3232

Mapping Metabolite and Neurotransmitter Alterations in Mesial Temporal Lobe Epilepsy Using Optimized Multi-TE MRSI: A Preliminary Study

Zepeng Wang^1,2, Ruiyang Zhao^2,3, Xinyu Li^1,2, Aaron Anderson^2,4, Graham Huesmann^4,5, and Fan Lam^1,2,3
¹Department of Bioengineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, ²Beckman Institute for Advanced Science and Technology, University of Illinois at Urbana-Champaign, Urbana, IL, United States, ³Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States, ⁴Neuroscience Institute, Carle Foundation Hospital, Urbana, IL, United States, ⁵School of Molecular and Cellular Biology, University of Illinois at Urbana-Champaign, Urbana, IL, United States

Synopsis

Keywords: Epilepsy, Epilepsy, Spectroscopic imaging, mTLE, multi-TE MRSI

Motivation: Multimodal imaging integrating molecular information promises to improve seizure foci lateralization and disease progression monitoring for mesial temporal lobe epilepsy (mTLE).

Goal(s): To evaluate the feasibility of a multi-TE MRSI method optimized for simultaneously mapping metabolites and neurotransmitters in mTLE.

Approach: Volumetric multi-TE MRSI and structural MRI data were acquired and analyzed for an mTLE cohort. Regional analysis was performed to quantified molecular maps and hippocampus volumetrics.

Results: Reduced NAA/Cr ratios were observed for the ipsilateral hippocampous (epileptic side), consistent with prior findings and overall reduced hippocampus volumes. Reduced Glx/Cr ipsilaterally was observed. Interesting differences in anterior vs. posterior hippocampus were also reported.

Impact: We demonstrated the feasibility of a new MRSI technique for mapping metabolic alterations in mTLE patients. Initial results support promise for improved lateralization and better understanding biochemical mechanisms of the disease using MRSI.

Introduction

Mesial Temporal Lobe Epilepsy (mTLE) is the most common form of pharmacologically refractory epilepsy^1,2, which is frequently associated with progressive mesial temporal sclerosis (MTS) in the hippocampus and eventually leads to bilateral seizure onset^3,4. Although clinical MRI plays a critical role in pre-surgical evaluation for the lateralizing epileptogenic zone for dissection and post-treatment monitoring of mTLE, using anatomical features alone led to around 33% mis/false-identification of lesions⁵. Metabolic changes^6-8 and tissue stiffness⁴ have demonstrated potential in improving epileptic foci localization. The recent advance in high-resolution MRSI⁶ also provides opportunities to study multi-regional changes in limbic networks. But molecular information provided by MRSI has mostly been limited to major metabolites such as NAA, Cr and Cho. There is an increasing interest in measuring neurotransmitter components, e.g., Glx vs. GABA to investigate excitatory/inhibitory imbalance and their connections with the progressive pathology^9,10. In this study, we evaluated the feasibility of a recently developed 3D multi-TE MRSI method for simultaneously mapping metabolites and neurotransmitters in mTLE brains. While the study is ongoing, initial results from a small cohort showed promising results, e.g., reduced NAA/Cr ratios in the ipsilateral hippocampus consistent with prior findings and reduced Glx/Cr in the sclerosis side. Interesting differences in anterior vs. posterior hippocampus are also reported.

Methods

Patient Recruitment and Demographics:
11 patients with unilateral moderate to severe mTLE participated in this study (Table. 1). Clinical inclusion criteria were based on clear clinical diagnosis of mTLE by clinical epileptologists. mTLE severity was determined based on the clinical findings on a combination of clinical presentation, duration of seizures, EEG, T1w and T2w MRI, and/or PET. EEG findings of frontotemporal slowing and sharp waves, MRI evidence of hippocampal volume and internal architecture loss, and hypometabolism on PET.
High-resolution Multi-TE MRSI:
All patient data were collected on a Siemens Prisma 3T system with local IRB approval. The MRI/MRSI protocol included T1w-MPRAGE (1.0×1.0×1.0mm³), T2 FLAIR (0.9×0.9×5mm³), a 2-TE MRSI scan with FOV = 240×220×64 mm³, 3.8×3.4×6.4 mm³ voxel size, TR/TE = 1100/[65, 80]ms optimized for metabolite, Glx and GABA quantification¹² and a B₀ field map for MRSI processing (2.2×2.2×2.0 mm³). Additional sets of lower-resolution MRSI data were acquired (6.8×6.9×8.0mm³) for learning population-specific subspace for spatiospectral reconstruction. Subspace-based processing steps, i.e., nuisance water/lipid removal, augmented multi-TE subspace-based spatiospectral reconstruction¹¹, and a multi-step metabolite quantification strategy were performed¹².
Data Analysis:
ROI masks of the hippocampus and its subdivisions (i.e., anterior and posterior) from both ipsilateral (sclerosis side) and contralateral hippocampus were generated using FreeSurfer from T1w-MPRAGE and co-registered to MRSI volume¹³. Gray-matter-dominant voxels (GM ratio >= 0.6) were selected from the quantified metabolite maps for regional and comparative analysis. Hippocampus volumetrics were also performed (Fig. 2). Wilcoxon matched-pairs signed rank test was conducted to compare metabolic changes in the ipsilateral and contralateral hippocampus.

