Tag Archive for: aotw 24-07-19

Posts

Article of the week: Biparametric vs multiparametric prostate MRI for the detection of PCa in treatment‐naïve patients

Every week, the Editor-in-Chief selects an Article of the Week from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.

In addition to the article itself, there is an editorial written by a prominent member of the urological community, and a video produced by the authors. These are intended to provoke comment and discussion and we invite you to use the comment tools at the bottom of each post to join the conversation. 

If you only have time to read one article this week, it should be this one.

Biparametric vs multiparametric prostate magnetic resonance imaging for the detection of prostate cancer in treatment-naïve patients: a diagnostic test accuracy systematic review and meta-analysis

Mostafa Alabousi*, Jean-Paul Salameh†‡, Kaela Gusenbauer§, Lucy Samoilov, Ali Jafri**, Hang Yu§ and Abdullah Alabousi††

 

*Department of Radiology, McMaster University, Hamilton, Department of Clinical Epidemiology and Public Health, University of Ottawa, The Ottawa Hospital Research Institute, Clinical Epidemiology Program, Ottawa, §Department of Medicine, McMaster University, Hamilton, Department of Medicine, Western University, London, ON, Canada, **Department of Medicine, New York Institute of Technology School of Osteopathic Medicine, Glen Head, NY, USA, and ††Department of Radiology, St Joseph’s Healthcare, McMaster University, Hamilton, ON, Canada

Abstract

Objective

To perform a diagnostic test accuracy (DTA) systematic review and meta‐analysis comparing multiparametric (diffusion‐weighted imaging [DWI], T2‐weighted imaging [T2WI], and dynamic contrast‐enhanced [DCE] imaging) magnetic resonance imaging (mpMRI) and biparametric (DWI and T2WI) MRI (bpMRI) in detecting prostate cancer in treatment‐naïve patients.

Methods

The Medical Literature Analysis and Retrieval System Online (MEDLINE) and Excerpta Medica dataBASE (EMBASE) were searched to identify relevant studies published after 1 January 2012. Articles underwent title, abstract, and full‐text screening. Inclusion criteria consisted of patients with suspected prostate cancer, bpMRI and/or mpMRI as the index test(s), histopathology as the reference standard, and a DTA outcome measure. Methodological and DTA data were extracted. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS)‐2 tool. DTA metrics were pooled using bivariate random‐effects meta‐analysis. Subgroup analysis was conducted to assess for heterogeneity.

Results

From an initial 3502 studies, 31 studies reporting on 9480 patients (4296 with prostate cancer) met the inclusion criteria for the meta‐analysis; 25 studies reported on mpMRI (7000 patients, 2954 with prostate cancer) and 12 studies reported on bpMRI DTA (2716 patients, 1477 with prostate cancer). Pooled summary statistics demonstrated no significant difference for sensitivity (mpMRI: 86%, 95% confidence interval [CI] 81–90; bpMRI: 90%, 95% CI 83–94) or specificity (mpMRI: 73%, 95% CI 64–81; bpMRI: 70%, 95% CI 42–83). The summary receiver operating characteristic curves were comparable for mpMRI (0.87) and bpMRI (0.90).

Conclusions

No significant difference in DTA was found between mpMRI and bpMRI in diagnosing prostate cancer in treatment‐naïve patients. Study heterogeneity warrants cautious interpretation of the results. With replication of our findings in dedicated validation studies, bpMRI may serve as a faster, cheaper, gadolinium‐free alternative to mpMRI.

 

Editorial: Dropping the GAD – just a fad?

It is not without irony that, at the very moment that the UK’s National Institute for Health and Care Excellence (NICE) is poised to ratify the recommendation that multiparametric MRI (mpMRI) be introduced into the prostate cancer diagnostic pathway, we are seeking to significantly modify the very intervention on which they are about to provide judgement on [1].

