The Impact and Reliability of Tissue Segmentation on In Vivo Magnetic Resonance Spectroscopy Metabolite
Jessica Archibald1, Kay Chioma Igwe2, Antonia Kaiser3
1Department of Radiology, Weill Cornell Medicine, New York, New York, USA.
Magnetic Resonance in Medicine
|April 10, 2026
Summary
Brain tissue segmentation significantly impacts magnetic resonance spectroscopy (MRS) metabolite quantification. Using different segmentation tools introduces substantial variability, affecting metabolite concentration estimates and reproducibility in MRS research.
Area of Science:
- Neuroimaging
- Biomedical Engineering
- Metabolomics
Background:
- Metabolite quantification using Magnetic Resonance Spectroscopy (MRS) relies on accurate tissue composition estimation within voxels.
- Variability in MRS-derived concentration estimates can arise from differences in brain tissue segmentation algorithms.
Purpose of the Study:
- To investigate the impact and reliability of different brain tissue segmentation tools on metabolite quantification using MRS.
- To assess how segmentation variability influences the accuracy and reproducibility of metabolite concentration estimates.
Main Methods:
- Three segmentation tools (ANTs, FSL, SPM) were evaluated on an in vivo test-retest MRI/MRS dataset.
- Voxelwise gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) fractions were used to calculate tissue-corrected total creatine (tCr) concentrations.
- Linear mixed-effects modeling and variance-component partitioning quantified tool-, session-, and participant-related variance, with comparisons to manual segmentation.
Main Results:
- Systematically different tissue fractions from segmentation tools led to variations in tCr concentration estimates.
- Segmentation tool contributed significantly to total tCr variability (50-56.8%), while participant factors accounted for 28.5-36.2%.
- Test-retest reliability decreased from high (ICC > 0.8) to moderate (~0.5) when both segmentation and MRS variability were considered.
Conclusions:
- Brain tissue segmentation is a major source of variability in MRS metabolite concentration estimates.
- Transparent reporting of segmentation methods and data sharing are crucial for enhancing reproducibility and cross-study comparability in MRS research.
