Wei T Yang1, Jay R Parikh1, A Thomas Stavros2, Pam Otto2, Greg Maislin3. 1. 1 Department of Diagnostic Radiology, Division of Diagnostic Imaging, The University of Texas MD Anderson Cancer, 1515 Holcombe Blvd, Unit 1459, Houston, TX 77030. 2. 2 Department of Radiology, The University of Texas Health Science Center at San Antonio, San Antonio, TX. 3. 3 Biomedical Statistical Consulting, Wynnewood, PA.
Abstract
OBJECTIVE: This article describes the definition and proposed utilization of negative likelihood ratios (NLRs) as statistical parameters in breast imaging. Examples with calculations are provided using BI-RADS category 4 subcategories. CONCLUSION: By auditing individual performance early and often against American College of Radiology benchmark positive predictive value ranges for the BI-RADS category 4 subcategories, and by fully understanding NLRs and their application in breast imaging, radiologists can minimize false-positive findings and unnecessary biopsies.
OBJECTIVE: This article describes the definition and proposed utilization of negative likelihood ratios (NLRs) as statistical parameters in breast imaging. Examples with calculations are provided using BI-RADS category 4 subcategories. CONCLUSION: By auditing individual performance early and often against American College of Radiology benchmark positive predictive value ranges for the BI-RADS category 4 subcategories, and by fully understanding NLRs and their application in breast imaging, radiologists can minimize false-positive findings and unnecessary biopsies.
Keywords:
BI-RADS; Bayes theorem; breast imaging biopsy; negative likelihood ratio; posttest probability; pretest probability