Comparison of cardiac diffusion MRI using multiple prospective respiratory motion correction techniques.

Abstract

Purpose: A novel prospective motion correction control system with slice tracking (MNav-CoS) was compared with three other prospective respiratory motion correction techniques in performing free-breathing cardiovascular diffusion tensor imaging (cDTI) acquisitions.

Methods: Ten healthy volunteers underwent cDTI using an M2SE sequence. The performance of the proposed MNav-CoS was compared with three respiratory compensation techniques: multiple breath-holds (BH), free breathing with respiratory gating (Gate), and free breathing with single navigator slice tracking (Nav). Data for five diffusion weightings were acquired in a single mid-ventricular slice in end systole. MD, FA, and HA maps were calculated for each technique and combinations of low and high b-values. Data from the respiratory navigators were used to estimate the total amount of cardiac through-plane motion during free breathing.

Results: The metrics derived from the diffusion tensor for MNav-CoS with b_low|b_high = 50|450 s/mm² were MD: $1.48\pm 0.10\mu m⁄{\text{ms}}^2$ , FA: $0.39\pm 0.07$ , and HAg: $-0.82\pm 0.22°⁄\%$ . All of the other respiratory compensation techniques produced a similar range of results to the MNav-CoS technique. On average, the free-breathing acquisitions with slice tracking were three times shorter than using BH. The total amount of cardiac through-plane motion during the free-breathing acquisitions ranged from 4 to 10 mm with an average of $6.2\pm 1.7\text{mm}$ .

Conclusion: The MNav-CoS technique performed comparably to other commonly used respiratory compensation techniques. Prospective respiratory motion compensation, such as the slice tracking used with MNav-CoS, is a useful tool that offers time-saving benefits and compensates for through-plane motion present during free breathing. These techniques may be beneficial for performing longer cDTI acquisitions providing increased utility in a clinical context.