Predictive Value of PI-RADS Is Low and Variable. Or Is It? Lessons We Can Learn from BI-RADS

Abstract

HomeRadiology: Imaging CancerVol. 2, No. 6 PreviousNext Research HighlightsFree AccessPredictive Value of PI-RADS Is Low and Variable. Or Is It? Lessons We Can Learn from BI-RADSMateus TaveiraMateus TaveiraMateus TaveiraPublished Online:Nov 27 2020https://doi.org/10.1148/rycan.2020209033MoreSectionsPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked InEmail Take-Away Points■ Major Focus: The Prostate Imaging Reporting and Data System (PI-RADS) classification system for prostate MRI established a standardized five-point scale for likelihood of clinically significant prostate cancer from 1 (very low) to 5 (very high). The true positive predictive value (PPV) of each PI-RADS category remains poorly validated.■ Key Result: In a retrospective cross-sectional study across 26 centers, PPV of PI-RADS ≥3 lesions (ie, intermediate to very high chance of malignancy) was 35% with large variability among centers and within a single center.■ Impact: PPV of PI-RADS classification system varies widely, but some centers have commendable precision and accuracy. Identifying factors contributing to high performance in some centers will improve global reproducibility and accreditation for prostate MRI.Because of adverse consequences of overdiagnosis and overtreatment of prostate cancer, imaging societies have developed standardized classification and reporting systems for prostate MRI (PI-RADS). PI-RADS version 2.1, updated in 2019, has been implemented worldwide. Despite single-center published data supporting the value of PI-RADS, concerns persist regarding its cross-center comparability and, more importantly, if the published performances could be replicated in community centers where most cancer screening occurs.Westphalen et al concluded that the PPV of PI-RADS was low and varied widely across centers. A historical comparison with BI-RADS, a much more mature reporting system for breast cancer screening, offers hope that the PPV of PI-RADS, even in the less accurate centers, is not that bad. Twenty-three years after initial publication of BI-RADS, an analysis by Salim et al showed that even among 24 high-volume readers, the PPV of an abnormal screening examination (BI-RADS categories 0, 4 or 5) was 8% and ranged from 4% to 14%. High-volume readers all had similar specificity (93%–99%), and variations in PPV arose from differences in reader sensitivity (53%–86%).The historical evolution of BI-RADS and breast cancer screening provides a path to make PI-RADS more accurate in a much shorter time frame. First, systematic biopsy of the prostate allows more opportunities to evaluate MRI findings against pathology samples, as compared with breast biopsy that only samples the suspicious area. Second, unlike mammograms, multiparametric MRI offers not only morphologic but also functional information, and MRI technologies for prostate MRI continue to advance more rapidly than mammography. Finally, the prostate cancer community can learn from BI-RADS strategies for audits and accreditation to kickstart the PI-RADS qualification process. Radiologists are not the only professionals involved in prostate cancer screening, but their involvement in the field through research, education, and quality control will be paramount for improving the lives of these patients while minimizing the risks of overdiagnosis and overtreatment.Highlighted ArticlesWestphalen AC, McCulloch CE, Anaokar JM, et al. Variability of the positive predictive value of PI-RADS for prostate MRI across 26 centers: experience of the Society of Abdominal Radiology Prostate Cancer Disease-focused Panel. Radiology 2020;296(1):76–84. doi:10.1148/radiol.2020190646Salim M, Dembrower K, Eklund M, Lindholm P, Strand F. Range of radiologist performance in a population-based screening cohort of 1 million digital mammography examinations Radiology 2020;297(1):33–39. doi:10.1148/radiol.2020192212Highlighted ArticlesWestphalen AC, McCulloch CE, Anaokar JM et al. Variability of the positive predictive value of PI-RADS for prostate MRI across 26 centers: experience of the Society of Abdominal Radiology Prostate Cancer Disease-focused Panel. Radiology 2020;296(1):76–84. doi:10.1148/radiol.2020190646 Link, Google ScholarSalim M, Dembrower K, Eklund M, Lindholm P, Strand F. Range of radiologist performance in a population-based screening cohort of 1 million digital mammography examinations Radiology 2020;297(1):33–39. doi:10.1148/radiol.2020192212 Link, Google ScholarArticle HistoryPublished online: Nov 27 2020 FiguresReferencesRelatedDetailsRecommended Articles Influence of Risk Category and Screening Round on the Performance of an MR Imaging and Mammography Screening Program in Carriers of the BRCA Mutation and Other Women at Increased RiskRadiology2017Volume: 286Issue: 2pp. 443-451ACR BI-RADS Assessment Category 4 Subdivisions in Diagnostic Mammography: Utilization and Outcomes in the National Mammography DatabaseRadiology2018Volume: 287Issue: 2pp. 416-422Additional Breast Cancer Detection at Digital Screening Mammography through Quality Assurance Sessions between Technologists and RadiologistsRadiology2020Volume: 294Issue: 3pp. 509-517Multicenter Evaluation of Breast Cancer Screening with Digital Breast Tomosynthesis in Combination with Synthetic versus Digital MammographyRadiology2020Volume: 297Issue: 3pp. 545-553Closing the Chapter on Supplemental Breast Cancer Screening with USRadiology2021Volume: 298Issue: 3pp. 576-577See More RSNA Education Exhibits Non-Contrast-Enhanced Breast MR Screening for Women with Dense BreastsDigital Posters2019The First 60 Days of COVID-19: A Mandatory Rapid and Unusual Distillation of Clinical Acuity of Breast Disease and Management ImplicationsDigital Posters2020BIRADS Mammography: What Do Residents Need to Know?Digital Posters2017 RSNA Case Collection Invasive Lobular CarcinomaRSNA Case Collection2021Invasive ductal carcinoma as developing asymmetryRSNA Case Collection2021Anterior transition zone prostate adenocarcinomaRSNA Case Collection2020 Vol. 2, No. 6 Metrics Downloaded 92 times Altmetric Score