The repeatability of bilateral diffusion tensor imaging (DTI) in the upper leg muscles of healthy adults

Abstract Objectives Assessment of the repeatability of diffusion parameter estimations in the upper leg muscles of healthy adults over the time course of 2 weeks, from a simultaneous bilateral upper leg DTI measurement. Methods SE-EPI DTI datasets were acquired at 3 T in the upper legs of 15 active adults at a time interval of 2 weeks. ROIs were manually drawn for four quadriceps and three hamstring muscles of both legs. The following DTI parameters were analyzed: 1st, 2nd, and 3rd eigenvalue ( λ 1 , λ 2 , and λ 3 ), mean diffusivity (MD), and fractional anisotropy (FA). DTI parameters per muscle were calculated with and without intravoxel incoherent motion (IVIM) correction together with SNR levels per muscle. Bland-Altman plots and within-subject coefficient of variation (wsCV) were calculated. Left-right differences between muscles were assessed. Results The Bland-Altman analysis showed good repeatability of all DTI parameters except FA for both the IVIM-corrected and standard data. wsCV values show that MD has the highest repeatability (4.5% IVIM; 5.6% standard), followed by λ 2 (4.9% IVIM; 5.5% standard), λ 1 (5.3% IVIM; 7.5% standard), and λ 3 (5.7% IVIM; 5.7% standard). wsCV values of FA were 15.2% for the IVIM-corrected data and 13.9% for the standard analysis. The SNR (41.8 ± 16.0 right leg, 41.7 ± 17.1 left leg) and wsCV values were similar for the left and right leg and no left-right bias was detected. Conclusions Repeatability was good for standard DTI data and slightly better for IVIM-corrected DTI data. Our protocol is suitable for DTI of the upper legs with overall good SNR. Key Points • The presented DTI protocol is repeatable and therefore suitable for bilateral DT imaging of the upper legs. • Additional B1 + calibrations improve SNR and repeatability. • Correcting for perfusion effects improves repeatability.