Variability of Water T2 Times in Quantitative Muscle MRI: Considerations and Methodological Factors—A Narrative Review

Abstract

Quantitative muscle imaging (qMRI) is an established method for detecting muscular changes, particularly in the diagnosis and follow-up of conditions as muscular injuries and neuromuscular diseases. While global T2 reflects all tissue components and cannot reliably indicate disease activity in fat-replaced muscles, water T2 times (wT2) can serve as a biomarker for inflammatory processes and intramuscular oedema and provide valuable information for disease activity. Proton-MR-spectroscopy (¹H-MRS) is considered the reference standard for accurate wT2 measurements. However, its limited spatial coverage makes it unsuitable for whole-leg or whole-body analyses. T2-mapping offers an alternative by extracting wT2-values over larger regions. While monocentric assessment of individual changes is well established, comparing results across studies is difficult because of methodological differences. This narrative review summarizes 54 studies published between 1991 and 2025 reporting on wT2-values in healthy skeletal muscle. It evaluates the measurement methods used and their impact on measured values. This includes the use of ¹H-MRS as a reference, sequence parameters, fitting models and their correction methods, and inter-vendor variations. Finally, it gives recommendations for future studies investigating wT2 in muscle tissue. Since healthy muscle tissue is generally consistent across populations, similar wT2-values are expected between vendors. Compared with the ¹H-MRS ground truth of approximately 28.1 ms, reported wT2-values in literature range from 22.9 to 43.0 ms at 3 T (45 studies) and 27.6 to 41.9 ms at 1.5 T (nine studies), far exceeding the pathological sensitivity threshold of 1–2 ms. This review shows that, even with comparable methods and confounder correction, the observed differences in wT2-values exceed the pathological range. The review recommends including control groups and extended-phase-graph–based fitting with fat correction to improve comparability.