The value of quantitative features in multimodal MRI images to construct an imaging histological model for breast cancer diagnosis
Objective To analyze the value of quantitative features in multimodal magnetic resonance imaging(MRI)images to construct imaging histological models for the diagnosis of breast cancer.Methods Ninety-five patients with breast-related diseases from January 2020 to January 2021 were selected,and the pathological findings were divided into a benign group(n=57)and a malignant group(n=38),and the cases were divided into a training group(n=66)and validation group(n=29)at the time of inclusion.All the cases were examined by T1-weighted imaging(T1WI),T2-weighted imaging(T2WI),diffusion-weighted imaging(DWI),dynamic contrast enhancement(DCE),and apparent diffusion coefficient(ADC)multimodality MRI.The clinicopathological data of the two groups were compared.The MRI findings were analyzed based on the pathological findings.The support vector machine(SVM)classifier was used construct an imageomics model for diagnosis of breast cancer based on multiple imageomics indicators of patients in the training group.Breast cancer imaging histological model based on multiple imageomics indicators of patients in the training group.The validation group analyzed the diagnostic efficacy of the model,and the diagnostic efficacy was analyzed by constructing receiver operator characteristic(ROC)curves.Results Fibroadenoma accounted for 49％of benign breast diseases in this experiment,and invasive ductal carcinoma accounted for 74％of malignant breast diseases.The age and mass size of patients in the malignant group were higher than those in the benign group(P＜0.05).The sensitivity of T1WI,T2WI,ADC,DCE and DWI,in diagnosing breast cancer was found to be 61％,67％,71％,79％,and 86％using the four-grid table method.Three kinds of features were extracted from the ROI area of patients in the training group for MRI examinations of plain,diffusion,and enhancement,including morphological features,first-order features,and textural features.ROC curve analysis found that the AUC of T1WI,T2WI,DWI,ADC,and DCE for the diagnosis of breast cancer was 0.715,0.769,0.785,0.835,and 0.792,respectively.The AUCs of the multiple imaging histological models of plain,diffuse,enhanced,plain+diffuse,plain+enhanced,enhanced+diffuse,and plain+enhanced+diffuse for the diagnosis of breast cancer were 0.746,0.798,0.816,0.839,0.890,0.906.and 0.927,respectively.Conclusion The construction of imaging histological models by quantitative features in multimodal MRI images is valuable in the diagnosis of breast cancer.The value of multi-image histological models such as pan+enhancement+diffusion is higher in diagnosing breast cancer and can be widely used in clinical practice.
万方数据
维普
