Prostate-specific antigen density as the best predictor of low- to intermediate-risk prostate cancer: a cohort study.

Abstract

Evaluation of prostate cancer (PCa) when serum prostate-specific antigen (PSA) level is vaguely elevated is complicated. This is because serum PSA levels only reflect the number of prostate epithelial cells. We aimed to compare PSA and various prostate volume-related factors to determine which one can best predict PCa in patients with a PSA level of 2.5-20 ng/mL. Patients who underwent transrectal ultrasound (TRUS)-guided prostate biopsy at the Inje University Sanggye Paik Hospital between January 2018 and July 2021 and who had a PSA level of 2.5-20 ng/mL were retrospectively identified (n=275). Among them, based on biopsy results, patients were divided into cancer group and non-cancer groups, and age, PSA, total prostate volume (TPV), peripheral zone volume (PZV), peripheral zone PSA density (PZ-PSAD), transitional zone-PSAD (TZ-PSAD), and PSAD were compared and analyzed using receiver operating characteristic (ROC) and univariate analyses. The areas under ROC curves (AUCs) for age, total PSA, TPV, PZV, PZ-PSAD, TZ-PSAD, and PSAD for predicting PCa in patients with a PSA level of 2.5-20.0 ng/mL were 0.678, 0.680, 0.671, 0.639, 0.731, 0.736, and 0.764, respectively. In univariate and multivariate analysis, all categorical variables were divided based on the cut-off value and used to predict PCa. Those with a PSAD of ≥0.218 ng/mL2 were found to be at an increased risk of PCa than those with a PSAD of <0.218 ng/mL2 [odds ratio (OR) =3.51; 95% confidence interval (CI): 1.306-9.415], which was the best result, followed by TZ-PSAD with a cut-off value of 0.353. At a PSAD level of 0.218 ng/mL2, 85.0% of the PCa group could avoid unnecessary biopsy and 61.4% of the non-PCa group could reduce missed diagnosis when the TRUS findings were inaccurate. PSAD may inform biopsy decisions as the best predictor of PCa when TRUS findings are ambiguous in patients with a PSA level of 2.5-20.0 ng/mL.