Assessment of Motion patterns in free breathing MRI of the abdomen using continuously tagged imaging

Introduction The entire abdomen is subject to complex motion patterns [1,2], comprising of breathing, heart and peristaltic motion. Insight into the contribution of each source is important to motility research and motion correction. This study proposes the use of continuously tagged imaging for motion assessment during free breathing, non-triggered, non-gated, dynamic MRI of the entire abdomen. Data acquisition Tagging was originally developed for cardiac imaging and is linked to the cardiac motion cycle using ECG triggering [3-5]. In the present work each tagging prepulse is followed by a fast 3D read-out, this is repeated in a dynamic series resulting in a continuous registration of tag patterns and the motion associated with the tag patterns. Hence, tagged dynamic data can be acquired without any assumption or demand on the motion of interest. A delay must be inserted between the tagging prepulse and dynamic readout to allow motion to deform the tag pattern or tag sheets. The sampling duration is defined by the sum of this delay and the readout duration itself. The readout sequence must be fast to enable high sampling rates and avoid T1-decay of the tag pattern. After approval of the Institutional Review Board a free breathing dynamic set of 150 dynamics was acquired in one healthy consenting volunteer with a delay of 50 ms between tagging prepulse and readout. The readout duration was as short as 85 ms by using a combination of parallel and half-Fourier imaging, such that the sampling rate was 7 Hz. The voxel size was 3 mm isotropic, the FOV was 400x400 with 6 slices. Motion extraction Tag patterns per dynamic were determined automatically using a scale-space based algorithm [6] (Figure 1) to render high detail deformation fields (Figure 2). Accumulating the difference in deformation over time resulted in a velocity map (Figure 3), in which heart pulsation and breathing motion can be recognized (Figure 4).