Clinical Application of MR Fingerprinting in Epilepsy Advances With NIH Grant Renewal

Researchers at Cleveland Clinic and Duke University have received a $3.2 million, five-year R01 renewal grant from the National Institutes of Health (NIH) to further advance the clinical application of magnetic resonance fingerprinting (MRF) in epilepsy. The renewed project aims to improve postsurgical seizure outcomes in patients with pharmacoresistant focal epilepsy, particularly those with nonlesional MRI, by integrating quantitative imaging, deep learning and intracranial EEG.

“This renewal allows us to move beyond development and into real clinical translation,” says Irene Wang, PhD, Research Director at Cleveland Clinic’s Epilepsy Center and contact PI of the grant. “We aim to make MRF a routine part of presurgical evaluation by building a fully automated, quantitative imaging pipeline with AI models that directly support surgical planning and outcome prediction.”

Quest to uncover elusive epileptogenic abnormalities

The project addresses a critical need: approximately 40% of patients undergoing evaluation for epilepsy surgery show no visible lesions on conventional clinical MRI, yet many harbor subtle cortical malformations, such as focal cortical dysplasia (FCD), that cause seizures. MRF is a novel quantitative MRI method that produces maps of tissue properties specific to microstructural pathology, providing higher sensitivity for detecting such lesions.

The team’s previous R01 grant focused on the technical development and initial validation of an MRF framework on 3T clinical MRI scanners, specifically tailored for epilepsy imaging. In this renewal, the researchers will:

• Deploy an end-to-end, motion-robust 3D MRF pipeline, making quantitative maps available to clinicians for review within minutes
• Apply machine and deep learning models trained on MRF data from seizure-free surgical patients to localize epileptogenic regions with high accuracy
• Integrate MRF with stereoelectroencephalography (SEEG) to explore the relationship between quantitative imaging and electrophysiological activity, generating co-localized anatomo-functional atlases of brain structure and function

“Our central hypothesis is that quantitative MRF can uncover subtle epileptogenic tissue abnormalities invisible on conventional clinical MRI, and that resecting these regions improves seizure outcomes,” says Dr. Wang. “By combining SEEG and MRF data from the same patients, we’ll explore how tissue properties relate to brain activity. This is a rare opportunity, since such paired data are uniquely available during intracranial EEG evaluations for epilepsy surgery. With this, we plan to create anatomo-functional atlases that show how structure and function interact in both healthy and epileptic brain regions, helping to advance future research in epilepsy and other neurological disorders.”

Building on promising foundational findings

This grant renewal builds on a highly productive previous funding cycle by the research team, with studies published in Proceedings of the National Academy of Sciences,¹ Annals of Neurology,² Epilepsia^3-5 and Cerebral Cortex.⁶

Notably, one study², highlighted on the cover of the November 2024 Annals of Neurology, developed a machine learning framework leveraging high-resolution 3D MRF data to accurately characterize FCD lesions. In a cohort of 119 subjects, the MRF machine learning model achieved over 94% sensitivity and 96% accuracy in distinguishing FCD from healthy and disease controls, far surpassing standard radiology reports, which identified only 48% of lesions. Remarkably, the model succeeded even in cases where both clinical MRI and post-processing were negative. It also differentiated FCD subtypes with high accuracy, demonstrating its potential as a powerful, noninvasive tool for epilepsy surgical planning.

References

1. Jordan SP, Hu S, Rozada I, et al. Automated design of pulse sequences for magnetic resonance fingerprinting using physics-inspired optimization. Proc Natl Acad Sci. 2021;118(40):e2020516118.

2. Su TY, Choi JY, Hu S, et al. Multiparametric characterization of focal cortical dysplasia using 3D MR fingerprinting. Ann Neurol. 2024;96(5):944-957.

3. Choi JY, Krishnan B, Hu S, et al. Using magnetic resonance fingerprinting to characterize periventricular nodular heterotopias in pharmacoresistant epilepsy. Epilepsia. 2022;63(5):1225-1237.

4. Ding Z, Ting- SH, Su Y, et al. Combining magnetic resonance fingerprinting with voxel-based morphometric analysis to reduce false positives for focal cortical dysplasia detection. Epilepsia. 2024;65(6):1631-1643.

5. Tang Y, Su TY, Choi JY, et al. Characterizing thalamic and basal ganglia nuclei in medically intractable focal epilepsy by MR fingerprinting. Epilepsia. 2022;63(8):1998-2010.

6. Choi JY, Hu S, Su T yu, et al. Normative quantitative relaxation atlases for characterization of cortical regions using magnetic resonance fingerprinting. Cereb Cortex. 2023;33(7):3562-3574.