MR Fingerprinting Added Diagnostic Value in 46% of 100 Presurgical Epilepsy Cases

Magnetic resonance fingerprinting (MRF) delivered additional radiologic findings beyond conventional 3T MRI in 46% of 100 patients undergoing presurgical epilepsy evaluation at Cleveland Clinic, according to a retrospective study published in Epilepsia. The cohort, reviewed between 2017 and 2024, represents the largest clinical evaluation of MRF in epilepsy surgery candidates to date.

"We conducted the first large-scale study of MRF in patients undergoing evaluation for epilepsy surgery, analyzing 100 cases, the largest cohort to date," says senior author Irene Wang, PhD, Research Director of the Cleveland Clinic Epilepsy Center. "We found that MRF improved the ability to detect and clearly define subtle brain abnormalities, particularly in challenging epilepsy subtypes. These findings lay the foundation for integrating MRF into routine clinical care."

What the protocol delivered

The investigators compared standard 3T MRI clinical protocols with 10-minute, high-resolution 3D MRF acquisitions that simultaneously generated T1, T2 and gray/white matter fraction maps. Unlike conventional T1-weighted or FLAIR sequences, which rely on relative signal intensities vulnerable to scanner hardware, acquisition settings, patient positioning and field inhomogeneity artifacts, MRF produces absolute quantitative measures of intrinsic tissue properties by matching acquired signal patterns against a dictionary of simulated tissue signatures.

Among the 32 patients whose initial conventional MRI reports were read as negative, MRF identified abnormalities or supplied new insights in 38% of cases — a directly relevant figure for surgical conferences where a "nonlesional" designation often limits confidence in seizure-onzone localization and worsens expected surgical outcomes.

The added value broke down into four categories: enhanced lesion conspicuity (n = 16), better visualization of malformation boundaries for resection planning (n = 11), detection of mesial temporal asymmetry through more uniform left-right hippocampal and amygdaloid comparisons with fewer bias-field artifacts (n = 9), and differentiation of which of multiple lesions — as in tuberous sclerosis or heterotopia — was most relevant to the seizure onset zone (n = 5).

Subtype-specific yield

The diagnostic impact was most striking in histologic subgroups where conventional MRI struggles. MRF added value in 78% of focal cortical dysplasia (FCD) type IIa cases, compared with only 17% of type IIb cases — consistent with type IIb's typical visibility on standard sequences and type IIa's subtlety. For periventricular nodular heterotopia, MRF provided additional information in 88% of cases.

"MRF proved especially valuable in defining subtle brain abnormalities, particularly in conditions such as periventricular nodular heterotopia and FCD type IIa, which are often difficult to identify but critically important for surgical planning," says first author Maksim Parfyonov, MD, a staff pediatric epileptologist in the Epilepsy Center.

Two illustrative cases anchor the findings. In a 30-year-old woman with left mesial temporal epilepsy whose conventional MRI was completely nonlesional, SEEG demonstrated left mesial temporal onset, and MRF T1 maps showed increased signal in the same region; she became seizure-free after SEEG-guided surgery. In a 27-year-old woman with mild malformation of cortical development with oligodendroglial hyperplasia and epilepsy (MOGHE), whose conventional MRI was initially read as negative, MRF T1 and white-matter fraction maps better delineated lesion extent and abnormal white-matter architecture; she too became seizure-free after resection.

Practical considerations

The biophysical basis explains much of the performance. Hypomyelination, common in epileptogenic lesions, lengthens T1 relaxation times — and T1 maps indeed outperformed T2 maps in this cohort, providing superior diagnostic utility in 32 patients versus nine for T2. The 3D protocol also incorporates integrated B1 mapping to correct field inhomogeneities, particularly in the inferior temporal lobes where many seizure foci arise, yielding "flatter" images in which a regional signal change more reliably reflects true tissue difference.

For clinical workflow, the researchers reviewed the quantitative maps directly in their standard PACS, using custom colorized scales to maximize dynamic range and reveal subtle features such as gray-white matter junction blurring.

Limitations remain. Reconstructing MRF maps demands substantial computational power, large data dictionaries and specialized expertise. Current MRF sequences lack the fluid suppression of FLAIR, and skull signal variability can complicate assessment of lesions near the brain-bone interface. With recent FDA clearance of MRF for initial implementation, however, the authors argue the path to routine clinical use is narrowing — with the goal, as Dr. Wang puts it, of "improving diagnostic accuracy, guiding more precise surgical treatment and ultimately enhancing outcomes for patients with epilepsy."