Differentiation Of Atypical Fibroadenoma From Invasive Breast Cancer And Classification Of Invasive Breast Cancer Based On MRI Imaging

Objective: To investigate the value of radiomics model based on multiparameter magnetic resonance T2-weighted imaging(T2WI)and diffusion-weighted imaging(DWI)in differentiating atypical breast fibroadenoma(ABF)from invasive breast cancer(IBC).Materials and methods: The imaging data of 154 patients with breast mass who were admitted to the First Affiliated Hospital of Bengbu Medical College from January 1,2019 to January 1,2021 were retrospectively analyzed.Among them,43 cases of ABF diagnosed as category 4 by Breast Imaging Reporting and Data System(BI-RADS)but confirmed as fibroadenoma by pathological biopsy and 111 cases of IBC by pathological biopsy were included.The region of interest(ROI)of the largest breast mass was selected for delineation in all cases.The dataset was randomly divided into training set and test set according to 4 ∶ 1 using Darwin platform.The radiomics features on the delineated lesions were extracted.The differences of parameters between ABF group and IBC group were compared,and the optimal radiomics feature set was selected.The receiver operating characteristic curve(ROC),sensitivity,specificity and accuracy were analyzed to explore the differential value of T2 WI alone,DWI alone and T2 WI combined with DWI image between ABF group and IBC group.Results: Finally,9 radiomics features were extracted from T2 WI model alone,9radiomics features from DWI model alone,and 19 radiomics features were extracted from T2 WI combined with DWI model.The area under the ROC curve of T2 WI and DWI were 0.80(0.65,0.94)and 0.75(0.62,0.88)for the training set,and 0.75(0.62,0.88)for the test set.The area under the ROC curve of T2 WI image combined with DWI image model was 0.85(0.75,0.96)in the training set and 0.79(0.54,1)in the test set.The sensitivity of T2 WI alone,DWI alone and combined model was 0.75,0.76,0.76 in the training set,and 0.76,0.71,0.75 in the test set,respectively.The training set specificity was 0.85,0.86,0.86,and the test set specificity was 0.86,0.80,and 0.84.The accuracy of the training set is 0.75,0.77,0.76,and the accuracy of the test set is0.77,0.68,0.74.Conclusion: T2 WI combined with DWI image radiomics model has certain value in differentiating atypical fibroadenoma and invasive breast cancer,which can assist clinical diagnosis of tumor type more accurately and avoid unnecessary biopsy as much as possible.Objective:To investigate the value of dynamic contrast-enhancement MRI(DCE-MRI)and DWI radiomics models in the pathological grading of IBC.Materials and methods: The imaging data of 142 patients with pathologically confirmed invasive breast cancer in the First Affiliated Hospital of Bengbu Medical College from January 1,2019 to June 1,2021 were retrospectively analyzed.According to pathological grade,the patients were divided into low-grade group(n=93)and highgrade group(n=49).The features of DCE-MRI and DWI images of patients were extracted using Darwin research platform.Logistic regression analysis was used to model the lesions.The receiver operating characteristic(ROC)curve and accuracy were analyzed to explore the value of DCE,DWI and their combined radiomics labels in predicting the pathological grade of IBC.Results: Finally,5 radiomics features were extracted from DCE model alone,5 imaging features were extracted from DCE combined with DWI model,and 10 radiomics labels were extracted from DWI model alone.The area under the AUC curve of DCE graph and DWI graph training set was 0.82(0.68,0.96)and 0.77(0.62,0.91),and the accuracy was 0.74 and 0.68,respectively.The area under the AUC curve of the test set was 0.79(0.62,0.97)and 0.66(0.46,0.86),respectively,and the accuracy was 0.72 and 0.60.The AUC of the model training set of DCE combined with DWI graph was 0.82(0.69,0.95),and the accuracy was 0.74.The AUC of the test set was 0.76(0.59,0.96),and the accuracy was 0.65.Conclusions: DCE image alone and DCE combined with DWI image radiomics label have certain value in predicting high and low grade IBC.DCE radiomics model alone is superior to DCE combined with DWI radiomics model.