Category: Imaging & Image Guided Therapy: New Therapies

MP4-7 - The Accuracy of Prostate Magnetic Resonance Imaging Interpretation: What is the Impact of the Individual Radiologist?

Fri, Sep 21
10:00 AM - 12:00 PM

Introduction & Objective : Prostate multiparametric magnetic resonance imaging (MRI) has emerged as a useful tool for the management of prostate cancer (PCa).  Despite attempts at standardization with the Prostate Imaging-Reporting and Data System (PI-RADS), there remain differences among individual radiologists’ interpretations.  We aim to define the degree of variation in accuracy among individual radiologists as well as identify factors that predict the accurate interpretation of prostate MRI. 


Methods : We examined our prospectively maintained database of men who received prostate MRI prior to biopsy between September 2015 and December 2016 (n=459).  Prostate MRI was read as part of the clinical workflow of the abdominal imaging section at our institution, by nine fellowship-trained attending radiologists who were supervising fellows and residents.  Sensitivity, specificity, and accuracy were calculated for each attending radiologist, treating PI-RADS classification 4 or 5 as test positive and Gleason 7+ on biopsy as outcome positive.  


Results : No statistically significant differences were found in sensitivity, specificity, or accuracy among the individual radiologists, with overall sensitivity, specificity, and accuracy of 86%, 56%, and 65%, respectively.  Multivariate analyses are provided in Table 1.  Prior biopsy (OR 0.10, p=0.01), radiologist experience greater than 500 prostate MRI (OR 0.18, p=0.04), and lesion location (transition zone OR 0.10, p=0.04, posterior OR 0.04, p=0.03) were predictors of diminished sensitivity.  Lesion location in the transition zone (OR 2.53, p=0.04) was a predictor of improved specificity.  Prior biopsy status (OR=0.50, p<0.01) was a predictor of reduced accuracy.


Conclusions : In our experience, the variations in test performance (sensitivity, specificity, accuracy) with PI-RADS for prostate MRI were not significantly different among individual radiologists.  Prior biopsy status was a significant predictor of reduced test accuracy for prostate MRI, indicating a role for MRI in the biopsy-naïve setting.

Gerald Andriole

Chief of Urologic Surgery
Washington University School of Medicine
St. Louis, Missouri

Gerald L. Andriole, Jr., MD
Robert K. Royce Distinguished Professor
Chief of Urologic Surgery
Siteman Cancer Center
Barnes-Jewish Hospital
Washington University School of Medicine
St. Louis, Missouri

Gerald L. Andriole, Jr., MD, is the Robert K. Royce Distinguished Professor and Chief of Urologic Surgery at Barnes-Jewish Hospital, the Siteman Cancer Center and Washington University School of Medicine in St. Louis, Missouri.

Dr. Andriole received his medical degree from Jefferson Medical College in Philadelphia, Pennsylvania. He trained in surgery at Strong Memorial Hospital and the University of Rochester and completed urology residency at Brigham and Women’s Hospital and Harvard Medical School. Subsequently, he was a fellow in Urologic Oncology at the National Cancer Institute in Bethesda, Maryland.

Dr. Andriole has over 30 years of consistent contributions in the areas of BPH and prostate cancer screening and prevention research. He has contributed well over 300 peer-reviewed publications and serves on the editorial boards of several prestigious journals. He is Chairman of the Prostate Committee of the National Cancer Institute’s Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial and of the Steering Committee of the REDUCE Prostate Cancer Prevention Trial. He is a member of the American Urological Association, American Association for Cancer Research, the American Society of Clinical Oncology, the Prostate Cancer Early Detection Panel of the National Comprehensive Cancer Network, American Surgical Association and the American Association of Genitourinary Surgeons among other societies.

Eric H. Kim

Assistant Professor
Washington University in St. Louis
St. Louis, Missouri

Eric H. Kim, M.D., is an Assistant Professor of Surgery in the Division of Urologic Surgery at Washington University School of Medicine in St. Louis. Clinically, he specializes in robotic surgery for urologic cancers. His research focuses on imaging in prostate cancer.

Joel Vetter

Statistician
Washington University in St. Louis
Saint Louis, Missouri

Nicholas A. Pickersgill

Medical Student
Washington University School of Medicine in St. Louis
St. Louis, Missouri

Anup Shetty

St. Louis, Missouri