Clinical Research
Quantitative indicators of diagnosis research on BI-RADS 3~5 classification in the 2013 edition of BI-RADS
Weina Zhang, Mei Peng, Fan Jiang, Xinshu Zhang, Xiabi Wu, Tingting Wu
Published 2016-05-25
Cite as Chin J Ultrasonogr, 2016, 25(5): 392-395. DOI: 10.3760/cma.j.issn.1004-4477.2016.05.008
Abstract
ObjectiveTo explore the value of BI-RADS scoring system based on the sonographic features in the breast nodules.
MethodsIn order to build a Logistic regression model, regression was made to analyse 401 patients ' sonographic features of breast nodules. A scoring system was developed based on the results of regression's filter and the weight of each factor, used to score and classify the other 243 patients. It's diagnostic value was evaluated through comparing all types of theoretic risk ranges of BI-RADS.
ResultsAge, morphology, orientation, margin, echo pattern and microcalcifications in a mass were selected in the final step of the logistic regression analysis. By means of scoring system, the scores corresponding to BI-RADS 3, 4a, 4b, 4c, 5 classes are 6, 7-8, 9-15, 16-22 and ≥23 respectively. Case study comprehensive score of BI-RADS 3-5 classification's positive predictive values were 0, 4.17%, 21.43%, 84.85%, 100%, and the area under the ROC curve scoring system was 0.947.
ConclusionsThe scoring system can objectively score and classify breast nodules, and therefore provide an effective reference for clinical evaluation of benign and malignant breast.
Key words:
Ultrasonography; Beast diseases; Breast Imaging Reporting and Data System; Score
Contributor Information
Weina Zhang
Department of Ultrasound, the Second Affiliated Hospital of Anhui Medical University, Hefei 230601, China
Mei Peng
Fan Jiang
Xinshu Zhang
Xiabi Wu
Tingting Wu