Detection of High-grade Prostate Cancer Using a Urinary Molecular Biomarker-Based Risk Score

Abstract

Background: To reduce overdiagnosis and overtreatment, a test is urgently needed to detect clinically significant prostate cancer (PCa).

Objective: To develop a multimodal model, incorporating previously identified messenger RNA (mRNA) biomarkers and traditional risk factors that could be used to identify patients with high-grade PCa (Gleason score ≥7) on prostate biopsy.

Design, setting, and participants: In two prospective multicenter studies, urine was collected for mRNA profiling after digital rectal examination (DRE) and prior to prostate biopsy. The multimodal risk score was developed on a first cohort (n=519) and subsequently validated clinically in an independent cohort (n=386).

Outcome measurements and statistical analysis: The mRNA levels were measured using reverse transcription quantitative polymerase chain reaction. Logistic regression was used to model patient risk and combine risk factors. Models were compared using the area under the curve (AUC) of the receiver operating characteristic, and clinical utility was evaluated with a decision curve analysis (DCA).

Results and limitations: HOXC6 and DLX1 mRNA levels were shown to be good predictors for the detection of high-grade PCa. The multimodal approach reached an overall AUC of 0.90 (95% confidence interval [CI], 0.85-0.95) in the validation cohort (AUC 0.86 in the training cohort), with the mRNA signature, prostate-specific antigen (PSA) density, and previous cancer-negative prostate biopsies as the strongest, most significant components, in addition to nonsignificant model contributions of PSA, age, and family history. For another model, which included DRE as an additional risk factor, an AUC of 0.86 (95% CI, 0.80-0.92) was obtained (AUC 0.90 in the training cohort). Both models were successfully validated, with no significant change in AUC in the validation cohort, and DCA indicated a strong net benefit and the best reduction in unnecessary biopsies compared with other clinical decision-making tools, such as the Prostate Cancer Prevention Trial risk calculator and the PCA3 assay.

Conclusions: The risk score based on the mRNA liquid biopsy assay combined with traditional clinical risk factors identified men at risk of harboring high-grade PCa and resulted in a better patient risk stratification compared with current methods in clinical practice. Therefore, the risk score could reduce the number of unnecessary prostate biopsies.

Patient summary: This study evaluated a novel urine-based assay that could be used as a noninvasive diagnostic aid for high-grade prostate cancer (PCa). When results of this assay are combined with traditional clinical risk factors, risk stratification for high-grade PCa and biopsy decision making are improved.

Keywords: Biologic markers; Logistic models; Messenger RNA; PSA; Prostate neoplasms; Urine.