Cape Town - 2026 ISMRM-ISMRT Annual Meeting and Exhibition - Prognostication and Predictive Value of MR Biomarkers

Prognostication and Predictive Value of MR Biomarkers

Power Pitch

Analysis Methods

Thursday, 14 May 2026

Power Pitch Theatre 2

08:30 - 10:06

Moderators: Enamul Bhuiyan & Priyanka Bhat

Session Number: 652-01

No CME/CE Credit

Similar Sessions

The session focused on research towards the identification of MR biomarkers with prognostic and predictive value that can help improve patient outcome through optimised treatment planning.

Skill Level: Intermediate,Advanced

08:30		652-01-001. Harmonizing Patch-Based Radiomics in Longitudinal Glioblastoma MRI: Can Canonical Correlation Analysis Boost Interpretation? Marta Loureiro, Catarina Passarinho, Ana Matoso, M. Rosário Oliveira, Pedro Vilela, Patricia Figueiredo, Rita Nunes Instituto Superior Técnico, Universidade de Lisboa, Lisbon, Portugal Impact: Technical variations greatly influence MRI-derived radiomics features. We successfully harmonized radiomics features from small patches of longitudinal glioblastoma MRI using both ComBat and longComBat. Canonical Correlation Analysis suggested batch effects were more effectively removed with ComBat, while preserving biological information.
08:32		652-01-002. Prediction of High Mortality and Morbidity Incident Disease Using Machine Learning with Wholistic Imaging and Clinical Traits Karl Landheer, Benjamin Geraghty, Joseph Herman, Joshua Backman, Manuel Ferreira, Goncalo Abecasis, Jonathan Marchini Regeneron Genetics Center, LLC, Tarrytown, United States of America Impact: Machine learning models were trained to predict the risk of incident diseases with highest mortality and morbidity via a combination of wholistic imaging features and clinical biomarkers. This work investigates which diseases could benefit from the addition of imaging.
08:34		652-01-003. Integrating DCE-MRI and DW-MRI for Patient-Specific Prediction of Drug Transport Dynamics in Solid Tumors Hooman Salavati, Charlotte Debbaut, Pim Pullens, Wim Ceelen Ghent University, Ghent, Belgium Impact: This research demonstrates that patient-specific, image-based CFD models can capture tumor heterogeneity in interstitial fluid flow and drug distribution, offering a predictive tool to optimize chemotherapy delivery and support personalized treatment strategies in solid tumors.
08:36		652-01-004. Interpretable Machine Learning Model for predicting Neoadjuvant Chemotherapy Response in Advanced Olfactory Neuroblastoma Shanbin Sun, Xiwen Wang, yongchao wu, Dandan Zheng, Jing Peng, Zhaohui Liu Beijing Tongren Hospital, Capital Medical University, Beijing, China Impact: The fusion model integrating radiomic, clinical, and pathological features shows promise for predicting NACT response in advanced ONB, potentially aiding treatment stratification and personalized therapeutic decisions for this rare malignancy.
08:38		652-01-005. Integrated Machine Learning Model for Predicting Prostate Cancer Progression from mpMRI Radiomics and Clinical Data Sohaib Naim, Siriluck Satonkiatngam, Hyun Lim, Qi Miao, Kai Zhao, Katarina Chiam, Wayne Brisbane, Leonard Marks, Steven Raman, Holden Wu, KyungHyun Sung David Geffen School of Medicine at UCLA, Los Angeles, United States of America Impact: This study demonstrates the potential of integrating routinely collected clinical data with multiparametric MRI-based radiomics and delta radiomics to improve progression-free survival prediction in prostate cancer patients under active surveillance, supporting more precise risk stratification and improved clinical outcomes.
08:40		652-01-006. Incremental Value of Epicardial Adipose Tissue to Predict Major Adverse Cardiovascular Events in Hypertrophic Cardiomyopathy Yipei Song, Mengyao Hu, Qimin Fang, Jiankun Dai, Lianggeng Gong The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, China Impact: This study revealed additional applying epicardial adipose tissue (EAT) can improve clinical-LGE and ESC Risk-SCD models for predicting major adverse cardiovascular events, helping more precise stratification and highlighting EAT as a potential therapeutic target for patients with hypertrophic cardiomyopathy.
