CUOS 2019: MRI: Where Do We Stand in 2019?

Toronto, Ontario ( The current Cancer Care Ontario (CCO) guidelines in Canada does not recommend performing a multiparametric MRI (mpMRI) in patients with an elevated risk of clinically significant prostate cancer (according to PSA levels) who are biopsy naïve. In contrast, in patients who have had a prior transrectal-ultrasound (TRUS) guided systematic biopsy and demonstrated a growing risk of having clinically significant prostate cancer – mpMRI should be performed, and followed by a targeted prostate biopsy. In this scenario, mpMRI may be considered to help in detecting more clinically significant prostate cancer patients compared to repeated TRUS-guided systematic biopsy.

In the setting of active surveillance of low-grade prostate cancer, there is some similarity in the various guidelines regarding the recommendation of performing a mpMRI. CCO recommends performing mpMRI when a patient’s clinical findings are discordant with the pathologic findings. The national comprehensive cancer network (NCCN) recommends considering mpMRI if anterior and/or aggressive cancer is expected (when PSA is increased and systematic biopsies are negative). Lastly, the National institution for health and care excellence (NICE) guidelines recommend a mpMRI at enrollment to active surveillance.

In the setting of pre-treatment local staging of prostate cancer, the CCO states that it is reasonable to perform a mpMRI to assess extra-prostatic extension in intermediate and high-risk patients considered for radical therapy (if knowledge of extraprostatic extension will alter management). Centers following this recommendation must have a quality assurance program to measure the diagnostic performance of mpMRI.

The indication for mpMRI use according to the CCO include:

  1. High PSA with a prior negative systematic biopsy
  2. Active surveillance – in changing the risk profile
  3. Staging – intermediate and high risk if it will alter management
  4. Reassessment after local therapy
  5. Local recurrence - only if it will alter management
  6. High PSA with no previous biopsies – not recommended!
Figure 1 demonstrates the results of the latest European multicenter randomized controlled trials in biopsy naïve men, including the Precision1, 4M2, and MRI-FIRST3 trials. These trials show that usage of mpMRI before the biopsy, result in the reduction of the number of patients being biopsied, in the percentage of insignificant cancer diagnosed, and in the number of biopsy cores taken. In contrast, it leads to an increased rate of detection of patients requiring active treatment and results in greater precision in index tumor grade/volume determination.

Figure 1 - Latest European Multicenter Randomized Controlled Trials in Biopsy Naïve Men
UroToday CUOS19 trials in biopsy naive menUroToday CUOS19 trials in biopsy naive men

In the current paradigm the risk stratification is heavily based on human interpretation and performance:

  1. The radiologist who provided MRI interpretation
  2. The urologist/radiologist who performed the biopsy
  3. The pathologist – who is responsible for the Gleason score and volume
Since this could lead to a large potential for error, the future should lead to combining fluidics/genomics with imaging biomarkers (radiomics) to improve accuracy. An example is the Stockholm 3 model, which combines six plasma protein biomarkers, genetic polymorphisms and clinical variables that estimates the risk of Gleason score>=7.4  Furthermore, incorporating machine learning techniques to provide more specific and sensitive results, resulting in a reduced potential of human error, has already been performed by Google5, is part of the future.

Dr. Haier concluded his talk stating that there is also work and upgrades being currently performed to the PIRADS system used for diagnosis (version 2.1 and 3). This upgrade will include quality assurance, report templates, certification, standards for significant interval change in imaging, brain uptake of Gadolinium, biopsy qualification, and biparametric MRI. These anticipated changes and upgrades should, in theory, improve the diagnostic accuracy of mpMRI and make it less prone to errors.

Presented by: Masoom Haider, BAM, MD, FRCPC, Professor, Mi Research Lead - Radiomics/Quantitative Imaging, Toronto General Hospital, Dept of Medical Imaging, University of Toronto, Ontario

Written by: Hanan Goldberg, MD, Urologic Oncology Fellow (SUO), University of Toronto, Princess Margaret Cancer Centre, Twitter: @GoldbergHanan at the CUOS – Canadian Uro-Oncology Summit 2019, #CUOS19 January 10-12, 2019 Westin Harbour Castle, Toronto, Ontario, Canada

1. Kasivisvanathan V. et al. NEJM 2018
2. van der Leest M. et al. European Urology 2018
3. Rouvière O. et al. Lancet Oncology 2019
4. Gronberg H et al. European Urology 2018
5. Stumpe, M (2018, November, 16).  Improved Grading of Prostate Cancer Using Deep Learning [Blog post] Retrieved from