We examined the usefulness of the volume-adjusted prostate-specific antigen (PSA) parameters for prediction of T1c prostate cancer on 210 patients who had abnormal PSA levels but no abnormal findings in digital transrectal examination (DRE) or transrectal ultrasonography (TRUS). PSA, prostate volume (PV), transition zone volume (TZV), PSAD (PSA/PV) and PSATZD (PSA/TZV) were assessed with the receiver operating characteristic (ROC) curve and the area under the curve (AUC). Simple and stepwise logistic regression models were used to calculate the odds ratios of these parameters. Fifty-three (25.2%) of all 210 patients and 31 (19.9%) of 156 patients with intermediate PSA levels had biopsy-proved prostate cancer. The ROC curves of all patients revealed that PSA, PV, TZV, PSAD and PSATZD had significant predictive values, while AUCs of PV, PSAD and PSATZD had significant predictive values as compared to that of PSA. In the patients with intermediate PSA levels, the ROC curves revealed that PV, TZV, PSAD and PSATZD had significant predictive values, but there were no significant differences in AUCs among these parameters. The stepwise logistic regression analysis showed that PV and PSATZD were significant predictive parameters in all patients and that PSATZD was the only significant predictive parameter in the patients with intermediate PSA levels. In conclusion, not only PSAD and PSATZD but also PV and TZV had significant predictive values in discriminating prostate cancer. However, the multivariate analysis showed that PSATZD had the strongest predictive value in all patients and in those with intermediate PSA levels.