Capturing the breath of the heart by magnetic resonance imaging : five-dimensional cardio-respiratory resolved cine imaging during free breathing

Abstract

Cardiac imaging with cardiovascular magnetic resonance has the advantage of enabling dynamic visualizations, several quantifiable measures, and good soft-tissue contrast. However, a major drawback is the need for breath holding during many acquisitions to limit motion artifacts. Breath holding is often insufficient for artifact reduction and, furthermore, limits the image acquisition to only one respiratory phase. The aim of this work, therefore, was to develop a free breathing cardiovascular magnetic resonance imaging method for three-dimensional whole-heart cine imaging to measure the left ventricular (LV) volume in all combinations of cardiac and respiratory phases.

Three-dimensional free-running acquisition was implemented as a double golden-angle radial trajectory. Respiratory self-gating signals were extracted from the center of k-space and used in combination with the measured electrocardiogram to bin data into combinations of cardiac and respiratory phase. Image volumes covering the whole heart were reconstructed and the LV endocardial border was manually segmented.

In the first of four studies (Study I), the proposed technique was validated for LV volumes in end diastole (EDV) and systole in end expiration against gold standard two-dimensional breath held cine imaging in healthy volunteers. No difference was found between the two methods with regard to LV volumes and in test-retest variability of the volume measurements. Study II focused on the respiratory cycle where the same free breathing acquisition was used for measuring the respiratory-induced variation in LVEDV and the acquisition time was reduced from 40 minutes to 4.5 minutes by implementing parallel image reconstruction. The maximum respiratory variation in LVEDV was found to be 5-6% for healthy volunteers. An extension to the double golden-angle acquisition was implemented in Study III to reduce eddy current-induced artifacts, which improved image quality. A projection-based respiratory self-gating strategy was added to the improved trajectory in Study IV and used to measure respiratory variation in LVEDV in consecutive patients referred for cardiac evaluation with cardiovascular magnetic resonance imaging. In patients, the k-space based self-gating failed to measure the respiratory variation in LVEDV, whereas with the projection based self-gating yielded an 8% respiratory variation in LVEDV.

In conclusion, this work presents a novel method for measuring the LVEDV resolved in a two-dimensional cardio-respiratory phase map and provides three-dimensional isotropicresolution image volumes enabling multiple view angles of the heart in one acquisition.