Establishing models based on ultrasonographic and demographic characteristics for predicting breast imaging reporting and data system classification
10.13929/j.1003-3289.201904139
- Author:
Peng WANG
1
Author Information
1. Medical Examination Center, Peking University Third Hospital
- Publication Type:Journal Article
- Keywords:
Breast neoplasms;
Logistic models;
Ultrasonography
- From:
Chinese Journal of Medical Imaging Technology
2019;35(9):1341-1345
- CountryChina
- Language:Chinese
-
Abstract:
Objective: To explore the value of models based on characteristics of ultrasonography and demography for predicting breast imaging reporting and data system (BI-RADS) classification. Methods: Breast ultrasonic data and demographic data of 5 324 females who underwent health screening were retrospectively analyzed. Multivariate Logistic regression analysis was used to establish model 1 based on breast ultrasonic characteristics, and model 2 based on breast ultrasonic characteristics as well as demographic characteristics. ROC curve was used to analyze the efficacy of the two models for BI-RADS≥4a breast lesions. Results: Ultrasound showed 5 019 (5 019/5 324, 94.27%) BI-RADS≤3 and 305 (305/5 324, 5.73%) BI-RADS ≥4a grade breast lesions. Logistic regression analysis showed that the number of nodules, morphology, echo, blood flow signal, age and body mass index (BMI) were independent predictors of BI-RADS≥4a lesions (all P<0.05). Regression model 1 was constructed based on nodule number, morphology, echo and blood flow signals, with AUC of predicting BI-RADS≥4a grade 0.821 (P<0.05), specificity of 90.58%, sensitivity of 61.25% and accuracy of 88.13%. Regression model 2 was constructed based on nodule number, morphology, echo, blood flow signal, age and BMI, with AUC of predicting BI-RADS≥4a grade 0.874 (P<0.05), the specificity, sensitivity and accuracy was 93.69%, 68.75%, and 91.80%, respectively. Conclusion: Models based on ultrasonic features and demographic characteristics have certain predictive value for BI-RADS classification.
