ISMRM25 - Pearls of Wisdom: Uncovering Diagnostic & Prognostic Gems with AI

ISMRM & ISMRT Annual Meeting & Exhibition • 10-15 May 2025 • Honolulu, Hawai'i

Pearls of Wisdom: Uncovering Diagnostic & Prognostic Gems with AI

Power Pitch

AI & Machine Learning

Monday, 12 May 2025

Power Pitch Theatre 2

13:45 - 15:45

Moderators: Maria Eugenia Caligiuri & Jamal Derakhshan

Session Number: PP-04

No CME/CE Credit

13:45		Screen Number: 26 0220. Leveraging the Untapped Potentials of Incomplete MRI Sequences for Glioma Grading and IDH Mutation Status Prediction View Presentation Video F. Liang, J. Yan, J. Lin, R. Wei, Y. Xu, X. Zhen, R. Yang The Second Affiliated Hospital, School of Medicine, South China University of Technology, Guangzhou, China Impact: The DISCERN model demonstrates significant potentials for real-world clinical applications in noninvasive glioma grading and IDH mutation status prediction with incomplete mp-MRI data, offering a robust tool for clinical decision-making and personalized treatment planning.
13:47		Screen Number: 27 0221. HomoNet: Enhancing Multi-Center Alzheimer's Disease Classification via Disentangling Homogeneous Features in Structural MRI L. Gao, C. Lin, Y. Dong, J. Zhang, J. Wang Ningbo University, Ningbo, China Impact: HomoNet enhances multi-center sMRI image classification, improving diagnostic accuracy for AD and MCI. It effectively addresses data heterogeneity, increases model generalizability, and provides a scalable, re-training-free solution, making it highly applicable for real-world clinical imaging.
13:49		Screen Number: 28 0222. Multi-Parametric Subtype and Stage Inference improves spatiotemporal subtyping of Alzheimer's Disease View Presentation Video Y. Huang, H. Xu, Y. Chen, S. Li, Y. Shen, Y. Shen, Z. Cao, Z. Zhao, M. Li, D. Wu Zhejiang University, Hangzhou Shi, China Impact: We identified three subtypes of Alzheimer's disease with distinct progression patterns and morphological characteristics by using multiple brain morphologic biomarkers and a modified Subtype and Stage Inference model, which can analyze the spatiotemporal heterogeneity of neurodegenerative diseases.
13:51		Screen Number: 29 0223. MLC-GCN: Multi-Level Connectomes Based GCN for AD Detection View Presentation Video Y. Fu, X. Zeng, J. Zhang, J. Detre, Z. Wang Univerisity of Maryland, Baltimore, United States Impact: The MLC-GCN classifier significantly enhances Alzheimer’s disease detection by exploiting multi-level connectomes. The clinically meaningful classifier features suggest a potential of localizing disease-related nodes or regions, facilitating clinical diagnosis and future targeted interventions.
13:53		Screen Number: 30 0224. Deep Learning Radiomics Model Based on Multiparametric MRI to Predict Extrathyroidal Extension in Papillary Thyroid Carcinoma View Presentation Video X. Li, Y. Song, H. Wang, L. Tang, X. Xie, A. Mao, Q. Chen, B. Song Fudan University Minhang Hospital, Shanghai, China Impact: This is the first DL radiomics model based on multiparametric MRI for prediction of ETE in PTC, and it could be used as a complement to ultrasound evaluation in clinical practice for PTC patients.
13:55		Screen Number: 31 0225. WITHDRAWN
13:57		Screen Number: 32 0226. Identifying Brain Regions Vulnerable and Resistant to Aging Using AI-Derived Cerebral Blood Volume from T1-Weighted MRI View Presentation Video Z. Li, Y. Zhang, A. Cao, S. Small, J. Guo Columbia University, New York, United States Impact: This non-invasive AICBV mapping technique offers an efficient alternative to traditional CBV fMRI scans, enabling large-scale and longitudinal studies of brain aging. It allows early detection of high-risk regions and guides targeted interventions to preserve cognitive health and resilience.
13:59		Screen Number: 33 0227. WITHDRAWN
14:01		Screen Number: 34 0228. Radiopathological System for Quantifying the Tumor Microenvironment and Predicting Prognosis in Osteosarcoma: A Multicenter Study X. Zhang, X. Chen, R. She, S. Li, Q. Feng, Y. Zhao School of Biomedical Engineering, Southern Medical University, Guangzhou, China Impact: This model provides a comprehensive prognostic assessment, essential for advancing predictive tools and extensive validation before clinical application.
14:03		Screen Number: 35 0229. Predicting Microsatellite Instability in Endometrial Cancer by Multi-modal MRI-based Radiomics Combined with Clinical Risk Factors View Presentation Video Q-y Wei, Y. Li, X. Huang, C. Yang, H. Qin, J-y Liao Department of Radiology, The First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi, China Impact: These machine learning models can perform non-invasive MSI status assessment for patients who are not suitable for surgery, help clinical better assess patient prognosis, evaluate the feasibility of immunotherapy in advanced patients, and provide clinical decision support.
