Demonstration of a Sliding-Window Diffusion Tensor Technique for Temporal Study of Post-Exercise Skeletal Muscle Dynamics
Conrad P Rockel1,2 and Michael D Noseworthy1,2,3

1School of Biomedical Engineering, McMaster University, Hamilton, ON, Canada, 2Imaging Research Centre, St Josephs Healthcare, Hamilton, ON, Canada, 3Electrical and Computer Engineering, McMaster University, Hamilton, ON, Canada

Synopsis

A novel sliding-window DTI analysis strategy, aimed at achieving both temporal resolution and valid spatial representation, was tested on one human subject pre- and post-exercise (plantar flexion) across 4 sets of different intensity. Temporal diffusion measures comprised of 3- and 15-directions (ADC and MD/FA, respectively) were assessed, as well as signal intensity of accompanying T2-weighted images (S0). Peroneus longus demonstrated increase in MD, ADC and S0, the peak and duration of which reflected exercise intensity. FA appeared noisy, although demonstrated large decreases following higher intensity exercise. While further work is needed, this method shows promise in measuring skeletal muscle dynamics.

Purpose

To evaluate a custom gradient acquisition strategy for dynamic measures of post-exercise human calf muscle using Diffusion Tensor Imaging (DTI) that enables temporal resolution, reliable spatial representation, and concurrent estimates of T2 change.

Background

Diffusion and T2 values of active muscle are known to temporarily increase immediately following exercise1-4 although their differing rates suggest different origins of contrast1,2. DTI, using at least 6 diffusion-weighted (DW) directions, produces better spatial representations of anisotropic tissue than single or 3-direction diffusion estimates5, although at the expense of temporal resolution6-8. This study tested the sensitivity of a novel sliding-window gradient scheme, designed to achieve both temporal resolution and spatial accuracy, in detecting changes within human calf muscle following exercise.

Methods

(Exercise Protocol) Lower leg muscles evaluated using a 3T GE MR750 and 8-channel extremity coil (GE Healthcare, Milwaukee WI). DTI was acquired using dual echo SE-EPI (b=0 or 450s/mm2, TR/TE=4000/69.6, custom 15-dir gradient scheme, 16 slices 4mm thick, 16cm FOV, 64x64, ASSET = 2). Baselines consisted of 40-60 volumes (10-15 timepoints), while post-exercise consisted of 60-120 volumes (15-30 timepoints, ~4-8 min). Experiment duration (scanning, exercise, rest) was ~90 min. Diffusion volumes were acquired before and immediately following in-bore exercise (plantar flexion @0.5 Hz for 20sec or 2min) using an MRI-compatible ergometer. It was expected lateral gastrocnemius (LG) and peroneus longus (PER) would be active in this exercise, while anterior tibialis (ATIB) would not9 (Fig 1). Four exercise/scanning sets were performed, each separated by 10min rest. Exercise intensity was based on maximum weight single flex (MWSF) of volunteer, and increased across sets by adding weight or number of flexes. The four exercise sets were: 10 flexes @10%MWSF; 60 flexes @10%MWSF; 10 flexes @40%MWSF; and 60 flexes @40%MWSF.

(Analysis) This study used a custom 15-dir gradient table developed to allow analysis of multiple time scales (Fig 2). Following acquisition, each pre- and post-exercise 4D volume was registered to a common space using FLIRT10, then broken into subunits of 3 or 15 DW images succeeding each b=0 volume (Fig 2). Apparent diffusion coefficients (ADCs) were calculated for each 3-dir subunit11, and diffusion tensors were calculated for each 15-dir subunit using FSL12, to produce mean diffusivity (MD) and fractional anisotropy (FA) maps. Signal intensity of each b=0 (S0) was also assessed. Regions-of-interest (ROIs) were drawn for LG, PER and ATIB, each three 2x2-voxel squares across 5 axial slices (Fig 1) and applied to all subunits for each measure (MD/ADC, FA, S0). Unless specified, all calculations used Matlab (v7.9; Mathworks, Natick MA).

Results

MD/ADC. (Overall) MD and ADC demonstrated similar values for each muscle, although ADC showed more variation across time (Fig 3). MD/ADC returned to baseline values during the post-exercise scanning period (~4-8 min), and within-muscle baselines were similar across all exercise sets. (Between Muscles) The PER showed largest increase in MD/ADC, while LG and ATIB showed minimal change, except after the 10-flex 10% set. ADC of PER increased slightly less than MD, and peaked sooner (Timepoint 2 vs 6). (Exercise Intensity) MD increased in all muscles following the first and least intensive exercise set (10-flex 10%) although ADC did not. The post-exercise increase in MD/ADC of PER showed similarity of peak height within MWSF sets (10% vs 40%), with greater MD/ADC for the 40% sets. The post-exercise increase in PER 40% 60-flex demonstrated greater initial MD/ADC than did the 40% 10-flex set, and also longer sustained increase.