08:42		652-01-007. MRI-based Assessment of Tumor Aggressiveness in Nasopharyngeal Carcinoma: Risk Stratification and Survival Prediction Fan Yang, zhenyu Huo, Haoran Wei, Meng Lin, Hongmei Zhang National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China Impact: Critical imaging parameters, including Radiological depth, Tumor enhancement margin, and Hypointense on T2WI , serve as significant predictors of long-term survival.
08:44		652-01-008. Alterations in Functional Connectivity Associated with Prognosis in Unilateral Sudden Sensorineural Hearing Loss Jin Sun, Yifei Zhang, Zihan Li, Yufei Shen, Shaoqiang Han, Yong Zhang, Baohong Wen The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China Impact: STG-seeded FC patterns stratify the prognosis of sudden sensorineural hearing loss (SSNHL) prior to therapy, highlighting maladaptive cross-modal and default-mode/attention network engagement. These neuroimaging markers may enable earlier patient counseling and inform trials of targeted neuromodulation alongside time-sensitive standard care.
08:46		652-01-009. Long-term prognostic value of left ventricular trabeculae fractal analysis in patients with known or suspected coronary arter Naoki Hashimoto, Masaki Ishida, Haruno Ito, Satoshi Nakamura, Masafumi Takafuji, Yasutaka Ichikawa, Shiro Nakamori, Tairo Kurita, Kaoru Dohi, Hajime Sakuma Mie University Hospital, Tsu, Japan Impact: Left ventricular cavity trabecular complexity has mild but significant prognostic value for cardiovascular events, even in patients with known or suspected coronary artery disease. This distinct phenotypic endocardial structure itself may contribute to long-term cardiac dysfunction.
08:48		652-01-010. Pulmonary Transit Time Assessed by Cardiac Magnetic Resonance: A Prognostic Marker in Dilated Cardiomyopathy Wendi Zhang, Xu Xu, Wanlin Peng, Xinyang Lyu, Ke Shi, Chunchao Xia, Zhenlin Li West China Hospital, Sichuan University, China Impact: Pulmonary transit time offers incremental prognostic value beyond cardiac function and strain parameters in dilated cardiomyopathy. As a parameter conveniently obtained from cardiac magnetic resonance first-pass perfusion sequences, it can be easily integrated into the management workflow for dilated cardiomyopathy.
08:50		652-01-011. AI-driven Automated Risk Prediction of Ventricular Arrhythmias and Sudden Death from Late Gadolinium Enhancement Cardiac MRI Joao Santinha, Miguel Antunes, Minh Nhat Trinh, João Augusto, Sílvia Rosa, Teresa Correia Champalimaud Foundation, Lisboa, Portugal Impact: This work introduces a fully automated radiomics-based survival model that predicts ventricular arrhythmias in HCM by assessing myocardial quality. This paradigm shift from scar quantity to quality enables a non-invasive, personalized, and dynamic risk assessment, improving upon current static models.
08:52		652-01-012. Deep Learning Prediction of Patient Biological Profile from Knee MR Imaging Yu-Cherng Chang Jackson Memorial Hospital/University of Miami, United States of America Impact: Successful prediction of patient biological profiles from knee MRI images supports understanding of the range of normal knee physiology, which in turn, allows greater discrimination of patient-incongruous anatomy potentially relevant to early or subtle pathological changes.
08:54		652-01-013. Predicting Aggressive Cribriform Prostate Cancer Using Pre-operative Multiparametric MRI Nader Gharbia, Yasmine Saad, Aymen Kammoun, Kays Cheker, Yassine Nouira Faculty of medicine of Sfax, Tunisia Impact: Pre-operative mpMRI features, including ADC ratio, multizonal involvement, and DCE patterns, can non-invasively predict aggressive cribriform prostate cancer. This enables improved surgical planning, risk stratification and opens avenues for further research on imaging biomarkers of high-risk prostate subtypes.