Abstracts

Rationale: Neuroimaging of focal neuroinflammation (NI) with MRI is at present limited to contrast-enhanced techniques that target the blood-brain-barrier. However, brain temperature measurements with volumetric magnetic resonance spectroscopic imaging (MRSI-t) may be used as a proxy for NI and reflect focal NI in epilepsy. Whole-brain MRSI-t noninvasively calculates voxel-level brain temperature from the chemical shift difference between water and creatine (T_CRE=−102.61×Δ_H20-CRE+206.1^oC). In healthy controls, temperature measurements (T_CRE) with MRSI-t have been shown to be location-specific and stable over at least 12 weeks (Sharma et al., 2020). Our objective was to investigate whether T_CRE elevations in patients with TLE are localized to clinically relevant brain regions.

Methods: 19 healthy controls (HC, age 22-52) and 19 TLE patients (age 20-71) were scanned on a 3T Siemens Magnetom Prisma using a 20-channel head coil. MRSI-t scans were collected using a 3D echo planar spectroscopic imaging sequence. Structural images were acquired to co-register and normalize MRSI-t data to standard space. Image reconstruction, spectral processing, and temperature mapping were completed within the Metabolite Imaging and Data Analysis System. Mean T_CRE was estimated and analyzed for AAL-defined regions of interest (ROIs). For left TLE, data were flipped such that all putative ictal onset zones were R-sided (data on matched HCs were also flipped due to existing slightly asymmetric MRSI-t T_CRE derivations). Voxelwise independent samples t-tests computed group T_CRE differences. Statistical models were assessed with permutation-based testing using the Threshold-Free Cluster Enhancement toolbox, p < 0.05 corrected for multiple comparisons using family-wise error.

Results: TLEs’ T_CRE maps showed statistically significant T_CRE increases in two large clusters in the R middle and inferior TL (Fig1A – close to the putative ictal onset zone). Analyses of mean T_CRE in 47 ROIs showed a significant main effect for group F(1, 1692)=36.06, p < 0.0001 and for brain region F(46, 1692)=2.60, p < 0.0001. The group*ROI interaction was not significant, indicating group had the same effect across all brain regions. TLEs had an overall higher global T_CRE (M=37.3±0.4) than HCs (M=37±SD=0.3) (Fig1B) and higher focal T_CRE in the hypothesized R hippocampus (p=0.04, CI 0.03-1.45). TLEs’ T_CRE was also significantly higher in small clusters in the L/R calcarine, L caudate, L parietal and L postcentral ROIs (all p’s < 0.05) by ~1^oC.

Conclusions: This study indicates that brain temperatures measured with standard MRI equipment may be specific to seizure-generating brain regions in TLE. With the already established high stability/repeatability of MRSI-t based brain temperature in HCs, our findings indicate MRSI-t could be used to measure brain temperature as an aid for localizing the ictal onset zone and potentially for tracking disease progression. Correlations with other measures of focal NI are needed.

Funding: Please list any funding that was received in support of this abstract.: This was supported by General Funds from the State of AL (Carly’s Law), AMC21 funds from UAB, an NIH/NINDS institutional training grant (5T32NS061788-13), and funds from UAB Division of Epilepsy.