AUTHOR=Xu Aqiao , Chu Xiufeng , Zhang Shengjian , Zheng Jing , Shi Dabao , Lv Shasha , Li Feng , Weng Xiaobo 

TITLE=Prediction Breast Molecular Typing of Invasive Ductal Carcinoma Based on Dynamic Contrast Enhancement Magnetic Resonance Imaging Radiomics Characteristics: A Feasibility Study

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.799232

DOI=10.3389/fonc.2022.799232

ISSN=2234-943X

ABSTRACT=Objective: To investigate the feasibility of radiomics in predicting molecular subtype of breast invasive ductal carcinoma（IDC）based on dynamic contrast enhancement magnetic resonance imaging (DCE-MRI). Methods: A total of 303 cases with IDC pathologically confirmed from January 2018 to March 2021 were enrolled in this study, including 223 from Fudan University Shanghai Cancer Center (training/test set) and 80 from Shaoxing Central Hospital (validation set). All the cases were classified as HR+/Luminal, HER2-enriched and TNBC according to immunohistochemistry. DCE-MRI original images were treated by semi-automated segmentation to initially extract wavelet texture features. The least absolute shrinkage and selection operator (LASSO) regression algorithm was applied to identify the optimal texture features, which were then used to establish predictive models combined with significant clinical risk factors. Receiver operating characteristic curve (ROC), calibration curve, and decision curve analysis were adopted to evaluate the effectiveness and clinical benefit of the models established. Results: Of the 223 cases from Fudan University Shanghai Cancer Center, HR+/ Luminal cancers were diagnosed in 116 cases (52.02%), HER2-enriched in 71 cases (31.84%) and TNBC in 36 cases (16.14%). Based on the training set, 788 texture features in total were obtained and 8 optimal features were further identified by LASSO algorithm after dimensionality reduction, including 2 first-order features, 1 gray-level run length matrix (GLRLM), 4 gray-level co-occurrence matrices (GLCM) and 1 3D shape feature. Logistic regression was used to construct three types of predictive models: clinical model (age, tumor location, Ki-67, histological grade, and lymph node metastasis), radiomic model and combined model. The areas under the ROC curve (AUC) for the three models were 0.75, 0.81, 0.84 in the training set, 0.77, 0.81, 0.83 in the test set, and 0.77, 0.82, 0.83 in the validation set, respectively. Conclusion: The DCE-MRI radiomic features are significant markers in differentiating between molecular subtypes of breast cancer, which is non-invasive. Notably, the predictive performance can be better when combined with significant clinical factors.