ID #592 Spinal cord tumour Quantitative Imaging with Super High-field strength magnetic resonance spectroscopy for accurate non-invasive diagnosis in Youngsters (SQUISHY spine project)
Reeya Patel, Angela Walls, Amanda Luck, Hannah Rose Wardill, Jordan Hansford, Ryan O’Hare DoigAbstract
Background
Quantitative magnetic resonance imaging (qMRI) techniques, including diffusion tensor imaging (DTI), functional MRI (fMRI), and proton MR spectroscopy (1H-MRS), offer significant potential to improve paediatric spinal cord tumour (SCT) diagnosis. However, our scoping review shows that ∼99% of paediatric SCT studies rely on conventional MRI, with minimal use of qMRI and no reported application of MRS. The small size and motion sensitivity of the paediatric spinal cord limit resolution and acquisition stability, contributing to the lack of validated imaging biomarkers. We have recently established cervical and thoracolumbar DTI and fMRI pipelines in clinical cohorts, demonstrating the feasibility of advanced spinal cord qMRI. The next phase is to develop and optimise a paediatric-adapted1H-MRS pipeline for SCT characterisation.
Methods
A paediatric specific MRS protocol is being developed on a Siemens MAGNETOM Cima.X 3T system, focusing on single-voxel PRESS/STEAM acquisitions in cervical and thoracolumbar cord segments. Parameters under optimisation include voxel dimensions, flip angles, TR/TE settings, shim optimisation strategies, and physiological-motion compensation. In parallel, to identify novel metabolites suitable for targeted in vivo MRS, SCT samples from paediatric and adult donors will undergo spatial metabolomics, using matrix-assisted laser desorption/ionisation (MALDI) imaging. Metabolite profiles will be quantified to compare biomarker abundance across tumour types and assess concordance between MALDI-derived and MRS-detectable signatures.
Results
Findings will inform construction of a metabolic phantom for systematic MRS optimisation. To date, multiple pilot scans across varying cord sizes and acquisition parameters, demonstrate stable signal, improved contrast, and early feasibility of achieving reliable spectra in small-calibre spinal cord regions. Full optimisation and validation are ongoing.
Conclusion
Our work outlines a translational pipeline integrating qMRI, spatial-omics, and phantom-based optimisation to develop a clinically deployable paediatric MRS protocol for SCT diagnosis and stratification, addressing a key gap in paediatric neuro-oncology.