Regional Histopathology and Prostate MRI Positivity: A Secondary Analysis of the PROMIS Trial

Abstract

Background The effects of regional histopathologic changes on prostate MRI scans have not been accurately quantified in men with an elevated prostate-specific antigen (PSA) level and no previous biopsy. Purpose To assess how Gleason grade, maximum cancer core length (MCCL), inflammation, prostatic intraepithelial neoplasia (PIN), or atypical small acinar proliferation within a Barzell zone affects the odds of MRI visibility. Materials and Methods In this secondary analysis of the Prostate MRI Imaging Study (PROMIS; May 2012 to November 2015), consecutive participants who underwent multiparametric MRI followed by a combined biopsy, including 5-mm transperineal mapping (TPM), were evaluated. TPM pathologic findings were reported at the whole-prostate level and for each of 20 Barzell zones per prostate. An expert panel blinded to the pathologic findings reviewed MRI scans and declared which Barzell areas spanned Likert score 3-5 lesions. The relationship of Gleason grade and MCCL to zonal MRI outcome (visible vs nonvisible) was assessed using generalized linear mixed-effects models with random intercepts for individual participants. Inflammation, PIN, and atypical small acinar proliferation were similarly assessed in men who had negative TPM results. Results Overall, 161 men (median age, 62 years [IQR, 11 years]) were evaluated and 3179 Barzell zones were assigned MRI status. Compared with benign areas, the odds of MRI visibility were higher when a zone contained cancer with a Gleason score of 3+4 (odds ratio [OR], 3.1; 95% CI: 1.9, 4.9; P < .001) or Gleason score greater than or equal to 4+3 (OR, 8.7; 95% CI: 4.5, 17.0; P < .001). MCCL also determined visibility (OR, 1.24 per millimeter increase; 95% CI: 1.15, 1.33; P < .001), but odds were lower with each prostate volume doubling (OR, 0.7; 95% CI: 0.5, 0.9). In men who were TPM-negative, the presence of PIN increased the odds of zonal visibility (OR, 3.7; 95% CI: 1.5, 9.1; P = .004). Conclusion An incremental relationship between cancer burden and prostate MRI visibility was observed. Prostatic intraepithelial neoplasia contributed to false-positive MRI findings. ClinicalTrials.gov registration no. NCT01292291 © RSNA, 2022 Supplemental material is available for this article. See also the editorial by Harmath in this issue.

Figure 1

(A) Flowchart shows participant population and MRI-transperineal mapping (TPM) alignment. The study included 161 nonpilot Prostate MRI Imaging Study (PROMIS) participants from University College London Hospitals (UCLH) who underwent multiparametric MRI followed by a combined biopsy procedure and detailed per-zone recording of Gleason grade and maximum cancer core length (MCCL) per the International Society of Urological Pathology (ISUP) definition. A consensus multidisciplinary panel, blinded to TPM findings, reviewed MRI scans and aligned any lesions with a Likert score greater than or equal to 3 to specific Barzell zones before the pathologic status was revealed. For example, the left peripheral zone lesion shown on the right was aligned to zones 13,14,17, and 18. (B) Spineplots show the percentage of zones deemed visible at MRI according to the overall Gleason score at TPM. The consensus panel determined the MRI positivity of 3179 zones; this was slightly less than the expected 3220 (161 participants x 20 zones) due to small prostate size in five men, which prevented full sampling of all Barzell zones. Of 3179 zones, 2516 were benign and, of those with cancer, 301 had a Gleason score of 3+3, 271 had a Gleason score of 3+4, and 91 had a Gleason score greater than or equal to 4+3. In total, 595 zones were MRI-positive (18.7%), although the proportion that were MRI-visible rose with increasing Gleason grade and with each additional millimeter of MCCL, motivating a zonal pathologic finding-based model of MRI positivity. ADC = apparent diffusion coefficient, b1400 = b value of 1400, csCa = clinically significant cancer, DCE = dynamic contrast-enhanced, T2WI = T2-weighted imaging.

Figure 2

(A) Boxplot shows a moderate age difference between Gleason score groups (P = .04, Kruskal-Wallis analysis of variance) that was primarily driven by the lower median age of the Gleason 3+3 group (n = 21). (B) Boxplot shows men with an overall Gleason score greater than or equal to 4+3 at transperineal mapping (TPM) had low prostate volumes compared with other groups (P < .001, Kruskal-Wallis analysis of variance and adjusted pairwise comparisons). (C) Boxplot shows men with an overall Gleason score greater than or equal to 4+3 at TPM had high prostate-specific antigen (PSA) density (PSAD) (P < .001, Kruskal-Wallis analysis of variance and adjusted pairwise comparisons) and low prostate volume (as shown in B) compared with other groups. In men who were TPM-negative, this relationship was reversed; prostate volume was highest and PSA density lowest. These findings imply the existence of two distinct pathologic states in biopsy-naive men that, although both manifest as an elevated PSA level indicating the need for biopsy, differ in terms of the mechanism generating the increase in PSA.

Figure 3

(A) Graphs show the predicted probabilities of a Barzell zone being visible at MRI when using the zonal Gleason grade, maximum cancer core length (MCCL), and binary logarithm (log₂) of prostate volume (in milliliters) as predictors in a mixed model with random intercepts for participants (3179 zones in 161 men). The probability progressively increases with every increment in Gleason grade or MCCL, while prostate volume increase has the opposite effect. (B) Horizontal boxplots show the relationship between Likert scores and cancer burden. After selecting the Barzell zones spanning a specific lesion, the expert panel also agreed on a single zone having the “best” alignment. In total, 115 cancerous zones best-aligned with a lesion (index or secondary), and their pathologic classification against Likert scoring is shown. Greater MCCL was associated with higher radiologic scores (Likert score 4-5), particularly when Gleason pattern 4 was present (P < .05, Kruskal-Wallis). An MCCL less than approximately 5 mm was mostly associated with Likert 3 lesions (the main arena of true- and false-positive MRI distinction), regardless of cancer grade. ISUP = International Society of Urological Pathology.