Quantitative MRI Reveals Microstructural Changes in the Upper Leg Muscles After Running a Marathon

Abstract

Background: The majority of sports-related injuries involve skeletal muscle. Unlike acute trauma, which is often caused by a single traumatic event leading to acute symptoms, exercise-induced microtrauma may remain subclinical and difficult to detect. Therefore, novel methods to detect and localize subclinical exercise-induced muscle microtrauma are desirable.

Purpose: To assess acute and delayed microstructural changes in upper leg muscles with multiparametric quantitative MRI after running a marathon.

Study type: Longitudinal; 1-week prior, 24-48 hours postmarathon and 2-week follow-up POPULATION: Eleven men participants (age: 47-68 years).

Field strength/sequence: Spin-echo echo planar imaging (SE-EPI) with diffusion weighting, multispin echo, Dixon, and fat-suppressed turbo spin-echo (TSE) sequences at 3T. MR datasets and creatine kinase (CK) concentrations were obtained at three timepoints.

Assessment: Diffusion parameters, perfusion fractions, and quantitative (q)T₂ values were determined for hamstring and quadriceps muscles, TSE images were scored for acute injury. The vastus medialis and biceps femoris long head muscles were divided and analyzed in five segments to assess local damage.

Statistical tests: Differences between timepoints in MR parameters were assessed with a multilevel linear mixed model and in CK concentrations with a Friedman test. Mean diffusivity (MD) and qT₂ for whole muscle and muscle segments were compared using a multivariate analysis of covariance (MANCOVA).

Results: CK concentrations were elevated (1194 U/L [166-3906], P < 0.001) at 24-48 hours postmarathon and returned to premarathon values (323 U/L [56-2216]) at 2-week follow-up. Most of the MRI diffusion indices in muscles without acute injury changed at 24-48 hours postmarathon and returned to premarathon values at follow-up (MD, RD, and λ3; P < 0.006). qT₂ values (P = 0.003) and perfusion fractions (P = 0.003) were higher at baseline compared to follow-up. Local assessments of MD and qT₂ revealed more pronounced changes than whole muscle assessment (2-3-fold; P < 0.01).

Data conclusion: Marathon running-induced microtrauma was detected with MRI in individual whole upper leg muscles and even more pronounced on local segments.

Level of evidence: 2 TECHNICAL EFFICACY STAGE: 3 J. Magn. Reson. Imaging 2020;52:407-417.

Keywords: DTI; inflammation; microstructure; muscle injury; perfusion.

Figure 1

Study design. (a) Timeline indicating marathon and timepoints of MRI and blood sampling. (b) Schematic view of the scan geometry with three stacks of 31 slices and 30 mm section overlap covering the upper legs. Bottom right: Dixon images at mid‐thigh level with manually drawn ROIs for eight upper leg muscles: 1 = BFS, 2 = BFL, 3 = ST, 4 = SM, 5 = VM, 6 = VI, 7 = VL, 8 = RF.

Figure 2

Boxplots showing the CK concentrations (U/L) in the venous blood for all participants at baseline, postmarathon, and follow‐up (a) showing all participants and (b) with exclusion of participants with overt injuries. Significant differences are indicated with an asterisk (*).

Figure 3

Raw and processed example images of a representative subject (47 years/male) for the three timepoints. Left column: baseline, middle column: postmarathon and right column: follow‐up. Top: water images from Dixon sequence. Middle: raw diffusion images for b = 0 s/mm² and MD maps. Bottom: raw images from first TE of T₂ sequence and processed EPG T₂ maps.

Figure 4

Time courses of FA, MD, λ3, qT₂, and perfusion fraction for all upper leg muscles separately (left and right leg taken together; left column) and for all subjects separately for the BFL and VM muscle. Each muscle (left column) or subject (middle and right column) is shown in gray, while the mean and standard deviation are shown in black. Asterisks indicate an overall time‐effect in the left column and a significant change between timepoints in the middle and right column. The timepoints are defined as 0 = baseline, 1 = postmarathon, and 2 = follow‐up.

Figure 5

Comparisons between whole muscle volume and local measurements. Left: boxplots showing MD and qT₂ for the BFL (top) and VM (bottom) muscle for all subjects based on a whole volume assessment. Middle: line graphs of MD and qT₂ for five segments of the BFL and VM muscle averaged over all subjects at baseline (black), postmarathon (red), and follow‐up (gray). Asterisks indicate significant time‐effects between baseline and postmarathon. Right: showing the location of the five muscle segments for the BFL and VM muscle. Note that the difference in MD and qT₂ values between timepoints is more pronounced for the localized measurements, indicating the higher sensitivity to detect changes.

Figure 6

Comparisons between whole muscle volume and local measurements in percentage change. Scatterplots showing mean diffusivity (MD) (top) and qT₂ (bottom) for the BFL (left) and VM (right) muscle for all subjects based on a whole volume assessment and the five local measurements. Relative change was determined per subject between baseline and postmarathon measurements. Each gray dot reflects the relative change between baseline and postmarathon for an individual subject, with the group mean and standard deviation shown in black. The red dotted line represents the average Δ% for the whole muscle volume measurements between baseline and postmarathon.