The modification proposed is both compelling and plausible, as it renders the process of imaging the prostate in order to detect and localise clinically significant prostate cancer; simpler, quicker, safer and cheaper. It entails dropping the most complex and time‐consuming component of the three multiparametric sequences, the dynamic (time‐dependent) T1‐weighted gadolinium‐enhanced (GAD) sequence. This was a sequence that was, in the early days of MRI, imbued to have biological significance because it was capable of exploiting the differences in the microvascular architecture and function that we have tended to associate with cancer and non‐cancer in order to discriminate between the two. Or so we thought [2].

The systematic review in this issue of the BJUI by Alabousi et al. [3] explores, via the process of systematic review, whether the omission of the T1‐GAD sequence results in any clinically important reduction in test performance when compared with the full sequence scan comprising traditionally of T2, diffusion and T1‐GAD sequences. It did not.

By any stretch this is a tough analysis to pull‐off, as T1‐GAD sequences are not standardised in terms of acquisition or reporting. Every group seems to manage the dynamic images in a different way. As such they tend to suffer from quality control issues, possibly to a greater extent than the T2 and diffusion sequences. The verification of the signal by biopsy strategy and sampling intensity will have varied across studies, as will the threshold of the definition of clinically significant prostate cancer. These inherent methodological problems are all familiar to readers and issues that are pertinent to any imaging study in the detection of prostate cancer. However, there are two issues that make any current assessment of GAD vs no GAD really problematic. The first is the almost exclusive reliance on single‐centre retrospective data. In the few studies that claim a prospective design no comparative data were available. Studies of this type are typical in the early phase of exploring a clinical question and will, in time, be corrected. The other, largely hidden, hardly discussed and truly problematic issue relates to the manner by which we synthesise an overall risk score from the MRI sequences that we derive. The near ubiquitous use of the Prostate Imaging‐Reporting and Data System (PI‐RADS) scoring system introduces a systematic bias by the manner in which a Boolean form of logic is used to decide on the degree of influence that each sequence has in relation to the overall score. According to the manner by which PI‐RADS is applied, it tends to render the T1‐GAD sequence subordinate (only relevant in a minority of cases), contingent (to T2/diffusion) and disparate (dependent on prostate zone) in the way it is invoked [4]. The result is, that within the PIRADS framework, the T1‐GAD sequence is destined to play a relatively small role in driving the overall summary score of risk. It might, therefore, not be too surprising if its removal made little difference to the overall detection of clinically significant prostate cancer.

So what are the next steps? Clearly this is a very important issue and a simpler, quicker, safer and cheaper MRI would be desirable from multiple perspectives. It would render what is currently a complex intervention that comprises an invasive component into a totally passive image acquisition in which no medically trained health professional need be present. It is almost certainly a pre‐requisite for adoption in resource-poor jurisdictions and for entertaining the role of MRI as a primary population‐based screening test.

It took a large number of randomised trials to get mpMRI accepted into the prostate cancer diagnostic pathway. What is the minimum amount of evidence required to disinvest in one of its key components? In other words how many clinically significant cancers would we tolerate missing in order to offer the less complex test?

A direct (head‐to‐head) non‐inferiority randomised comparative study would, following some of our own recent calculations, require >3000 men to participate, which might just prove a little too challenging. An alternative approach is a study in which men would have lesions declared using a Likert score, thereby making no prior assumptions on the role and utility of any single sequence, by traditional mpMRI (standard) but also by a T2‐diffusion MRI (experimental) with appropriate blinding. Some lesions would be private to either standard or experimental imaging but most, it is likely, would be shared. All would require sampling. The yield, the misses, the test accuracy for each approach, could be calculated with necessary adjustments for the inevitable incorporation and verification biases.

It is interesting to observe that in many parts of the world mpMRI was introduced by clinicians before a large body of evidence was accumulated because they felt it was the right thing to do [5]. It may well be the case that ‘dropping the GAD’ will be subject to the same decision‐making process and precede any definitive judgement based on reliable evidence. Recent activity on PubMed would suggest that this might already have happened [6].

