Introduction:

Historically, approximately 70% of men with an elevated PSA did not have any prostate cancer (PCa) following trans-rectal ultrasound guided systematic biopsy (SB) . Today, the consensus amongst the urology community is the need to improve the specificity of PSA testing for detecting clinically significant prostate cancer (csPCa). mpMRI and/or various biomarkers are being recommended as reflex tests to better select men with elevated PSA levels for prostate biopsy with the objective of increasing specificity of PSA screening for detecting csPCa. The advantage of mpMRI is that PIRADS scoring is directly proportional to risk of detecting csPCa and MRI guided biopsy increases detection of csPCa and decreased risk of low risk cancer compared to SB.  The advantages of both blood-based and urine-based PCa biomarkers include lower costs, simplicity of testing, and a standard, quantitative result. The optimal utilization of mpMRI and biomarkers to inform decision to perform prostate biopsy is not well defined. The objective and unique aspect of the present study is to compare the performance of two biomarkers (4Kscore Test and SelectMDx) individually and in combination to predict PI-RADS score in a group of men with an elevated serum PSA under consideration for prostate biopsy.

Methods:

Between  November 2018 to April 2019, all men presenting with the diagnosis of elevated PSA to two uro-oncologists (JW and HL) were advised to undergo a 4Kscore Test, SelectMDx, and mpMRI in order to inform decision whether to perform prostate biopsy. All patients who completed all three tests were included in the study. The total PSA level reported as part of the 4Kscore represented the baseline serum PSA. mpMRI was performed using the 3T clinical instrument and an external phased –array coil and included T2 weighted images, axial diffusion weighted imaging using b values of 50 and 1000 s/mm2 with generation of the apparent diffusion coefficient maps and calculated b-1500 images and dynamic contrast enhancement imaging after intravenous gadolinium chelate. A PI-RADS score was assigned by one of several board certified radiologists with expertise interpreting prostate mpMRI. Risk cut-off points of 7.5% for detecting a csPCa (i.e.,  GGG >1) derived from the 4Kscore and SelectMDx tests were compared for their ability to predict mpMRI PI-RADS scores. Positive Predictive Value, (PPV), Negative Predictive Value (NPV), Sensitivity,  and Specificity were calculated, setting PI-RADS >2 as a positive mpMRI result. The predictive performance of these biomarkers, as well as PSA and PSAD, were evaluated using the area-under-the receiver operator curve.  To combine these two continuous biomarker tests into a signal predictor, a predicted probability was calculated using logistic regression for both mpMRI cut-offs, and then this predicted probability was evaluated as the AUC of the ROC curve.

Results:

Of  the124 men who were advised to undergo both biomarker tests and mpMRI, 103 (83%) completed all three tests. The median age was 63 years (IQR: 58-69) and median PSA  was 5.61ng/ml (IQR: 4.12-8.26).  Thirty men had a  family history of PCa and 16 patients had  abnormal DRE. Biomarker characteristics included median 4Kscore of 11% risk of csPCa (IQR: 6-24) and median SelectMDx risk  score of <2% csPCa (IQR: <2 – 13).  After mpMRI, 30, 22, 36, 13, and 2 patients were found to have PI-RADS scores of 1, 2, 3, 4 and 5, respectively. 

The positive predictive value (PPV), negative predictive value (NPV), sensitivity, and specificity for the individual biomarkers to predict PI-RADS scores are shown in Table 1. The positive and negative predictive value of the individual  biomarkers to predict mpMRI was consistently between 50 to 60%. The superior sensitivity the 4Kscore Test and specificity of SelectMDx, respectively is likely attributed to twice  number of 4K tests that were positive (69% vs. 29%). The ROC curves and AUC values are shown for the serum PSA, PSAD, 4Kscore Test and SelectMDx to discriminate between PI-RADS 1-2 and 3-5 is shown in Figure 1. The rank order of AUC (highest to lowest) was the combination of 4Kscore Test + SelectMDx, 4Kscore Test, PSAD, SelectMDx, and PSA. The increase in the AUC by adding of SelectMDx to the 4Kscore Test was marginal and clinically insignificant.

Conclusion:

Our study provides compelling evidence that biomarkers do not reliably inform decisions whether to obtain an mpMRI in men with elevated PSA. The present study was not designed to determine whether biomarkers or mpMRI is the preferred reflex test to inform decisions whether to proceed with prostate biopsy in men with elevated PSA since it would require all these men undergoing  biopsy independent of mpMRI and biomarker findings. Our study was designed to address whether biomarkers can be used in order inform decisions whether to proceed with an mpMRI assuming discriminating between PI-RADS 1,2 vs 3-5 was clinically relevant. Additional follow-up and a prospective trial design will be necessary to provide insight about optimal implementation of biomarkers and mpMRI in order to maximize the sensitivity and specificity of prostate biopsy for detecting csPCa. 

Funding: N/A