Synopsis

Keywords: Skeletal, Muscle

Diffusion tensor imaging (DTI) has been shown to be highly sensitive to subtle pathological alterations caused by multiple diseases. Utilizing this capability of DTI, in this study we characterized post-exercise hyperemic responses of calf muscles. We found that multiple DTI measures, including ADC, λ_1-3, and RD, increased significantly in activated muscles, and more than in those in non-activated muscles. Of particular interest was the significant increase in λ₃ and RD of the activated muscle, indicating a relatively higher degree of increase in water diffusion along the cross-section direction of muscle fibers.

Introduction

Clinical T₁- and T₂-weighted MRI are capable of detecting significant changes in skeletal muscles induced by inflammation, trauma or atrophy, but not earlier pathological alterations on cellular or fascicular level¹. Diffusion tensor imaging (DTI) has been shown to be sensitive to changes in the tissue microstructure, and enables the characterization of muscle fiber architecture¹. Herein, we aimed to explore the value of DTI in quantifying exercise-induced changes in calf muscles, so as to better understand exercise-stimulated muscle hyperemia from the perspective of water diffusion.

Methods

The study was approved by our institutional review board. We recruited 6 healthy subjects (male, age 22±3 years, weight 75±13 kg), and informed consent was signed before the experiment. On a 3T MRI scanner (United Imaging Healthcare, China), the subject was positioned head-first, with a 12-channel flex coil wrapping around one calf. With a custom-built apparatus, the subject performed plantar flexion with a frequency of 1 Hz for 6 minutes and a load of 18 lbs. Before and immediately after the exercise, DTI based on echo planar sequence was performed with following parameter values: b values 0 and 400 s/mm², TE/TR 70/4000 ms, 6 directions, FOV 250×250×64 mm, average number 9.
With each set of the acquired DTI data, voxelwise fitting of exponential decay was performed to obtain maps of apparent diffusion coefficient (ADC), three eigenvalues (λ₁, λ₂, λ₃) and radial diffusivity (RD), and 3D tractography was constructed to visualize muscle fibers. On the image of b = 0 s/mm², regions of interest (ROI) were manually drawn to include lateral gastrocnemius (LG) and soleus (SOL), respectively. For each ROI, cautions were taken to exclude fascia, blood vessels. The ROIs were applied to the above parameter maps to obtain the averaged unweighted (b  =  0 s/mm²) signal intensity (S₀), ADC, λ_1-3 and RD for each muscle group. To quantitatively characterize diffusivity anisotropy in skeletal muscles, we computed ellipsoid eccentricity (e) as (1-λ₃/λ₁)^-1/2 for each muscle group. Here λ₁, λ₂, and λ₃ represent the diffusive transport along the long axis of a muscle fiber, within the endomysium perpendicular to the long axes of the muscle fibers, and within the cross-section of a muscle fiber, respectively³. Hence, an increased value of e corresponds to a change in muscle morphology from a spherical to a more longitudinal shape.

Results

Representative maps for S₀, λ_1-3, ADC and tractography before and after exercise are shown in Fig. 1. Before exercise, LG and SOL did not show any significant difference between their S₀ values, or between their ADC values (P>0.50). With exercise, S₀ of LG increased significantly from 74.1±9.0 to 121.4±32.3 (64.51% on average, P=0.01), while S₀ of SOL did not change significantly (84.36±31.33 to 95.68±30, P=0.47) (Fig. 2a). Similar patterns of exercise-induced responses were observed for the ADC values (Fig. 2b): ADC of LG increased significantly from 1813.0±166.5 to 2145.0±268.4 µm²/s (15.4% on average, P=0.04), and for SOL, from 1722.0±94.5 to 1807.0±218.4 µm²/s (P=0.45).None of the eigenvalues (λ_1-3) was significantly different between LG and SOL prior to exercise (P>0.57). All eigenvalues of the two muscle groups increased with exercise, but only the increase of λ₃ in LG was significant (1376.0±122.1 to 1655.0±134.5 µm²/s, 20.5%, P =0.008) (Fig. 3a). The percentage increases in λ_1-3 are shown in Fig. 3b, where obviously increases in LG (14.5%, 18.3%, 20.5%) were significantlysignificiantly higher than the corresponding increases in SOL (3.6%, 4.9%, 7.1%)。The exercise-induced changes in the three eigenvalues were also reflected in the estimates of RD and ellipsoid eccentricity (Fig. 3c and 3d). For LG with exercise, its RD increased significantly from 1600.0±135.0 to 1907.0±223.6 µm²/s, P=0.03) and ellipsoid eccentricity had a trend of decrease. Neither of the two parameters changed significantly for SOL with exercise.

Discussion

In this work, we explored how intensive plantar flexion would impact the DTI metrics in calf muscles. It was found that ADC, λ_1-3 and RD increased significantly in the activated gastrocnemius muscle, but not in the soleus. In skeletal muscles, myofilaments inside muscle cells are organized in a geometry that restricts water diffusion⁴. With exercise to induce hyperemia and thus significant enlargement of muscle intracellular space⁶, water diffusion along all directions may be accelerated. This may explain the increase of λ₁ (the principal direction) and λ_2-3 (the perpendicular directions)⁵. The increase of T₂-weighted signal intensity S₀ also supports the increase of water content in such a scenario. The relatively more increase in λ₃ led to a decrease in ellipsoid eccentricity for both LG and SOL with exercise. This indicates a transition of muscle diffusion from a highly anisotropic state (mostly along the muscle fibers) at rest to a less anisotropic state in exercise. A possible explanation for the such trend may be the increased exchange of water and oxygen along the cross-section direction of muscle fibers, more than that along the long axis of a muscle fiber.

Conclusion

DTI revealed significant differences in diffusion between activated and non-activated muscles, and the difference between along the long axis of muscle fiber and perpendicular to it, which is highly related to the functional structure of muscle fibers. Such a phenomenon demonstrates that DTI has a promise for monitoring subtle changes in muscle pathology.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant No. 82171924).