Utility of the transition zone index for identification of prostate cancer in Chinese men with intermediate PSA levels.

Purpose: We aimed to investigate the utility of the transition zone index (TZI) for identification of prostate cancer (PCa) in Chinese men with prostate-specific antigen (PSA) levels of 4-10 ng/mL.

Methods: In this retrospective cohort study, results of transrectal ultrasonography (TRUS)-guided biopsy were assessed in 616 consecutive Chinese men; all subjects had intermediate serum PSA levels and normal digital rectal examination findings. The prostate and transition zone volumes were determined by TRUS. A TZI cutoff value of 0.47 produced the best sensitivity and specificity rates in receiver operating characteristic (ROC) curve analysis and thus was used to classify the study subjects into two subgroups: group 1 with TZI ≤ 0.47 and group 2 with TZI > 0.47. Logistic regression analysis was used to predict outcomes. The variables that were statistically significant in the stepwise logistic regression analysis were assessed using the ROC curve and the area under the curve.

Results: Overall, 166 of the 616 patients (26.9%) had histologically confirmed PCa. A total of 238 (38.6%) patients were classified into group 1, of whom 97 (40.8%) exhibited a positive biopsy; and 378 (61.4%) patients were classified into group 2, of whom 69 (18.3%) exhibited a positive biopsy. The stepwise logistic regression analysis revealed that PSA density (PSAD) exhibited the strongest predictive value in the overall population and in group 1, whereas PSA transition zone density (PSATZD) was the optimal predictor in group 2. The ROC curve analysis revealed that when using the TZI-specific 100% sensitivity cutoffs, 17.7% and 25% of the biopsies were unnecessary and could be avoided in the overall patient population prior to and following the division into groups, respectively (P = 0.002). Using an individually generated 95% sensitivity cutoff of 0.12 ng/mL(2) for PSAD and a cutoff of 0.179 ng/mL(2) for PSATZD for TZI-stratified cohorts of TZI ≤ 0.47 and TZI > 0.47, a more consistent specificity of 44% and 46.9%, respectively, for each cohort was observed.

Conclusions: The optimal predictor for PCa differs between various TZI levels. The combination of PSAD in patients with TZI ≤ 0.47 and PSATZD in patients with TZI > 0.47 helps to identify potentially unnecessary biopsies compared to the use of a single PSAD for the entire patient population. The discrepancies regarding an optimal predictor in published reports are most likely due to the differing TZI levels among the cases. In this study, we demonstrated improved identification of PCa using TZI-adjusted cutoffs for PSAD and PSATZD.