Intended for healthcare professionals


Future of screening for prostate cancer

BMJ 2017; 358 doi: (Published 19 September 2017) Cite this as: BMJ 2017;358:j4200
  1. Philipp Dahm, professor of surgery
  1. Urology Section, Minneapolis VA Medical Center, University of Minnesota, Minneapolis, MN, USA
  1. pdahm{at}

Multiparametric magnetic resonance imaging is likely to be part of it

Few topics in medicine stir up as much controversy as prostate cancer screening. Prostate cancer is a highly prevalent, potentially lethal disease, making the enthusiasm of the public and health professionals for screening and treatment understandable.1 However, many screen detected cancers have a protracted and indolent natural course with no adverse effects for decades, and patients are at greater risk from the ensuing cascade of diagnostic imaging and unnecessary treatments than from the disease itself.

Current best evidence suggests that prostate specific antigen (PSA) testing provides only a small reduction in prostate cancer mortality and no reduction in all cause mortality, while at the same time exposing healthy individuals to the risk of overdiagnosis and overtreatment.2 Recent draft recommendations by the United States Preventive Services Task Force against population based screening (grade C), emphasise the importance of shared decision making between men and their healthcare providers but fall short of providing other actionable guidance.3 Although the results of the CAP trial, a cluster randomised controlled trial of PSA screening in primary care clinics in the UK, will provide further evidence, it is unlikely to settle the controversy.4

Better tests

Meanwhile, developments are under way that may allow us to improve diagnostic accuracy and potentially lessen the burden of testing and treatment associated with prostate cancer screening. Tools such as the Prostate Health Index and the 4Kscore, which calculate risk from measurements of multiple biomarkers, may enable better risk stratification for men who choose screening, although no evidence based guidelines currently endorse their routine use.

Perhaps more promising in the short term is the development of multiparametric magnetic resonance imaging (mpMRI), which combines morphological assessment of the prostate with elements of functional and physiological assessment. There has also been a concerted, multiorganisational effort to standardise the acquisition, interpretation, and reporting of mpMRI in the form of the Prostate Imaging Reporting and Data System (PIRADS).5

The most notable study so far to place mpMRI on the map as a potential tool for identifying men with clinically important prostate cancer is PROMIS.6 This prospective cohort study from the UK enrolled 740 screened men with PSA concentrations up to 15 ng/mL and scheduled for their first prostate biopsy. The authors compared the diagnostic accuracy of mpMRI with traditional transrectal ultrasound (TRUS) guided biopsy, by testing them both against a reference standard test—template prostate mapping (TPM) biopsy, which samples tissue every 5 mm through the prostate.

A total of 576 patients completed all three tests. The primary outcome was clinically significant cancer defined as a Gleason score ≥4+3 or a maximum cancer length in any biopsy core ≥6 mm.

Based on the reference standard biopsy, 230 men (40%) had clinically significant prostate cancer and an additional 178 men (31%) had clinically insignificant cancers. The sensitivity and specificity of mpMRI for identifying the primary outcome were 93% (95% confidence interval 88% to 96%) and 41% (36% to 46%), respectively, compared with 48% (42% to 55%) and 96% (94% to 98%), respectively, for TRUS biopsy.

Using the prevalence found in the study, the 27% of men with a negative mpMRI result would be able to avoid a biopsy, accepting a 11% probability of a false negative MRI result. Patients with suspicious MRI findings would have a roughly 50% chance of having clinically significant prostate cancer.

Reducing the need for prostate biopsy is a major step forward that will be welcomed by all men, and especially those who have had the procedure. One source of bias in this study relates to the relatively large proportion of enrolled patients (22%) who did not complete the study, most commonly because their prostate was too large for a TPM biopsy (>100 cm3). Secondly, interpretation of mpMRI is known to be subject to major interobserver variability.7 Unless rigorous training and quality assurance is implemented, centres adopting mpMRI are unlikely to match the diagnostic accuracy reported by these experienced investigators. Lastly, mpMRI was not used to identify men for TRUS biopsy in this study. The results therefore provide only indirect evidence about how MRI imaging would perform when used as a triage test in the diagnostic pathway.

The study authors should therefore be congratulated for taking the critical next step and conducting a randomised controlled trial of mpMRI triage. The PRECISION (Prostate Evaluation for Clinically Important Disease: Sampling Use Image-Guidance or Not) trial is in the process of randomising 470 men to mpMRI triage versus standard 12 core TRUS biopsy.8 In the former group, a prostate biopsy is performed only after a suspicious mpMRI, which is then used to target the biopsy. The primary endpoint is once again clinically significant prostate cancer.

Men at risk of prostate cancer, their families, and the health professionals who care for them should all be encouraged by the development of mpMRA and the rigorous evidence based approach being taken to evaluate new screening strategies. While not solving all current problems, both hold much promise for the future of prostate cancer screening.


  • Competing interests: I have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract