MIAMI—While the 4Kscore test has been shown to accurately detects aggressive prostate cancer and reduces unnecessary biopsies overall, its performance in African-American (AA) men remained unclear. To remedy that, a study led by researchers from the Miami VAMC and the University of Miami assessed the performance of the 4Kscore test in a cohort of men with a majority of African-Americans.
According to the study published in the Journal of Urology, men referred for a biopsy of the prostate at eight VAMCs were prospectively enrolled. All men underwent phlebotomy for 4Kscore test assessment prior to biopsy of the prostate, with the primary outcome defined as the detection of Grade Group 2 (GG2) or higher cancer on biopsy. With 366 veterans enrolled, 205 (56%) were AA, and 131 (36%) had GG2 or higher prostate cancer.1
Researchers assessed the discrimination, calibration and clinical utility of the 4Kscore test for predicting GG2 or higher prostate cancer, comparing it to a base model consisting of age, digital rectal exam and PSA. At the same time, they compared the performance of the 4Kscore test in AA and non-AA men.
Results indicated that the 4Kscore test exhibited better discrimination (AUC: 0.81 vs. 0.74, p<0.01) and higher clinical utility on decision curve analysis than the base model, closely approximating the observed risk of GG2 or higher prostate cancer. No difference was documented in the performance of the 4Kscore test in AA and non-AA men (0.80 vs. 0.84; p=0.32), and the 4Kscore outperformed the base model in both groups.
“The 4Kscore test accurately predicts aggressive prostate cancer for biopsy decision making in both AA and non-AA men,” the study authors concluded.
1. Punnen S, Freedland SJ, Polascik TJ, Loeb S, Risk MC, Savage S, Mathur SC,Uchio E, Dong Y, Silberstein JL. A Multi-institutional Prospective Trial in theVeterans Affairs Health System confirms the 4Kscore maintains its Predictivevalue among African American Men. J Urol. 2017 Dec 6. pii: S0022-5347(17)78032-9.doi: 10.1016/j.juro.2017.11.113. [Epub ahead of print] PubMed PMID: 29223389.