S0. Baselines increased with successive exercise in PER, but not LG or ATIB (Fig 4). PER showed large increase in peak height for both 40%MWSF sets, with greater and sustained increase observed for the 60-flex set. All muscles showed a temporary increase in S0 following the first exercise (10-flex 10%MWSF).

FA. Variability in FA across time made interpretation difficult (Fig 5), although PER showed large post-exercise decreases in both 40%MWSF sets, the 60-flex set showing a greater sustained decrease.

Discussion/Conclusion

Post-exercise diffusion and S0 changes in an expected active muscle (PER) reflected the intensity of exercise sets, while the inactive muscle (ATIB) showed little change. Minimal post-exercise LG change may be due to limb positioning during exercise13. Further work exploring the capabilities of this technique will incorporate other muscles, additional subjects, stricter control and ordering of exercise, and investigation into DTI with <15-dir subunits. Nonetheless, this DTI technique shows promise for representing temporal and spatial aspects of skeletal muscle dynamics.

Acknowledgements

National Sciences and Engineering Research Council of Canada (NSERC) Discovery grant (MDN)

NSERC PGS-D (CPR)

References

1. Morvan D, Leroy-Willig A. Simultaneous Measurements of Diffusion and Transverse Relaxation in Exercising Skeletal Muscle. Magn Reson Imaging 1995; 13(7):943-948.

2. Nygren AT, Kaijser L. Water exchange induced by unilateral exercise in active and in active skeletal muscles. J Appl Physiol 2002; 93:1716-1722.

3. Rockel C, Davis A, Wells G, Noseworthy MD. Monitoring exercise-induced muscle changes using Diffusion Tensor Imaging, with and without caffeine. Proceedings of the International Society of Magnetic Resonance in Medicine, Melbourne Australia, 2012. Poster #1425.

4. Baete SH, Cho GY, Sigmund EE. Dynamic diffusion-tensor measurements in muscle tissue using the single-line multiple-echo diffusion-tensor acquisition technique at 3T. NMR Biomed 2015; 28:667-678.

5. Basser PJ, Mattiello J, Le Bihan D. MR diffusion tensor spectroscopy and imaging. Biophys J 1994; 66:259-267.

6. Damon BM. Effects of Image Noise in Muscle DT-MRI Assessed Using Numerical Simulations. Magn Reson Med 2008; 60(4):934-944.

7. Froeling M, Nederveen AJ, Nicolay K, Strijkers GJ. DTI of human skeletal muscle: the effects of diffusion encoding parameters, signal-to-noise ratio and T2 on tensor indices and fiber tracts. NMR Biomed 2013; 26:1339-1352.

8. Rockel C, Noseworthy MD. An Exploration of Diffusion Tensor Eigenvector Variability Within Human Calf Muscles. J Magn Reson Imaging 2015; E-Pub: May 27, 2015.

9. Snell RS. Clinical Anatomy by Regions. 9th ed. Philadelphia: Lippincott Williams & Wilkens, 2012.

10. Jenkinson M, Bannister P, Brady M, Smith S. Improved optimization for the robust and accurate linear registration and motion correction of brain images. Neuroimage 2002; 17(2):825-41.

11. Le Bihan D, Breton E, Lallemand D et al. MR Imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology 1986; 161(2):401-407.

12. Behrens TE, Woolrich MW, Jenkinson M, et al. Characterization and propagation of uncertainty in diffusion-weighted MR imaging. Magn Reson Med 2003; 50(5):1077-88.

13. Miaki H, Someya F, Tachino K. A comparison of electrical activity in the triceps surae at maximum isometric contraction with the knee and ankle at various angles. Eur J Appl Physiol 1999; 80:185–91.

Figures

Figure 1. Regions of interest (ROIs) measured within the right human calf: anterior tibialis (ATIB), peroneus longus (PER) and lateral head of gastrocnemius (LG).


Figure 2. (A) Schematic of acquisition scheme using 15 non-collinear diffusion directions with embedded b=0 images every 3 DW volumes. (B) Subunits of 3 orthogonal gradient vectors, each preceded by a b=0 volume. (C) Subunits of 15-directions, demonstrating “sliding-window” temporal combination.

Numbers refer to timepoints aligned by last DW volume.


Figure 3. Pre- and post-exercise timecourse of diffusivity measures for anterior tibialis (ATIB), lateral gastrocnemius (LG) and peroneus longus (PER). Top row: 15-dir MD. Bottom row: 3-dir ADC.

Figure 4. Pre- and post-exercise timecourse of signal intensity of b=0 volumes (S0) for anterior tibialis (ATIB), lateral gastrocnemius (LG) and peroneus longus (PER).

Figure 5. Pre- and post-exercise timecourse of 15-dir FA for anterior tibialis (ATIB), lateral gastrocnemius (LG) and peroneus longus (PER).



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
0014