Evaluation of magnetic resonance imaging and targeted biopsy: The difficulty of finding the right reference standard.

Abstract

We read with interest the study by Truong et al,1 who have proposed a novel tool for predicting benign prostate pathology after a prior negative 12-core systematic biopsy. The study included 285 patients, and 46.3% had benign histological findings after targeted biopsy despite abnormal magnetic resonance imaging (MRI). This was defined as a false-positive MRI. This approach highlights the most challenging and controversial aspect of evaluating a targeted biopsy: defining the clear reference standard. Because targeted biopsy is used in an interdisciplinary setting, all physicians involved (eg, urologists, radiologists, and pathologists) have a particular learning curve and must contend with uncertainty in the interpretation of their results.2 Gaziev et al3 elucidated this in a recent study showing that radiologists and urologists as well as technological differences could affect targeted biopsy results. In other words, using targeted biopsy results as the reference standard has clear limitations. If prostate MRI reveals suspicious lesions and the subsequent targeted biopsy yields benign histologic findings, there are theoretically 2 possible sources of error: the Prostate Imaging Reporting and Data System score is falsely high, or the targeted biopsy has failed to accurately sample the lesion in question. To underscore this, Cash et al4 analyzed a subgroup of 61 patients with suspicious MRI findings, a negative targeted biopsy, and an additional random systematic biopsy, which still revealed a high rate of clinically significant prostate cancer. In this study, both the failure of the targeted approach to catch the suspicious lesion and the inaccurately positive MRI scores contributed to these findings. Despite these limitations, the objective that Truong et al1 tried to investigate with their study is still useful: persistent imaging and laboratory evidence for prostate cancer despite a negative biopsy is a common and vexing problem. The ideal setting for creating a predictive nomogram to this end would involve MRI analysis and corresponding biopsy results together with whole-mount specimens of the prostate to truly evaluate the diagnostic accuracy of the new technology. The drawback is that this would sample only those patients who ultimately went on to undergo radical prostatectomy, and thus a clear selection bias would be introduced. Thompson et al5 studied this question with such a patient cohort, but they did not seek to develop a nomogram. The goal of developing and validating a standardized tool to help to reduce the number of unnecessary biopsies is clearly worthwhile, but we must take into account the underlying limitations of how these data were obtained. Sebastian Berg is supported by a grant from the German Research Foundation. Quoc-Dien Trinh is supported by an unrestricted educational grant from the Vattikuti Urology Institute, a Clay Hamlin Young Investigator Award from the Prostate Cancer Foundation, and a Genentech BioOncology Career Development Award from the Conquer Cancer Foundation of the American Society of Clinical Oncology. Sebastian Berg reports a grant from the German Research Foundation. Quoc-Dien Trinh reports honoraria from Bayer and Astellas Pharma US, outside the submitted work, a Clay Hamlin Young Investigator Award from the Prostate Cancer Foundation, and a Genentech BioOncology Career Development Award from the Conquer Cancer Foundation of the American Society of Clinical Oncology; he also receives research funding from Bayer and Astellas. Sebastian Berg, MD Department of Urology Marien Hospital Herne Ruhr-University Bochum Herne, Germany Sean A. Fletcher, BS Alexander P. Cole, MD Quoc-Dien Trinh, MD Division of Urological Surgery and Center for Surgery and Public Health Brigham and Women's Hospital Harvard Medical School Boston, Massachusetts