A comprehensive guide to MEGA-PRESS for GABA measurement

Abstract

The aim of this guideline is to provide a series of evidence-based recommendations that allow those new to using MEGA-PRESS to produce high-quality data for the measurement of GABA levels using edited magnetic resonance spectroscopy with the MEGA-PRESS sequence at 3T. GABA is the main inhibitory neurotransmitter of the central nervous system and has been increasingly studied due to its relevance in many clinical disorders of the central nervous system. MEGA-PRESS is the most widely used method for quantification of GABA at 3T, but is technically challenging and operates at a low signal-to-noise ratio. Therefore, the acquisition of high-quality MRS data relies on avoiding numerous pitfalls and observing important caveats. The guideline was developed by a working party that consisted of experts in MRS and experts in guideline development and implementation, together with key stakeholders. Strictly following a translational framework, we first identified evidence using a systematically conducted scoping literature review, then synthesized and graded the quality of evidence that formed recommendations. These recommendations were then sent to a panel of 21 world leaders in MRS for feedback and approval using a modified-Delphi process across two rounds. The final guideline consists of 23 recommendations across six domains essential for GABA MRS acquisition (Parameters, Practicalities, Data acquisition, Confounders, Quality/reporting, Post-processing). Overall, 78% of recommendations were formed from high-quality evidence, and 91% received agreement from over 80% of the expert panel. These 23 expert-reviewed recommendations and accompanying extended documentation form a readily useable guideline to allow those new to using MEGA-PRESS to design appropriate MEGA-PRESS study protocols and generate high-quality data.

Keywords: 1H-MRS; Edited-MRS; GABA; Guideline; MEGA-PRESS; MRS.

Graphical abstract

Fig. 1

Demonstrating the process followed to develop the guideline based on the ADAPTE process [32].

Fig. 2

Results from Round 1 of the Expert panel review, results displayed from highest to the lowest level of agreement and scaled from 50% to 100% agreement.

Fig. 3

Results from Round 2 of the Expert panel review, results displayed from highest to the lowest level of agreement and scaled from 50% to 100% agreement.

Fig. 4

Sample MEGA-PRESS data, and common quality issues that can occur, green demonstrating higher quality data, orange medium and red low. a) High-quality data with sufficient SNR, narrow linewidths, a well-defined edited signal at 3 ppm, and no substantial artefacts; b) very high noise levels due to low number of transients or small voxel volume; c) poor shim resulting in poor spectral resolution and lower SNR; d) severe subtraction artefacts due to scanner frequency drift; d) lipid contamination due to participant motion or voxel positioning too close to the skull; e) out-of-voxel echo (“ghost signal”); g) sensorimotor data and voxel location, usually a region that is easy to shim, but risks lipid contamination as demonstrated here; h) thalamus data and voxel location usually lower SNR, linewidths often greater due to iron deposition in deep regions; i) hippocampus/medial-temporal lobe data and voxel location, a region that is difficult to shim and prone to artefacts.

Multimedia component 5