Repeatability of diffusion kurtosis tensor parameters in muscles of the lower legs

The aim of this study was to provide measurements from and investigate the repeatability of diffusion kurtosis tensor parameters in the muscles of the lower legs.

Test-retest acquisition of a kurtosis tensor sequence was performed in 13 healthy volunteers. Quantitative kurtosis tensor parameters were derived, and repeatability of each parameter was evaluated by muscle group and over the whole muscle through intraclass correlation coefficient (ICC) and within-subject coefficient of variation (wsCV). Bland-Altman analysis was also conducted. Differences in parameter values by muscle group were investigated through an analysis of variance.

Axial kurtosis and radial kurtosis values from the test data were 0.63 ± 0.04 and 0.70 ± 0.05, respectively. Kurtosis tensor parameters from all muscle groups and over the whole muscle had wsCV below 15%. ICC for the parameters from most muscle groups was above 85%, with the lowest ICC over the whole muscle being 88.39%. The medial gastrocnemius and extensor digitorum longus showed highest repeatability. Mean, axial, and radial diffusivity had higher wsCV despite being lower-order terms than kurtosis.

This study sought to examine the repeatability of diffusion kurtosis tensor-derived parameters in the legs and verify that they could potentially be used as longitudinal imaging metrics. wsCV values from test-retest data indicated high repeatability throughout all examined muscle groups. There were minimal differences in kurtosis and diffusivity parameters between muscle groups in this healthy volunteer cohort.