References

  1. National Institute for Health and Care Excellence (NICE). Non‐invasive MRI scan for Prostate Cancer recommended by NICE. Available at: https://www.nice.org.uk/news/article/non-invasive-mri-scan-for-prostate-cancer-recommended-by-nice. Accessed May 2019.
  2. Little, RABarjat, HHare, JI et al. Evaluation of dynamic contrast‐enhanced MRI biomarkers for stratified cancer medicine: how do permeability and perfusion vary between human tumours? Magn Reson Imaging 20184698– 105
  3. Alabousi, MSalameh, JPGusenbauer, K et al. Biparametric vs multiparametric prostate magnetic resonance imaging for the detection of prostate cancer in treatment‐naïve patients: a diagnostic test accuracy systematic review and meta‐analysis. BJU Int 2019124209– 20
  4. Turkbey, BRosenkrantz, ABHaider, MA et al. Prostate Imaging reporting and Data System version 2.1: 2019 update of Prostate Imaging Reporting and Data System version 2. Eur Urol 2019 [Epub ahead of print]. https://doi.org/10.1016/j.eururo.2019.02.033
  5. Ahmed, HUKirkham, AArya, M et al. Is it time to consider a role for MRI before prostate biopsy? Nat Rev Clin Oncol 20096197– 20
  6. Xu, MFang, MZou, J et al. Using biparametric MRI radiomics signature to differentiate between benign and malignant prostate lesions. Eur J Radiol 201911438– 44

 

Video: Biparametric vs multiparametric prostate MRI for the detection of PCa in treatment‐naïve patients: a diagnostic test accuracy systematic review and meta‐analysis

Read the full article

Abstract

Objective

To perform a diagnostic test accuracy (DTA) systematic review and meta‐analysis comparing multiparametric (diffusion‐weighted imaging [DWI], T2‐weighted imaging [T2WI], and dynamic contrast‐enhanced [DCE] imaging) magnetic resonance imaging (mpMRI) and biparametric (DWI and T2WI) MRI (bpMRI) in detecting prostate cancer in treatment‐naïve patients.

Methods

The Medical Literature Analysis and Retrieval System Online (MEDLINE) and Excerpta Medica dataBASE (EMBASE) were searched to identify relevant studies published after 1 January 2012. Articles underwent title, abstract, and full‐text screening. Inclusion criteria consisted of patients with suspected prostate cancer, bpMRI and/or mpMRI as the index test(s), histopathology as the reference standard, and a DTA outcome measure. Methodological and DTA data were extracted. Risk of bias was assessed using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS)‐2 tool. DTA metrics were pooled using bivariate random‐effects meta‐analysis. Subgroup analysis was conducted to assess for heterogeneity.

Results

From an initial 3502 studies, 31 studies reporting on 9480 patients (4296 with prostate cancer) met the inclusion criteria for the meta‐analysis; 25 studies reported on mpMRI (7000 patients, 2954 with prostate cancer) and 12 studies reported on bpMRI DTA (2716 patients, 1477 with prostate cancer). Pooled summary statistics demonstrated no significant difference for sensitivity (mpMRI: 86%, 95% confidence interval [CI] 81–90; bpMRI: 90%, 95% CI 83–94) or specificity (mpMRI: 73%, 95% CI 64–81; bpMRI: 70%, 95% CI 42–83). The summary receiver operating characteristic curves were comparable for mpMRI (0.87) and bpMRI (0.90).

Conclusions

No significant difference in DTA was found between mpMRI and bpMRI in diagnosing prostate cancer in treatment‐naïve patients. Study heterogeneity warrants cautious interpretation of the results. With replication of our findings in dedicated validation studies, bpMRI may serve as a faster, cheaper, gadolinium‐free alternative to mpMRI.

 

 

View more videos

 

© 2024 BJU International. All Rights Reserved.