Abstract #0356

Exploring the Application of an Multi-Advanced-Diffusion-Model-Based Radiomics Method in Grading of Intestinal Fibrosis of Crohn's Disease

Li Huang¹, Xuehua Li¹, Siyun Huang², Mengzhu Wang³, Canhui Sun², Shiting Feng², Xu Yan³, Yang Song⁴, and Ziping Li¹

¹The First Affiliated Hospital of Sun Yat-Sen University, Guangzhou, China, ²The First Affiliated Hospital of Sun Yat-Sen University, guangzhou, China, ³MR Scientific Marketing, Siemens Healthcare, guangzhou, China, ⁴Shanghai Key Laboratory of Magnetic Resonance, East China Normal University, shanghai, China

In the present study, we investigated a multiple diffusion-model-based radiomics model in grading of bowel fibrosis of Crohn's disease (CD) patients. We used histogram features derived from parameters of DWI, IVIM and DKI models and chose SVM as the classifier to construct a prediction model. The results showed that the most accurate prediction was achieved by incorporating the following 6 features into a nomogram, including the DKI-related histogram parameters (mean D, mean K, 10th percentiles of K) and IVIM-related parameters (mean D_0-2000, mean D*_0-2000, 90th percentiles of f), with AUC and accuracy reached 0.835 and 0.833, respectively.

This abstract and the presentation materials are available to members only; a login is required.

Join Here