Diagnostic value of biparameter magnetic resonance imaging of PI-RADS and PSA related markers in first prostate biopsy

Abstract

Objective To evaluate the value of Prostate Imaging Reporting and Data System Version 2 (PI-RADS ) based biparametric magnetic resonance imaging (bpMRI) for predicting prostate biopsy results in patients with elevated prostate specific antigen (PSA). Methods The bpMRI from 539 patients who took transperineal template saturate biopsy from January 2015 to October 2017 were assessed retrospectively. The average age was 69.5 years old (44-88 years), with tPSA level of 7.23 ng/ml (4-10 ng/ml), f/t PSA of 0.183( 0.016-0.504), PSAD of 0.126 ng/ml2 ( 0.025-0.534 ng/ml2) , PV of 72.42 ml ( 18.71-199.51 ml). The age, PSA level, free/total PSA ratio, PSA density, prostate volume, and PI-RADS score of enrolled patients were analyzed for univariate analysis and their difference was compared by chi-square test, t-test. The multivariate logistic regression analysis was also performed through SPSS to select the independent risk factors for prostate cancer (PCa) and clinically significant cancer (csPCa). The receiver operating characteristic curves were also constructed to analyze the sensitivity and specificity of PI-RADS in PCa to explore the best cut-off value for the diagnosis of PCa and csPCa. Results A total of 539 patients were included in our study with 244 cases being positive and 295 cases being negative. In patients with positive results, 59 patients were diagnosed csPCa. According to univariate analysis results, the age(P<0.001) and PI-RADS score (P<0.001) of the positive patients were higher than the negative patients, and the difference was statistically significant. The age of the csPCa patients (P=0.023), PSAD (P=0.048) and PI-RADS scores (P<0.001) were higher than those of InsPCa patients, and f/t PSA (P=0.027) was lower than that of InsPCa patients with statistically significance. Multivariate logistic regression analysis demonstrated that f /t PSA (OR=2.283, P=0.049) and PI-RADS score (OR=9.046, P<0.001) were independent risk factors for positive biopsy results, while PSAD (OR=4.54, P=0.038) and PI-RADS score (OR=8.254, P<0.001) were independent risk factor for csPCa. The Yoden index analysis of different thresholds for prostate cancer detection indicated that PI-RADS 3 was the optimal threshold for the diagnosis of PCa, and PI-RADS 4 was the optimal threshold for the diagnosis of csPCa. Based on the combination of the above factors, the positive rate of prostate cancer was relatively high in patients with PI-RADS score ≥3 and f/t PSA<0.2 , which accounted for 86.6%(181/209). In contrast, the positive rate in patients with a PI-RADS score of ≤2 and f/t PSA≥0.2 was low, which accounted for 10.7%(6/56). The positive rate of csPCa was relatively high in patients with PI-RADS score≥4 and PSAD≥0.15 ng/ml2, which accounted for 76.0%(38/50). The positive rate of csPCa detected in patients with ≤3 and PSAD<0.15 ng/ml2 was low, which accounted for 0(0/359). Conclusions PI-RADS score could be used to reduce the unnecessary prostate biopsies in patients with elevated PSA when combined with other PSA related markers. Patients with a PI-RADS score of ≤3 and a PSAD ratio <0.15 ng/ml2 could avoid unnecessary biopsies. Key words: Prostatic neoplasms; Prostate Imaging Reporting and Data System (PI-RADS); Biparametric magnetic resonance imaging (bpMRI); Prostate specific antigen