Diagnostic value of CEM and MRI in differentiating benign and malignant breast lesions based on the 2013 BI-RADS lexicon and the supplement on CEM / 中华放射学杂志

ABSTRACT

Objective:To explore the diagnostic value of contrast-enhanced mammography (CEM) and MRI in differentiating benign and malignant breast lesions based on the 2013 breast imaging reporting and data system (BI-RADS) lexicon and the supplement on CEM.Methods:The clinical and imaging data of 83 patients with breast lesions from March 2019 to April 2022 in the Third Affiliated Hospital of Soochow University were retrospectively analyzed. Totally 100 breast lesions from 83 female patients aged 28 to 78 (49±14) years, were divided into benign lesions (50 lesions) and malignant lesions (50 lesions) according to the pathological results. The t-test, χ 2 test and Fisher′s exact test were used to compare the differences of clinical and imaging features between benign and malignant lesions, and these imaging features which had statistical differences were established CEM and MRI models by multivariate logistic regression analysis respectively. The receiver operating characteristic curves and the area under the curve (AUC) were used to assess the diagnostic efficacy of two models in differentiating benign and malignant breast lesions. Using the DeLong test compared the AUC. Results:Multivariate logistic regression analysis showed that associated features (OR=9.075,95%CI 1.430-57.570, P=0.019), lesion conspicuity (OR=6.180,95%CI 2.608-14.646, P<0.001), mass margin (OR=2.193,95%CI 1.405-3.422, P=0.001) and calcification distribution (OR=2.147,95%CI 1.157-3.986, P=0.015) were independent predictors of differentiating benign and malignant breast lesions in CEM, and then the predictive model of CEM was constructed. Time-signal intensity curve (OR=9.230, 95%CI 3.178-26.805, P<0.001), associated features (OR=5.289,95%CI 1.343-20.831, P=0.017) and mass margin (OR=2.192,95%CI 1.336-3.597, P=0.002) were independent predictors of differentiating benign and malignant breast lesions in MRI, and the predictive model of MRI was constructed. The AUC of CEM and MRI models for differentiating benign and malignant breast lesions were 0.947 and 0.930 respectively, and two models were no significant difference ( Z=0.68, P=0.494). Conclusion:The diagnostic efficacy of CEM and MRI in differentiating benign and malignant breast lesions is comparable based on the 2013 BI-RADS lexicon and the supplement on CEM.