Utility of noninvasive biomarker testing and MRI to predict a prostate cancer diagnosis

Abstract

Abstract Purpose To assess the diagnostic performance and utility of the ExoDx IntelliScore and an OPKO4K score to predict prostate cancer in men presenting with elevated PSA—both as independent predictors and in combination with clinical/MRI characteristics. Methods Patients with elevated PSA were retrospectively reviewed. Abnormal tests were defined as an OPKO4K score ≥ 7.5% and an ExoDx IntelliScore ≥ 15.6. Four regression models and ROC curves were generated based on: (1) age, PSA, and DRE, (2) model 1 + OPKO4K 4Kscore ≥ 7.5%, (3) model 2 + ExoDx IntelliScore ≥ 15.6, and (4) model 3 + MRI PIRADS 4–5. Results 359 men received an OPKO4K test, 307 had MRI and 113 had ExoDx tests. 163 men proceeded to prostate biopsy and 196 (55%) were saved from biopsy. Mean age was 65.0 ± 8.7 years and mean PSA was 7.1 ± 6.1 ng/mL. Positive biopsies were found in 84 (51.5%) men. The sensitivity and negative predictive value of an OPKO4K score were 86.7% and 72.3%; values for an ExoDx test were 76.5% and 77.1%, respectively. On regression analysis, clinical markers (Age, PSA, DRE) generated an AUC of 0.559. The addition of an OPKO4K score raised the AUC to 0.653. The stepwise addition of an ExoDx score raised the AUC to 0.766. The combined use of both biomarkers, patient characteristics, and MRI yielded an AUC of 0.825. Conclusion This analysis demonstrates the high negative predictive value of both the OPKO4K score and ExoDX IntelliScore independently while demonstrating that the combination of an OPKO4K score, an ExoDX IntelliScore, and MRI increases predictive capability for biopsy confirmed prostate cancer.