Heterogeneity in detection rates of higher grade prostate cancer by multiparametric MRI in an active surveillance cohort

View Poster


Multiparametric magnetic resonance imaging (mpMRI) has been shown to improve sensitivity for detection of higher grade prostate cancer (PCa). However, it&[prime]s utility in monitoring men on active surveillance (AS) is still unclear. We hypothesized that the utility of mpMRI would differ across risk strata in men appropriate for AS.


Between 2014 and 2017, we retrospectively identified 449 men with grade group (GG) 1 cancer (median AS follow-up 3 yrs., IQR 2 - 6 yrs.) from the Johns Hopkins AS registry with a mpMRI showing lesion(s) scored in PIRADS v2.0 and a follow-up targeted and/or systematic biopsy within a year. The study cohort was stratified into 4 sub-groups based on biopsy results prior to mpMRI: a) very-low-risk (≤2 positive biopsy cores and ≤50% core involvement, n = 212), b) low-risk (>2 cores or >50% core involvement, n = 237); and based on number of prior prostate biopsies: c) ≤2 biopsies (n = 220), d) >5 biopsies (n = 129). Rates of upgrading to GG≥ 2 (Gleason score ≥ 3+4) on follow-up biopsy were compared across PIRADS scores between each respective risk-subgroup.


Within each of the 4 sub-groups there was a significant increase in detection rates of GG≥ 2 with increasing PIRADS v2.0 score (all p 5 biopsies (8.0% PIRADS ≤3, 33% PIRADS 4, 50% PIRADS 5 vs. 10% PIRADS ≤3, 13% PIRADS 4, 25% PIRADS 5; p = 0.03). Overall, PIRADS 4/5 had significantly higher detection rates of GG≥ 2 in men with low-risk PCa than with very-low-risk PCa (38% vs. 11%, p


These data suggest heterogeneity in higher-grade cancer detection rates by mpMRI, which is contingent upon patient characteristics of cancer volume (low-risk vs. very-low-risk) and extent of under-sampling (prior number of biopsies). Further evaluation is needed to identify appropriate sub-groups of men in AS where mpMRI may be unnecessary.

Funding: None