Introduction

The 4Kscore test was previously validated in a large, prospective trial to predict aggressive prostate cancer, however, the study population had a limited number of African American (AA) men. We conducted an independent multi-institutional, prospective trial to validate the 4Kscore test within the Veterans Affairs (VA) Health System, where a large proportion of the men getting care are AA.

Methods

We prospectively enrolled men who were referred for biopsy of their prostate at 8 diverse VA sites throughout the nation. All men underwent phlebotomy for 4Kscore ascertainment prior to prostate biopsy. We assessed the discrimination, calibration, and clinical utility of the 4Kscore test for predicting Gleason 7 or higher (G7+) prostate cancer, and compared it to a base model consisting of age, digital rectal exam findings, and PSA. Additionally, we compared the performance of the 4Kscore test in AA and non-AA men.

Results

Among 403 men who were enrolled in the trial, we had 366 men with a 4Kscore and complete data available for analysis. Among these men, 208 (56%) were AA, and 134 (36%) had G7+ prostate cancer. The 4Kscore exhibited better discrimination (AUC: 0.81 vs. 0.74, p=0.011) and higher clinical utility on decision analysis than the base model for deciding on the need for biopsy. Calibration plots of the 4Kscore for the entire cohort afforded predictions that closely matched the observed risk of G7+ prostate cancer in the population (Figure 1). There was no difference in the discrimination of the 4Kscore test between AA and non-AA men (0.80 vs. 0.84; p=0.32). While we found some evidence that the 4Kscore underestimates the risk of G7+ prostate cancer in AA men, discrimination (0.80 vs. 0.72, p = 0.013) and clinical utility for the 4Kscore test were still higher than the base model.

Conclusions

In an independent, multi-institutional, prospective trial of the 4Kscore test in the VA health system, we confirmed that the 4Kscore accurately predicts the likelihood of aggressive prostate cancer and outperforms standard clinical information for biopsy decision making in both AA and non-AA men.

Funding

OPKO Diagnostics