14:05		Screen Number: 36 0230. Improving prognostic stratification in adult-type diffuse glioma using unsupervised spatiotemporal perfusion-based tumor habitat analysis View Presentation Video J. Lee, J. Jang, I. Hwang, S. H. Choi, K. S. Choi Seoul National University College of Medicine, Seoul, Korea, Republic of Impact: The Tumor Habitat Score represents an imaging biomarker that allows clinicians to stratify diffuse glioma patients more accurately based on survival risk. This facilitates personalized treatment decisions, potentially improving outcomes by identifying patients who might benefit most from targeted therapies.
14:07		Screen Number: 37 0231. Predicting Microvascular Invasion and Recurrence-Free Survival in Hepatocellular Carcinoma Using DCE-MRI Habitat Analysis and Deep Learning F. JIA, X. ZHAO, Y. JIANG, J. JIANG, Z. WANG, Y. XIONG, J. ZHANG The Second Hospital & Clinical Medical School, Lanzhou University, LANZHOU, China Impact: The fusion model's high accuracy in predicting MVI and RFS can significantly enhance preoperative planning for HCC patients, guiding personalized treatment strategies and improving prognosis. This approach opens new avenues for non-invasive cancer diagnostics and risk stratification.
14:09		Screen Number: 38 0232. Quantification of Extracellular Volume Fraction in Cardiac MRI without Blood Sampling Using Multi-Stage Training Deep Learning View Presentation Video Z. Li, K. Youssef, M. Amian, D. Yalcinkaya, V. Polsani, M. Elliott, R. Dharmakumar, R. Judd, D. Shah, O. Simonetti, M. Tong, B. Sharif Purdue University, Indianapolis, United States Impact: This study shows that incorporating additional features in a DL model enhances HCT prediction from CMR data, eliminating the need for blood sampling. This advancement could streamline ECV measurement, making it more accessible for diagnosing myocardial diseases in clinical settings.
14:11		Screen Number: 39 0233. Machine Learning Predicts Hemorrhagic Transformation after Stroke Using Diffusion and Perfusion Weighted MR Imaging View Presentation Video J. Wang, Y. Guo, S. Xu, Y. Luo Shanghai Fourth People's Hospital, Shanghai, China Impact: Practical model effectively predicts HT outcome using only three key features with high efficiency. Refined model, enhanced with domain knowledge, achieves higher accuracy and more reliable predictions, which makes it a valuable approach in decision making especially under complicated situations.
14:13		Screen Number: 40 0234. The value of machine learning approach based on echocardiography and CMR for diagnosing hypertrophic cardiomyopathy with HFpEF M. Hu, J. Wang, W. Zhu, C. Yang, Y. Leng, P. Yang, J. Dai, L. Gong the Second Affiliated Hospital of Nanchang University, Nanchang, China Impact: HCM patients with HFpEF demonstrated a markedly reduced survival rate compared to non-HFpEF. Our results suggested machine learning approaches based on echocardiography and CMR can be used to identify HFpEF which would be beneficial for the management of HCM patients.
14:15		Screen Number: 41 0235. Explainable AI Enables Early Prediction of Intramyocardial Hemorrhage Risk in Acute Myocardial Infarction (MI) Patients View Presentation Video K. Youssef, K. Vora, R. Gupta, R. Dharmakumar Indiana University School of Medicine, Indianapolis, United States Impact: By providing an accurate and interpretable method to predict hMI risk before reperfusion, this explainable-AI-based tool empowers clinicians to make informed, real-time decisions, potentially reducing complications and improving outcomes in patients mechanically revascularized for MI.
14:17		Screen Number: 42 0236. Multi-modal Adaptive Fusion Model for Breast Cancer Molecular Subtype Prediction Using Mammography and MRI View Presentation Video M. He, H. Chen, Y. Zheng, T. Tan, M. Ma Shengli Clinical College of Fujian Medical University & Department of Radiology, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital, Fuzhou 350001, China Impact: By uniquely combining 2D mammography and 3D MRI data, the multimodal deep learning model captures complementary tumor characteristics, supporting more accurate and nuanced classification across multiple subtypes, potentially aiding in treatment planning and improving patient outcomes.
