ZHAO Youfan,CHEN Zhongwei,ZHOU Jiejie, et al. Diagnostic value of nomogram based on the radiomics of DCE-MRI and clinical features in breast MRI BIRADS 4 lesions[J]. JOURNAL OF WEZHOU MEDICAL UNIVERSITY, 2021, 51(5): 375-380.
Abstract：Objective: To investigate the diagnostic value of nomogram based on radiomics of dynamic
contrast-enhanced magnetic resonance imaging (DCE-MRI) and clinical features in breast MRI BI-RADS 4
lesions. Methods: Totally 189 pathologically confirmed breast lesions with MRI BI-RADS 4 in the First Affiliated Hospital of Wenzhou Medical University from January 2017 to June 2019 were retrospectively analyzed, including 71 benign and 118 malignant lesions. All patients underwent DCE-MRI scan and blood biochemical examination before pathological examination. The parameters of DCE-MRI were calculated and the radiomics features were extracted from them. After feature selection, the rad-score was calculated by weighed sum of the selected features according to their coefficients. Univariate and multivariate logistic regression analyses were used to analyze the clinical risk factors of breast cancer. Finally, multivariate logistic regression was used to establish a nomogram based on clinical risk factors and the rad-score. Receiver operating characteristic (ROC) curve was used to evaluate the diagnostic performance of the model. Results: Univariate and multivariate analysis showed that age,low density lipoprotein cholesterol and total bilirubin levels were risk factors of breast cancer. The clinical model based on clinical risk factors yielded an AUC of 0.73 (0.58-0.87) in the test cohort. After feature selection, 11 radiomics features were selected to calculate the rad-score, yielding an AUC of 0.80 (0.68-0.92) in the test cohort.When combining the rad-score with the clinical model, the AUC was improved to 0.88 (0.79-0.97), showing statistical significance (P=0.037). Conclusion: The nomogram based on radiomics of DCE-MRI and clinical features can effectively differentiate benign and malignant breast lesions with MRI BI-RADS 4.
