Cape Town - 2026 ISMRM-ISMRT Annual Meeting and Exhibition - Advances in MRI Radiomics for Disease Characterization and Outcome Prediction

Advances in MRI Radiomics for Disease Characterization and Outcome Prediction

Oral

Analysis Methods

Monday, 11 May 2026

Meeting Room 1.40

08:20 - 10:10

Moderators: Shuncong Wang

Session Number: 307-02

No CME/CE Credit

Similar Sessions

This session surveys state-of-the-art MRI radiomics and deep learning methods that extract quantitative imaging biomarkers from diverse MRI contrasts (diffusion-weighted MRI, multi-parametric MRI, ASL perfusion, UTE-MRI, and high-resolution MRSI) to support precision diagnosis, molecular characterization, risk stratification, and treatment response prediction. Use cases span neuro-oncology (IDH status, glioblastoma heterogeneity), neurodegeneration (Alzheimer’s/non-AD dementia), vascular imaging (symptomatic carotid plaques), head & neck oncology (parotid tumors), gynecologic oncology (post-CCRT survival in cervical cancer), rheumatology (axial spondyloarthritis response), and bone health (osteoporotic microstructural deterioration). Emphasis is placed on interpretability/explainability, multicenter generalizability, and emerging paradigms such as federated learning and spatial statistics–driven radiomics.

Skill Level: Intermediate

08:20		307-02-001. Diffusion-Weighted MRI Radiomics Model in Predicting IDH Status of Non-Enhancing (Low-Grade-Appearing) Adult Diffuse Gliomas Yuhan Liang, Yanhong Liu, Zelong Chen, Menglin Ge, Yulin Wang Chinese PLA General Hospital, Beijing, China Impact: This radiomics model noninvasively predicts IDH status in non-enhancing gliomas (AUC >0.92), aiding preoperative molecular subtyping and personalized treatment planning. It supports the integration of imaging-based diagnostics into glioma management, potentially improving clinical decision-making.
08:31		307-02-002. Explainable Federated Multimodal MRI 3D-CNN for Predicting IDH Mutation Status in Brain Gliomas: A Multicenter Study Ao Liu, Xiuzheng Yue, Hao Yu Affiliated Hospital of Jining Medical College, jining, China Impact: This study enables privacy-preserving, multicenter collaboration for noninvasive glioma molecular profiling, providing clinicians with accurate IDH mutation prediction and reducing biopsy reliance, while encouraging researchers to explore explainable AI biomarkers and broader applications of federated learning in neuro-oncology.
08:42		307-02-003. Metabolomic imaging of the brain enabled by high-resolution MRSI Jiwei Li, Yibo Zhao, Wen Jin, Yudu Li, Rong Guo, Haiqing Zhang, Haokun Bu, Bingyang Cai, Yao Li, Zhi-Pei Liang, Jie Luo School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China Impact: Ultrafast high-resolution whole-brain MRSI provides a unique tool to enable spatial omics analysis of living human brains. It may play a key role in bridging brain molecular neurochemistry and macroscopic network organization, providing new insights into brain function and diseases.
08:53		307-02-004. Radiomics and Biomarker-Based Analysis of Arterial Spin Labelling MRI in Alzheimer’s Disease and non-AD Dementia Ha Young Kim, Ana Beatriz Solana, Edina Timko, David Shin, Julia Schnabel, PREDICTOM consortium Technical University of Munich, Munich, Germany Impact: The proposed approach enables clinicians to integrate conventional clinical biomarkers with non-invasive ASL MRI and perform a fast and computationally efficient analysis, establishing a foundation for using ASL-CBF as a biomarker in Alzheimer’s disease diagnosis.
09:04		307-02-005. Deep Learning Radiomics Signature from Multi-Contrast MRI for Automated Identification of Symptomatic Carotid Plaques Qun Gai, Jie Lu Xuanwu Hospital, Capital Medical University, Beijing, China Impact: This automated Deep Learning Radiomics Signature tool enables precise, non-invasive identification of stroke-prone carotid plaques, potentially improving risk stratification and guiding personalized prevention strategies for clinicians and patients.
09:15		307-02-006. Automated DL-Radiomics Model for Parotid Tumor Segmentation and Diagnosis on MRI Xiaofeng Tao, Ying Yuan Shanghai Ninth People's Hospital, Shanghai, China Impact: The DL-Radiomics model employed a two-step deep learning framework for segmentation and classification of parotid tumors. The model demonstrated high accuracy in distinguishing between benign and malignant PTs, with robust performance across both validation and external testing cohorts.
09:26		307-02-007. Local Moran’s I–based Radiomics Predicts Post-CCRT Survival in Locally Advanced Cervical Cancer: Multicenter Study Yidan Zhang, Jie Li, Qinghe Han The Second Hospital of Jilin University, Jilin, China Impact: Introducing Local Moran’s I into MRI radiomics yields explainable spatial habitats predicting post-CCRT survival in LACC. Enables transparent risk stratification and trial enrichment; motivates multi-center prospective validation and biologic correlates linking spatial patterns to tumor heterogeneity.
09:37		307-02-008. UTE-MRI Based Radiomic Analysis of Trabecular Bone Detects Osteoporotic Bone Deterioration in Femoral Head Jiyo Athertya, Jisook Yi, Arya Suprana, Jiayang Wu, Yajun Ma, Gina Woods, Jiang Du, Saeed Jerban University of California, San Diego, United States of America Impact: Radiomic texture analysis of 3D IR‑UTE MRI enables fast, noninvasive quantitative characterization of trabecular bone integrity. These imaging biomarkers may serve as indicators of bone fragility and complement traditional BMD assessments to improve clinical evaluation and management of osteoporosis risk.
09:48		307-02-009. Guided Discussion of Advances in MRI Radiomics for Disease Characterization and Outcome Prediction Shuncong Wang University of Cambridge, Cambridge, United Kingdom