Abstract: PD07-01
Sources of Funding: none

Introduction

A recent study confirmed the 4Kscore accurately predict aggressive prostate cancer. We assessed whether the performance of the 4Kscore Test differed by age or race. _x000D_ _x000D_

Methods

1312 men referred for prostate biopsy made up the cohort for these analyses. Differential calibration by age was assessed using logistic regression with an interaction between age and the 4Kscore Test for the outcome of a Gleason 3+4 cancer. We further categorized patients by age into subgroups of less than 55, 56-69, and 70 or older. The AUC for the 4Kscore Test was calculated in each age subgroup and confidence intervals for the difference in AUCs between groups were estimated. Similar analyses were performed to assess differential calibration and discrimination of the 4Kscore Test in African Americans versus the rest of the cohort. Finally, Decision curve analysis was used to assess the clinical utility of the 4Kscore Test for predicting aggressive prostate cancer within these sub groups of age and race.

Results

Among the cohort, 291 (22%) men were found to have high-grade cancer on prostate biopsy. There was no evidence to suggest a difference in discrimination of high-grade prostate cancer by either age or race, with the difference in confidence intervals surrounding the AUC’s overlapping zero. We found evidence of a difference in calibration by age when age was modeled as a continuous score (p=0.045). However, on sensitivity analysis looking at age categorized, we found no evidence of differential miscalibration (p=0.15). We found evidence of a difference in calibration by race (p=0.02), with scores slightly under predicting Gleason 7 cancer among African American men. Decision curve analysis found a higher net benefit for using the 4Kscore Test to decide on the need for biopsy in all of age and race subgroups.

Conclusions

We found no evidence of differential discrimination of Gleason 7 prostate cancer by age or race. There is some evidence to suggest miscalibration by age and race but this requires further assessment. Decision curve analysis suggests there is a clinical benefit to using the 4Kscore Test to decide on the need for a prostate biopsy in men of all ages and race.

Funding

none