Diffusion tensor imaging is increasingly popular for studying micro-architectural characteristics of human skeletal muscle associated with ageing and pathology¹. Mean diffusivity (MD), and fractional anisotropy (FA) are parameters most often considered in these studies. Previous work has reported changes in DTI-derived parameters in calf muscle with aging²; however, limited attention was given to the influence of lipid signals, which are known to have substantial impact on DTI analysis³. In this study, we investigate the influence of ageing on DTI measures using a robust protocol with improved fat-suppression applied to the thigh quadriceps muscle.Imaging was performed using a Philips Achieva 3T system (Philips Healthcare, Best, NL) with a 32-channel cardiac coil. Fifteen participants (11 male, median age=51, range 27-73yrs) were recruited as part of the Genetic and Epigenetic Signatures of Translational Aging Laboratory Testing (GESTALT) study. Participants were positioned feet-first in the scanner, with their legs aligned with the bore and the left thigh close to isocentre. A perfluorocarbon ‘SatPad’ (MRIequip.com, Brainerd, USA) was packed around the thigh to mitigate B₀-related image distortions. After localisers and second-order shimming, a multiecho two-point Dixon sequence was applied as a high-resolution anatomical reference: TR=5.8ms, TE=1.4 and 2.6ms, flip angle=6°, field-of-view=256mm×228mm, in-plane resolution=1mm×1mm, 60 slices, slice thickness=3mm, and sensitivity-encoding (SENSE) factor=2. This scan precisely overlapped the DTI stack, which was planned with its distal edge at the insertion of the vastus intermedius to sample the majority of the quadriceps muscle. DTI parameters were as follows: spin echo single-shot echo-planar imaging; TR/TE=3500/33ms; field-of-view = 256×225mm; 30 slices; slice thickness=6mm; in-plane resolution=2.56mm×2.61mm; 8 NSA; partial Fourier=0.6 in ky; SENSE factor=2; combined spectral adiabatic inversion recovery (TI/offset=70ms/250Hz) and slice-select gradient reversal for fat-suppression; and 15 diffusion gradient directions, with b=0, 450ms, δ=27ms, and Δ=35ms. DTI data were post-processed using FSL (FMRIB, University of Oxford, UK), where they were distortion- and eddy-current-corrected, and registered to anatomical images. Data were then exported to DSI-Studio (Fang-Cheng Yeh, Carnegie Mellon University, USA) where MD and FA were calculated. Regions-of-interest (ROIs) were drawn across each quadriceps head: rectus femoris (RF), vastus lateralis (VL), vastus medialis (VM), and vastus intermedius (VI). These ROIs were eroded by two pixels, to avoid partial-volume, and the mean and standard deviation of MD and FA were calculated for each. Both parameters were plotted against age, and data were tested for normality, after which Pearson’s test for correlation was applied. Subsequently, the DTI datasets were used to generate qualitative fibre tractography plots of the quadriceps, using fibre-tracking parameters as follows: 2mm seed-spacing, 0.2mm step-length, FA lower/upper threshold=0.1/0.5, and max. angle change=10°. Figures 1-4 show plots of FA and MD versus age for each of the quadriceps heads: RF, VL, VM, and VI, respectively. Trendlines are shown for selected data, with anatomical images to visualise ROI-positioning. These preliminary results show that FA in each quadriceps head tends to increase with age, agreeing well with previous studies^2,4, which attribute this to reduced fibre diameter and changes in the proportion of Type I to Type II fibres. The relationship is modest, with Pearson R² values from 0.17-0.27. The largest of these correlations indicated a strong, statistically-significant relationship between rectus femoris FA and age, with R²=0.27 (p=0.04), while FA in the other muscle heads approached significant correlations. In the whole quadriceps muscle, MD appeared to have a more complicated relationship with age, if any. This contrasts with the findings of Galbán et al., who indicated a pronounced decrease in calf muscle diffusion coefficients by up to 10% in older participants². We attribute this discrepancy to the improved fat-suppression used in the current work compared with that of Galbán et al., and the observed increase in adipose tissue with ageing⁵. Using simulated diffusion signals, with and without lipid components, we observed a negative bias error in MD estimates on the order of 10% with the addition of only 5% fat signal (data not shown), reinforcing the confounding influence of body composition described by Damon³. Fig. 5 shows example fibre tractography from one participant; the tracts are anatomically plausible, and well-suited to quantitative analysis of fibre pennation angle and curvature.We have demonstrated the requirement for high-quality fat-suppression to properly characterise MD in the context of ageing. With this, we have found a consistent increase in FA with respect to age across each of the quadriceps heads. Additional correlative data available within the GESTALT database will enable us to extend these results to establish relationships between FA and MD, and tractography measures, with measures of muscle quality and function.