AUTHOR=Ren Qingguo , Wang Yihua , Leng Shanshan , Nan Xiaomin , Zhang Bin , Shuai Xinyan , Zhang Jianyuan , Xia Xiaona , Li Ye , Ge Yaqiong , Meng Xiangshui , Zhao Cuiping TITLE=Substantia Nigra Radiomics Feature Extraction of Parkinson’s Disease Based on Magnitude Images of Susceptibility-Weighted Imaging JOURNAL=Frontiers in Neuroscience VOLUME=Volume 15 - 2021 YEAR=2021 URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2021.646617 DOI=10.3389/fnins.2021.646617 ISSN=1662-453X ABSTRACT=Background: It is reported that radiomic features extracted from quantitative susceptibility mapping (QSM) had promising clinical value for the diagnosis of PD. We aimed to explore the usefulness of radiomics features based on magnitude image to distinguish PD from non-PD controls. Methods: We retrospectively recruited PD patients and controls who underwent brain 3.0T MR including SWI. 396 radiomics features were extracted from the SN of 95 PD patients and 95 non-PD controls based on SWI. Intra-/inter-observer correlation coefficients (ICCs) were applied to measure the observer agreement for the radiomic feature extraction. Then the patients were randomly grouped into training and validation sets by the ratio of 7:3, in training set, the maximum correlation minimum redundancy algorithm (mRMR) and the least absolute shrinkage and selection operator (LASSO) were conducted to filter and choose the optimized subset of features and a radiomics signature was constructed. Moreover, radiomics signature were constructed by different machine learning models. AUC under the ROC curves were applied to evaluate the predictive performance of the models. Then correlation analysis was performed to evaluate the correlationship between the optimized features and clinical factors. Results: The intro-observer ICC were ranged from 0.82 to 1.0, and the inter-observer ICC were ranged from 0.77 to 0.99. The LASSO logistic regression model showed good prediction efficacy in the training set (AUC=0.82, 95% confidence interval (CI, 0.74-0.88)) and the validation set (AUC=0.81, 95% CI (0.68-0.91)). One radiomic feature showed a moderate negative correlation with Hoehn-Yahr stage (r= -0.49, P=0.012 ). Conclusion: Radiomic predictive feature based on SWI magnitude image could reflect the Hoehn-Yahr stage of PD to some extent.