Utilization of multiparametric prostate magnetic resonance imaging in clinical practice and focal therapy: report from a Delphi consensus project

Recent technological advancements in multiparametric magnetic resonance imaging (mpMRI) have resulted in improved detection of clinically significant prostate cancer (PCa) and are increasingly used in urological practice and for focal therapy (FT). MpMRI most commonly includes T1-2-weighted imaging, dynamic contrast-enhanced (DCE) and diffusion-weighted imaging (DWI), providing clinicians with meaningful information regarding lesion volume, morphology, location and disease extent. Dynamic three-dimensional lesion characterization and risk assessment with mpMRI is key for adequate patient selection and treatment planning for FT. Clinical guidelines for standardized reporting and acquisition (e.g., PI-RADS v2 [1] or Likert scale) of prostate mpMRI are nowadays advised for both research and clinical practice. When comparing mpMRI and pathology following radical prostatectomy (sliced by use of a customized 3D mold), the positive predictive value (PPV) for the detection of PCa in the peripheral zone, central zone and overall prostate was 98, 100 and 98 %, respectively, whereas the negative predictive value (NPV) was 90 % for all mpMRI sequences [2]. In another series, the positive and negative predictive values were 86 and 85 % for lesions >0.2 mL and 77 and 95 % for lesions >0.5 mL [3]. Noteworthy, most excellent results on mpMRI PCa detection are published by expert centers where the quality of the mpMRI is assured by standardized acquisition, interpretation and image-pathology feedback. Hence, results may not be reflective to general urological practice (for an overview table on mpMRI PCa detection see [4, 5]). To illustrate, a recent systematic review of available literature on the detection of significant PCa by mpMRI showed that the NPV ranged from 63 to 98 % [6]; however, the majority of the included studies used prostate biopsies for histopathological validation.

Of interest for FT of the index lesion, the PPV of mpMRI was reported to be 82.6 % [7], whereas 80 % of all index tumors and 72 % of Gleason ≥7 tumors on whole mount pathology were identified by mpMRI [8]. Moreover, mpMRI-transrectal ultrasound (TRUS) fusion (cognitive and system-based) targeted biopsies (TB) has shown to decrease the detection of clinically insignificant PCa and increase de detection rate of clinically significant PCa with an absolute difference of 6.8 % between mpMRI-TRUS fusion TB versus TRUS-guided biopsies [9]. However, not all studies reached a statistically significant difference [9‐15] (for systematic review see [9, 15]). Expert panels recommend to perform repeated mpMRI during surveillance following FT [4]. MpMRI-TRUS fusion TB has been shown to increase the detection of clinically significant PCa in the follow-up during active surveillance or during confirmatory biopsies of patients with previous negative TRUS-guided biopsies [16‐18]. With increasing evidence, the position of mpMRI and MR-targeted biopsies to the accepted standard of PCa diagnostics needs to be re-evaluated. Therefore, an international multidisciplinary consensus project was initiated using the Delphi method, aiming to define the use of mpMRI and MR-targeted biopsies in clinical practice and focal therapy of PCa.

This Delphi consensus project represents the opinion of 90 FT experts, experienced with mpMRI and MRI-TRUS fusion (TB), for all recommendations see Table 1. Despite experience, it should still be regarded as expert opinion, considered level 5 evidence, may be biased by personal enthusiasm and is potentially not reflective for community-based urologists. Nonetheless, these statements can provide clinicians guidance in areas where high-level evidence is sparse and provide a basis for standardization for clinical utilization of mpMRI and/or focal therapy that could result in improved interpretation of reported series.

The main challenge that was repeatedly encountered was the discrepancy regarding the consensus statements that can be made for expert centers versus non-expert centers. For experienced centers, where high-quality mpMRI is obtained by standardized reporting and acquisition, experience by radiologists, and where the own data are known, some excellent results have been published, establishing the mpMRI as a reliable diagnostic tool [2, 3, 6]. In community practice, this process may be less optimal, causing heterogeneous mpMRI quality with decreased sensitivity and specificity, resulting in clinically significant PCa being missed, understaging and a high number needed to image impairing the potential cost-effectiveness [20]. The panel argued that it should be our goal to guarantee high-quality mpMRI throughout the urological community before implementing it as standard of care. Important elements to achieve this are expert training, imaging-pathology feedback and/or certification. Moreover, cost-benefit studies must be performed per health care system and the availability of MRI scanners and logistics should improve [28]. Furthermore, it may be inevitable that FT can only be performed in centers where the mpMRI quality is guaranteed, since mpMRI and MRI-TRUS fusion biopsies should be performed in both the planning and follow-up of FT.

This consensus project recommended that mpMRI should be performed in patients with prior negative biopsies if clinical suspicion for PCa remains, which is stated more firmly than the EAU guideline recommendation that mpMRI may be used to evaluate the need to perform repeat biopsies [28]. Ahmed et al. presented during the 9th International Symposium on Focal Therapy and Imaging in Prostate and Kidney Cancer the first level 1 evidence that using mpMRI as a triage test, in all patients suspected of PCa (n = 576), can avoid unnecessary prostate biopsies and detected more clinically significant PCa (≥Gleason 4 + 3) than systematic (random) TRUS-guided biopsies. Based on their results, these authors calculated that in the United Kingdom it was cost-effective to perform mpMRI in all patients suspected of PCa (unpublished data, [20]). This underlines the panel statement that in expert centers mpMRI may be performed in all patients suspected of PCa. The outcomes regarding mpMRI acquisition and reporting are following the recommendations by the PI-RADS Steering Committee [1] that shows the potential that the PI-RADS recommendations have to effectuate uniform mpMRI.

Previous consensus meetings differed on the need for systematic biopsies when MRI-TRUS fusion TB were performed in biopsy-naïve patients [29, 30]. Since mpMRI quality is not constant throughout the field, abandoning systematic biopsies after MRI-TRUS fusion TB for FT planning in biopsy-naïve patients is not recommended by this consensus group. In line, systematic prostate biopsies should be taken during the follow-up after FT.

In expert centers mpMRI evaluation of the prostate is an established component in PCa diagnostics and this consensus project aimed to define the use of mpMRI in clinical care, especially in relation to FT. In FT, mpMRI should be used for patient selection, treatment planning or guidance, and post-treatment monitoring. Newly developed MRI-TRUS (system or cognitive) fusion is increasingly being used either during prostate biopsy procedures or during treatment planning or guidance. This consensus project was conducted during the turning point whether or not mpMRI of the prostate should be standard of care in all patients with or suspected of PCa. Standardization of acquisition and reading should be the main priority to guarantee consistent mpMRI quality throughout the field to substantiate the discrepancy between what you want to recommend and what you actually can recommend.