Prediction of High-grade Prostate Cancer Following Multiparametric Magnetic Resonance Imaging: Improving the Rotterdam European Randomized Study of Screening for Prostate Cancer Risk Calculators

Eur Urol. 2019 Feb;75(2):310-318. doi: 10.1016/j.eururo.2018.07.031. Epub 2018 Aug 3.

Abstract

Background: The Rotterdam European Randomized Study of Screening for Prostate Cancer risk calculators (ERSPC-RCs) help to avoid unnecessary transrectal ultrasound-guided systematic biopsies (TRUS-Bx). Multivariable risk stratification could also avoid unnecessary biopsies following multiparametric magnetic resonance imaging (mpMRI).

Objective: To construct MRI-ERSPC-RCs for the prediction of any- and high-grade (Gleason score ≥3 + 4) prostate cancer (PCa) in 12-core TRUS-Bx±MRI-targeted biopsy (MRI-TBx) by adding Prostate Imaging Reporting and Data System (PI-RADS) and age as parameters to the ERSPC-RC3 (biopsy-naïve men) and ERSPC-RC4 (previously biopsied men).

Design, setting, and participants: A total of 961 men received mpMRI and 12-core TRUS-Bx±MRI-TBx (in case of PI-RADS ≥3) in five institutions. Data of 504 biopsy-naïve and 457 previously biopsied men were used to adjust the ERSPC-RC3 and ERSPC-RC4.

Outcome measurements and statistical analysis: Logistic regression models were constructed. The areas under the curve (AUCs) of the original ERSPC-RCs and MRI-ERSPC-RCs (including PI-RADS and age) for any- and high-grade PCa were compared. Decision curve analysis was performed to assess the clinical utility of the MRI-ERSPC-RCs.

Results and limitations: MRI-ERSPC-RC3 had a significantly higher AUC for high-grade PCa compared with the ERSPC-RC3: 0.84 (95% confidence interval [CI] 0.81-0.88) versus 0.76 (95% CI 0.71-0.80, p<0.01). Similarly, MRI-ERSPC-RC4 had a higher AUC for high-grade PCa compared with the ERSPC-RC4: 0.85 (95% CI 0.81-0.89) versus 0.74 (95% CI 0.69-0.79, p<0.01). Unlike for the MRI-ERSPC-RC3, decision curve analysis showed clear net benefit of the MRI-ERSPC-RC4 at a high-grade PCa risk threshold of ≥5%. Using a ≥10% high-grade PCa risk threshold to biopsy for the MRI-ERSPC-RC4, 36% biopsies are saved, missing low- and high-grade PCa, respectively, in 15% and 4% of men who are not biopsied.

Conclusions: We adjusted the ERSPC-RCs for the prediction of any- and high-grade PCa in 12-core TRUS-Bx±MRI-TBx. Although the ability of the MRI-ERSPC-RC3 for biopsy-naïve men to avoid biopsies remains questionable, application of the MRI-ERSPC-RC4 in previously biopsied men in our cohort would have avoided 36% of biopsies, missing high-grade PCa in 4% of men who would not have received a biopsy.

Patient summary: We have constructed magnetic resonance imaging-based Rotterdam European Randomized study of Screening for Prostate Cancer (MRI-ERSPC) risk calculators for prostate cancer prediction in transrectal ultrasound-guided biopsy and MRI-targeted biopsy by incorporating age and Prostate Imaging Reporting and Data System score into the original ERSPC risk calculators. The MRI-ERSPC risk calculator for previously biopsied men could be used to avoid one-third of biopsies following MRI.

Keywords: Biopsy; Magnetic resonance imaging; Multivariable risk stratification; Prostate cancer; Risk calculator.

Publication types

  • Multicenter Study

MeSH terms

  • Aged
  • Biopsy, Large-Core Needle
  • Databases, Factual
  • Diffusion Magnetic Resonance Imaging*
  • Europe
  • Humans
  • Image-Guided Biopsy / methods*
  • Kallikreins / blood
  • Magnetic Resonance Imaging, Interventional*
  • Male
  • Middle Aged
  • Neoplasm Grading
  • Predictive Value of Tests
  • Prostate-Specific Antigen / blood
  • Prostatic Neoplasms / blood
  • Prostatic Neoplasms / diagnostic imaging*
  • Prostatic Neoplasms / pathology
  • Reproducibility of Results
  • Risk Assessment
  • Risk Factors
  • Ultrasonography, Interventional
  • Unnecessary Procedures

Substances

  • KLK3 protein, human
  • Kallikreins
  • Prostate-Specific Antigen