age, prostate volume, PSA, free-to-total PSA (f-PSA/t-PSA), PSA velocity, PSA density, results of digital rectal examination (DRE) and previous histology. We entered the variables stepwise into logistic regression models, and established a nomogram for the risk score on the probability of positive repeat biopsy, whose predictive value was assessed by receiver operating characteristic (ROC) analysis.

RESULTSProstate cancer was detected in 31.8% of the repeat biopsies (54/170). The most accurate predictive nomogram comprised age, PSA, f-PSA/t-PSA, PSA velocity, prostate volume, DRE and previous prostatic intraepithelial neoplasia (PIN) findings. The nomogram exhibited a high predictive value, with the area under the ROC curve (AUC) of 82.4%, significantly greater than that of the prediction based on PSA density (AUC: 66.9%), prostate volume (AUC: 72.6%), PSA velocity (AUC: 69.6%), f-PSA/t-PSA (AUC: 69.3%), or DRE (AUC: 58.5% ) alone.

CONCLUSIONThe nomogram is an accurate multi-variable predicting tool to determine the probability of positive repeat prostate biopsy.