14:19		Screen Number: 43 0237. CEST Radiomics for Accurate Differentiation of True Tumour Progression from Treatment Effects in Brain Metastases Using Machine Learning View Presentation Video D. Young, W. Lam, R. Chan, D. Sussman, P. Maralani, A. Sahgal, H. Soliman, G. Stanisz Toronto Metropolitan University, Toronto, Canada; Institute for Biomedical Engineering, Science and Technology (iBEST) at Toronto Metropolitan University, Toronto, Canada Impact: These findings suggest that AI-driven analysis of combined CEST and MRI features could serve as an effective diagnostic tool in clinical practice, potentially improving patient management without the need for further invasive procedures.
14:21		Screen Number: 44 0238. Ensemble Learning Stratification of Liver Histologic Fibrosis using Multi-Modal MRI and EHR Data in Pediatric and Adult Patients View Presentation Video H. Li, Z. Lu, S. Reeder, D. Harris, W. Masch, A. Aslam, K. Shanbhogue, A. Bernieh, S. Ranganathan, J. Dillman, L. He Cincinnati Children's Hospital Medical Center, Cincinnati, United States Impact: Our study demonstrated that an ensemble learning model had a moderate performance in stratifying liver fibrosis using clinical multi-modal MRI and EHR data. With further tuning, it provides a potential non-invasive means for monitoring and screening of liver fibrosis.
14:23		Screen Number: 45 0239. A BART machine learning model to classify lipomatous tumors in MR images View Presentation Video F. Godinez, N. Yuan, Y. Abdelhafez, A. Roy, H. Nalbant, C. Bateni, J. Qi, M. Zhang, S. Lee, A. Moawad, M. Guindani, L. Nardo University of California Davis, Sacramento, United States Impact: This method can assist radiologists in screening for ALT in regions with low incidence rates. By using machine learning less biopsies’ will be needed. We would like to implement this model in other detection tasks such as brain tumor classification.
14:25		Screen Number: 46 0240. An Interpretable and Generalizable Deep Learning Model for Non-Invasive Assessment of MVI in HCC Based on Preoperative MRI: A Multi-Center Study View Presentation Video X. Dong, J. Zhang, C. Ma, H. Yang, W. Zhang, X. Jia, D. Yang, Z. Yang Department of Radiology, Beijing Friendship Hospital, Capital Medical University, Yongan Road 95, West District, Beijing, 100050, China, Beijing, China Impact: Incorporating adversarial networks into the model development process may enhance the generalizability of the MVI prediction model, facilitating its practical implementation.
14:27		Screen Number: 47 0241. IVIM parameter estimation in brain tumors: Transformer-based deep learning shows potential for improved classification accuracy View Presentation Video M. Kaandorp, A. Jakab, C. Federau, P. While University Children’s Hospital Zurich, Zurich, Switzerland Impact: This study shows that transformer-based model fitting offers clinically valuable IVIM parameter estimates and potential for enhanced tumor classification accuracy. This advancement improves the noninvasive assessment of tumor heterogeneity, bringing IVIM closer to clinical use and supporting personalized treatment strategies.
14:29		Screen Number: 48 0242. Development of a Risk Calculator for Clinically Significant Prostate Cancer Using Biparametric MRI Deep Learning and Clinical Parameters L. Li, T. Peng, S. Lyu, M. Li, F. Shi, Y. Liu Affiliated Hospital of Chengdu University, Chengdu, China Impact: The developed risk calculator offers valuable insights for personalized treatment strategies and enhances prognosis evaluation in clinical practice. Its utilization can lead to more targeted biopsies, reducing unnecessary procedures and improving patient care.
14:31		Screen Number: 49 0243. Enhanced accuracy and stability in automated intra-pancreatic fat deposition monitoring of type 2 diabetes mellitus using Dixon MRI and deep learning View Presentation Video Y. Chen, Z. Pan, Q. Chen, H. Lin, B. Huang, W. Huang, F. Meng, Z. Zhong, W. Liu, Z. Li, H. Qin Shenzhen Traditional Chinese Medicine Hospital (The Fourth Clinical Medical College of Guangzhou University of Chinese Medicine), Shenzhen, China Impact: The DLR model demonstrated superior performance over radiologists, providing a more efficient, accurate and stable method for monitoring IPFD.
14:33		Screen Number: 50 0244. Habitat analysis based on mean apparent propagator-MRI for differentiating between glioblastoma and solitary brain metastasis G. Zhao, M. He, X. Ma, E. Gao, Y. Gao, X. Liu, J. Bai, M. Wang, Y. Zhang, J. Cheng The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China Impact: Morphological MRI often struggles to differentiate between glioblastoma (Gb) and solitary brain metastasis (SBM). This study seeks to identify differences between these two types of tumors through advanced diffusion imaging, with the aim of contributing to clinical practice.

Similar Session(s)