Introduction:

Active surveillance (AS) is the preferred management method for men with very low- and low-risk prostate cancers. Widespread adoption of AS has reduced over-treatment.  Currently, surveillance biopsies at regular intervals are the predominant method to identify upgrading, i.e. increase in biopsy Gleason Grade Group (GG).  However, reluctance to undergo repeated biopsies is one of the primary reasons men leave AS for definitive treatment in the absence of upgrading.  This problem has led to efforts to identify non-invasive biomarkers for detection of clinically significant prostate cancer (csPCa).  SelectMDx is a urine-based biomarker classifier combining urinary HOXC6 and DLX1 mRNA  levels, patient age, digital rectal examination results and prostate specific antigen density (PSAD) in a locked algorithm; values range from -5 to 5.  SelectMDx has been shown to identify csPCa in men prior to an initial biopsy.  We evaluated SelectMDx performance to monitor men in AS for upgrading.

Methods:

We identified Johns Hopkins AS patients diagnosed 2000-2016 with GG 1 at enrollment and very-low or low-risk PCa; upgrading is defined as a biopsy showing GG 2-5.  We retrospectively identified men in 3 groups:  Group 1:  men upgraded at their first surveillance biopsy (n=86); Group 2:  men upgraded during years 2-6 in AS (n=49); Group 3:  men who were never upgraded after >6 years in AS (n=50).  Urine samples from the time of AS enrollment were sent to MDxHealth, Inc. for analysis and SelectMDx score calculation.  Data analysis was performed by Johns Hopkins using polytomous logistic regression to compare SelectMDx scores among the 3 groups, adjusting for prognostic variables.  The impact of adding SelectMDx to a model with clinical variables was evaluated with Akaike Information Criterion (AIC), and AUC adapted for more than binary outcome. Funding was provided by MDxHealth, Inc.

Results:

The SelectMDx score differed significantly among the groups, with highest score in Group 1 (most rapid upgrade), and lowest in Group 3 (not upgraded).  The baseline clinical regression model included PSAD, time from diagnosis to AS enrollment, number of positive cores at enrollment, and year of urine collection.  Addition of SelectMDx score to the model significantly improved model fit, while PSA density became non-significant; AUC increased from 0.868 for clinical variables alone vs. 0.875 with addition of SelectMDx.  A one unit change in SelectMDx was associated with odds ratio (OR)=4.7 for Group 1 vs. Group 3, and OR=2.9 for Group 2 vs. Group 3 (see Table).  Risk associated with SelectMDx was even stronger if analysis was restricted to the 117 men who met criteria for very-low risk:  Group 1 vs. Group 3, OR=7.3; Group 2 vs. Group 3, OR=4.6.

Conclusion:

This is only the 2^nd study and the largest study to date to evaluate SelectMDx in an AS cohort.  SelectMDx score was associated with a more than 4-fold increase in risk of rapid upgrading, even after adjusting for known prognostic factors, and the effect was particularly strong in men meeting the criteria for very low risk (GG=1, no more than 2 positive cores, <50% of any core involved with cancer, and PSA density<0.15).  Evaluation of this non-invasive test is warranted as a potential means of monitoring AS patients during follow-up, to reduce biopsy frequency and the need for expensive MRIs.

Funding: MDxHealth, Inc. grant funding