Introduction:

Previous biomarker tests have failed to accurately predict upgrading in patients on active surveillance (AS) for low-risk prostate cancer. The 18-gene MyProstateScore 2.0 (MPS2) assay includes markers specifically overexpressed in high-grade relative to low-grade cancers, providing biologic rationale for potential effectiveness in AS. Using the MPS2 urinary testing panel, we developed and validated a novel testing model (MPS2-AS) to predict upgrading to GG≥3 PCa in patients with GG1 disease undergoing AS.

Methods:

A novel MPS2-AS testing model was established using the 18 MPS2 genes in a University of Michigan development cohort (2008-2020). The model was validated in a multi-center cohort of patients undergoing AS for GG1 disease. All patients underwent multiparametric magnetic resonance imaging (MRI) followed by systematic 12-core transrectal or transperineal biopsy, and patients with PI-RADS ≥3 lesions underwent additional targeted biopsies. First-catch (non-DRE) urine was collected at the time of surveillance biopsy for MPS2-AS testing. Because some patients with GG2 remain candidates for AS, the primary study outcome was significant upgrading to GG≥3 PCa on AS biopsy. Test performance measures and the clinical consequences (upgrading events detected, unnecessary biopsies avoided) of using MPS2-AS, MRI, or the PCPT risk calculator to determine the need for biopsy were calculated and compared.

Results:

The external validation cohort included 223 patients undergoing AS at one of nine urology clinics (3 academic, 6 non-academic) for an existing diagnosis of GG1 PCa. Median patient age was 68 years (IQR 63-73), median PSA was 7.2 ng/ml (4.4-9.2), and 142 (64%) patients had a positive pre-biopsy MRI (PI-RADS≥3). Overall, 84 patients (38%) had GG≥2 cancer and 20 (9.0%) had GG≥3 cancer on AS biopsy. Considering detection of GG≥3 as a primary aim of surveillance biopsies, use of MPS2-AS to determine the need for biopsy would have avoided 70% of unnecessary AS biopsies while detecting 95% of GG≥3 cancers. By contrast, use of PI-RADS ≥3 to determine the need for biopsy would have avoided 48% of unnecessary AS biopsies while capturing only 79% of GG≥3 cancers. Use of the PCPT risk calculator would have avoided 16% of unnecessary biopsies while capturing 95% of GG≥3 cancers.

Conclusion:

In this ongoing, prospective, multi-center study of patients undergoing AS for GG1 PCa, we found that use of a novel non-DRE urine test (MPS2-AS) to determine the need for AS biopsy appears to provide favorable clinical outcomes compared to MRI and PSA-based testing. Importantly, MPS2-AS provided actionable rule-out testing that would allow patients to forego more than two-thirds of unnecessary AS biopsies. While further validation is needed, these findings suggest MPS2-AS could be used to guide a non-invasive approach to monitoring, reduce the need for scheduled surveillance biopsies, and preserve the use of MRI for optimizing the yield of biopsies in higher-risk patients.

Funding: N/a