Impacts of retrospective lipid suppression on metabolite quantification in preclinical proton MR spectroscopic imaging

Abstract

Extracranial lipid contamination remains a challenge in proton magnetic resonance spectroscopic imaging (MRSI), especially in short acquisition delay MRSI, where broad lipid resonances overlap with metabolite and macromolecular signals. Although retrospective lipid suppression techniques are widely used in human MRSI, their effects on metabolite quantification in preclinical MRSI, which is more prone to lipid contamination, have not yet been examined. In this study, we assessed how the retrospective lipid suppression and spectral fitting range influence spectral quality, spatial metabolite mapping, and quantification variability using proton MRSI of rat brains at 14.1 T. Lipid suppression was applied via an orthogonal projection method to both fully sampled and compressed sensing datasets, each comprising data with minimal and pronounced lipid contamination. Spectral fitting was performed with both broad (4.1 - 0.2 ppm) and narrow (4.1 - 1.8 ppm) ranges. When lipid contamination was minimal, suppression caused only slight spectral and spatial changes, with consistent metabolite quantification across conditions. In contrast, datasets with pronounced lipid contamination exhibited notable spectral changes following suppression, consequently affecting spatial metabolite mapping and concentration estimates. Group analysis revealed that metabolites with low concentration estimates were most affected. Similar effects were observed in compressed sensing datasets. Our results provide a better understanding of the impact of retrospective lipid suppression on metabolite quantification in preclinical MRSI, thereby supporting future optimizations for its effective application.