Results and Discussion

The multimodal and multiplexed molecular imaging capabilities offered by our data are illustrated for a representative patient in Fig. 3. Hippocampus sclerosis was identified by the hyperintensity in the FLAIR image. Metabolic alterations, both in individual metabolite maps and ratios, can be observed in the ipsilateral side (sclerosis/epileptic side), within a large region of interest. Metabolite levels and their ratios in the hippocampus and its anterior and posterior subdivisions were compared between the ipsilateral and contralateral sides across all patients (Fig. 4). Reduced NAA/Cr ratios at the ipsilateral side can be observed, which is consistent with previous findings^6-10. Significantly reduced Glx/Cr in the ipsilateral hippocampus was also observed, which was rarely captured in previous studies, indicating neurotransmitter alterations related to epileptic regions. We also further stratify the patients based on severity levels assessed by our collaborating epileptologist (based on EEG findings, clinical MRI, and self-reported and clinical memory deficits). Differential metabolite changes were found in the more severe cases (Fig. 5). These observations need to be further validated after most patients are scanned and analyzed.

Conclusion

We demonstrated the feasibility of a multi-TE MRSI method optimized for simultaneous metabolite and neurotransmitter mapping in mTLE patients. Initial results from a small cohort show consistency with prior findings as well as interesting new observations. A more comprehensive analysis will be performed as more patients are recruited and scanned in this ongoing project.

Acknowledgements

This work was supported in part by NSF-CBET 1944249 and NIH-NIBIB R21EB029076.

References

[1] Semah F et al. Is the underlying cause of epilepsy a major prognostic factor for recurrence? Neurology.1998; 51:1256-62.

[2] Engel Jr J. Mesial temporal lobe epilepsy: what have we learned? The Neuroscientist. 2001 ;7:340-52.

[3] Coan AC et al. Epilepsy as progressive disorders: what is the evidence that can guide our clinical decisions and how can neuroimaging help? Epilepsy & Behavior. 2013; 26:313-21.

[4] Huesmann GR et al. Hippocampal stiffness in mesial temporal lobe epilepsy measured with MR elastography: Preliminary comparison with healthy participants. NeuroImage: Clinical. 2020; 27:102313.

[5] Muhlhofer W et al. MRI‐negative temporal lobe epilepsy—What do we know? Epilepsia. 2017; 58:727-42.

[6] Huang H et al. Simultaneous high-resolution whole-brain MR spectroscopy and [18F] FDG PET for temporal lobe epilepsy. European Journal of Nuclear Medicine and Molecular Imaging. 2023; 12:1-3.

[7] Capizzano AA et al. Multisection proton MR spectroscopy for mesial temporal lobe epilepsy. American Journal of Neuroradiology. 2002; 23:1359-68.

[8] Pan JW et al. Utility of magnetic resonance spectroscopic imaging for human epilepsy. Quantitative Imaging in Medicine and Surgery. 2015; 5:313.

[9] Simister RJ et al. A short‐echo‐time proton magnetic resonance spectroscopic imaging study of temporal lobe epilepsy. Epilepsia. 2002; 43:1021-31.

[10] Simister RJ et al. Proton MR spectroscopy of metabolite concentrations in temporal lobe epilepsy and effect of temporal lobe resection. Epilepsy Research. 2009; 83:168-76.

[11] Wang Z et al. High‐resolution, 3D multi‐TE 1H-MRSI using fast spatiospectral encoding and subspace imaging. Magnetic Resonance in Medicine. 2022; 87:1103-18.

[12] Wang Z et al. Optimized subspace-based J-resolved MRSI for simultaneous metabolite and neurotransmitter mapping. In Proc. of ISMRM, 2021, p. 72.

[13] Xie L et al. Accounting for the confound of meninges in segmenting entorhinal and perirhinal cortices in T1-weighted MRI. MICCAI, 2016. pp. 564-571.

Figures

Table 1: Demographics for the recruited mTLE patients who completed the MRI scans. Clinical MRI findings are observed in most patients. Laterality was determined based on EEG findings (not reported here), MRI findings, and Neuropsychological assessments. A severity assessment was provided by a collaborating epileptologist.

Figure 2: Summary of hippocampus and its subdivision's (anterior and posterior) volumes. Overall volume reduction at the ipsilateral side was observed with larger changes in the anterior part. Figure arrangement: ratio of volume loss ipsilateral to contralateral hippocampi (left), within hippocampi ratio of anterior to posterior (center), and across hippocampi for anterior and posterior (right). Subdivisions were generated using the ASHS atlas.

Figure 3: Multimodal imaging results from a representative patient: Hyper-intensity in the right hippocampus was observed in the T₂ FLAIR image on the left (red arrow). Reduction of NAA in the right hippocampus can be observed, as shown in the quantified metabolite maps (second column) and regional-averaged spectra (bottom right). Metabolite ratio distributions in the hippocampus region are shown on the right (top), exhibiting reduction of NAA/Cr and Glx/Cr on the ipsilateral side.

Figure 4: Comparisons of NAA/Cr (top) and Glx/Cr (bottom) ratios between the ipsilateral and contralateral hippocampi (column 1), in the anterior hippocampus specific region (column 2) and in the posterior hippocampus specific region (column 3), for all patients. Reduced NAA/Cr and Glx/Cr in ipsilateral side can be observed in the overall hippocampus, which seems to be driven mostly by the differences in the posterior regions (different from the volumetrics in Fig. 2). Wilcoxon signed-rank tests were performed, *, ** and *** represent p < 0.1, p<0.05 and p<0.01 respectively.

Figure 5: A more detailed look at the metabolite alterations for patients with different severity (M: moderate and minimal, S: severe; see texts for more information). Reduction of NAA/Cr was observed in the ipsilateral side in both cases, consistent with Fig. 4. Interestingly, increased NAA/Cr was observed in the ipsilateral side for the more severe cases (top left figure), which was mainly due to reduction of Cr (bottom right figure). Glx/Cr differences between moderate and severe cases were also shown (top right figure).

Proc. Intl. Soc. Mag. Reson. Med. 32 (2024)

3232

DOI: https://doi.org/10.58530/2024/3232