The Value Of Radiomic Model Based On Dynamic Contrast-enhanced MRI In Predicting Neoadjuvant Chemotherapy Pathological Response Of Breast Cancer

Objective: To develop and validate a radiomics nomogram integrating the radiomics features from dynamic contrast-enhanced MRI(DCE-MRI)and clinical factors to predict the pathological response of breast cancer patients to neoadjuvant chemotherapy(NAC).Materials and Methods: We retrospectively analyzed the clinical,pathological and post-NAC DCE-MRI data of 129 breast cancer patients confirmed by pathology and received NAC from February 2016 to May2020.The state of estrogen receptor(ER),progesterone receptor(PR),human epidermal growth factor receptor 2(HER-2),nuclear-associated antigen(Ki-67)were evaluated by the immunohistochemistry(IHC)SP methods.Miller-Payne grading system were used to access treatment response.Patients were divided into pathological complete response group of 42 cases and non-pathological complete response group of 87 cases.All cases were subdivided into primary cohort(n=88)and independent validation cohort(n=41)in a ratio of 7:3.The ROI of DCE-MRI images was manually delineated using Mazda software.A total of 386 feature parameters including first order statistics,absolute gradient,gray-level cooccurrence matrix,grey level run length matrix,wavelet parameters and autoregressive model parameters were extracted by Mazda software automatically.The least absolute shrinkage and selection operator(LASSO)regression method was used to screen the optimal features with the method of 10 fold cross-validation,so as to establish the radiomics signatures.The receiver working curve(ROC)were evaluated diagnostic efficiency of the radiomics signature primary cohort and validation cohort.Univariate logistic regression analysis was used to analyze the clinicopathological factors,Multivariable logistic regression analysis was used to establish a radiomics nomogram model based on radiomics signature and clinicopathological factors.The calibration curve was used to evaluate the model,and the area under the curve(AUC),sensitivity and specificity of the nomogram was calculated,and Decision curve analysis was conducted to determine the net clinical benefits of the radiomics nomogramResults: The average age of 129 patients was(60.3 ± 5.3)years.The386 radiomic features were extracted from the DCE-MRI sequence.A total of 9 optimal feature parameters were selected to establish radiomics signature.The radiomics signatures achieved moderate prediction efficacy with AUC 0.746 and 0.779 in the primary and validation cohorts,respectively.There was a significant difference in radiomics score,clinical stage,ER and PR between p CR and non-p CR group in both cohort(P<0.05).there was no significant difference in age,state of menstrual,HER-2 and Ki-67 between p CR and non-p CR group in both cohort(P>0.05).The AUC,sensitivity,and specificity of the prediction model of primary cohort combined with clinical stage,PR and radiomics signature were0.846,81.48%,and 77.05%,respectively.The AUC,sensitivity,and specificity of the validation cohort model were 0.882,86.67%,and 80.77%,respectively.The decision curve which was demonstrated the radiomics nomogram displayed good clinical utility ranged the whole threshold.Conclusion: In this study,9 radiomics features extracted from DCEMRI images can predict the efficacy of breast cancer NAC in the early stage of chemotherapy.The radiomics nomogram,based on radiomics signature and clinicopathological features,can be used as a quantitative tool to predict the pathological response of breast cancer to neoadjuvant chemotherapy,this noninvasive approach provides a new medical approach for guiding clinicians to develop personalized treatment programs.