Combination and multivariate analysis of PSA-based parameters for prostate cancer prediction

Abstract

The aim of this study was to evaluate the ability of prostate-specific antigen (PSA)-based parameters including PSA density (PSAD), PSAD of the transition zone (PSA-TZ), percent free PSA, PSA velocity, and their combination to enhance the specificity of PSA for prostate cancer detection in men with serum PSA levels between 4 and 10 ng/mL. We evaluated prospectively 559 consecutive men referred for early detection of prostate cancer who had serum PSA levels between 4 and 10 ng/mL. All men underwent prostatic ultrasonography and sextant biopsy with two additional TZ biopsies. In all cases, if first biopsies were negative an additional set of biopsies was obtained within 6 weeks. The ability of PSAD, PSA-TZ, PSA velocity, percent free PSA, and their combination to improve the detection of prostate cancer was evaluated by univariate and multivariate analysis as well as receiver operating characteristic (ROC) curves. In this prospective study of 559 patients, 217 had prostate cancer and 342 had histologically confirmed benign prostatic hyperplasia. Multivariate analysis and ROC curves showed that PSA-TZ and percent free PSA (f/t PSA) were the most powerful and highly significant predictors of prostate cancer. Areas under the ROC curve (AUC) for PSA-TZ and percent free PSA were 0.827 and 0.778, respectively (p = .01). Combination of f/t PSA with PSA-TZ (AUC = 88.1%) significantly increased AUC as compared to each of the other parameters alone as well as their combination (p = .02). The next best combinations were PSA-TZ + PSAD, PSA-TZ + PSA, and f/t PSA + PSA. PSA-TZ followed by f/t PSA and PSAD were the most powerful predictors of prostate cancer in referred patients with a serum PSA between 4 and 10 ng/mL. f/t PSA + PSA-TZ was the most effective combination. When volume-independent PSA parameters were taken into consideration, f/t PSA + PSA clearly outperformed the other options.