We read with interest the recent ‘Article of the Month’ by Park et al. in which they concluded that multi-parametric 3T-MRI can be used to predict adverse pathological features and to assess eligibility of patients for active surveillance (AS), in those initially meeting the PRIAS criteria . Nevertheless, we would urge a degree of caution before widespread adoption of this strategy in patient selection for AS.
Firstly, it is accepted that there are false positives with multi-parametric MRI, with the addition of contrast only leading to a minor increase in accuracy, due to increased sensitivity being offset by reduced specificity . Thus, it is imperative to at least make some effort to correlate tumour site on MRI with site on histopathology, which the authors acknowledge was not performed in their study.
Whilst realising that substantial technical difficulties arise in the correlation of imaging with radical prostatectomy specimens, we believe that, as a minimum, tumour side on MRI should be compared to tumour side on histopathology. This is relatively straightforward and could have been performed by Park et al., since 41% of the patients in the study were pathological stage T2a/b.
For example, an audit of 76 patients suitable for AS at our unit (Wirral University Teaching Hospital, UK), but electing for mapping transperineal template guided saturation biopsy, revealed that 53 patients had undergone MRI with diffusion weighted and 23 patients full multi-parametric dynamic contrast enhanced imaging, using a 1.5 T scanner. When analysed without correlation to tumour side the sensitivity was 83%, specificity was 68% and positive predictive value was 79%. However, when analysed with respect to tumour side on MRI with tumour side on histopathology the result becomes 73%, 61%, 79% respectively.
Park et al. concludes that MRI can be used to assess the eligibility of patients with PCa for AS, which is not backed up by their data. With only 11.7% exhibiting no tumour visible on imaging, the investigation will only exclude a relatively small proportion of patients from AS, whereas in the 88.3% with visible cancer on imaging, 50.2% did not have their cancer upgraded and 47.9% had favourable disease on final histology of the whole specimen. Thus, the authors have demonstrated a statistical significant correlation between identification of a lesion on MRI and the risks of upgrading and unfavourable disease, but not demonstrated that multi-parametric 3T MRI is a clinically useful investigation in this setting. In essence, introduction of MP-MRI would be likely to exclude more patients suitable for AS than those with adverse pathology. It seems more likely that it can only be built into a nomogram, rather than a stand-alone assessment tool.
Debashis Sarkar†*, Nijel J Parr**
*Research Fellow Urology, **Consultant Urologist, Wirral University Teaching Hospital, Upton, UK
†Correspondence: Debashis Sarkar, Research Fellow Urology, Wirral University Teaching Hospital,
Upton, UK. e-mail: [email protected]
- Park BH, Jeon HG, Choo SH et al. Role of multiparametric 3.0-Tesla magnetic resonance imaging in patients with prostate cancer eligible for active surveillance. BJU Int 2014; 113: 864–870
- Tanimoto A, Nakashima J, Kohno H, Shinmoto H and Kuribayashi S. Prostate cancer screening: The clinical value of diffusion-weighted imaging and dynamic MR imaging in combination with T2-weighted imaging. J Magn Reson Imaging 2007; 25: 146–152