Purpose or Case Report:
Conventional cardiac MRI (CMR) acquisitions in children are limited by long acquisition time, need for sedation, and an inefficient workflow. A conventional cine SSFP short-axis stack typically takes 5-7 minutes to complete and is ineffective for subjects with arrhythmias. Available alternatives like free-breathing, real-time cine (RT-cine) SSFP have suboptimal temporal and spatial resolution. Recent advances in compressed sensing (CS) MRI techniques may overcome this limitation. We hypothesize that a CS-based approach will allow the scan time for a short axis RT-cine stack to be shortened to less than a minute without compromising spatial or temporal resolution.
Methods & Materials:
Five patients (Age: 14-44 years ;4 males) underwent CMR study with both conventional breath-held cine SSFP (BH-cine) as well as CS-based free-breathing, cardiac-triggered RT-cine acquisitions (Ahmad and Schniter, DOI: 10.1109/TCI.2015.2485078). The short-axis of the ventricles was covered with 14 slices in each patient. The spatial resolution for both acquisitions was 1.8x1.8x8 mm³; temporal resolution was 32 ms (BH-cine) and 32 ms (RT-cine) per cardiac phase. End-diastolic volume (EDV), end-systolic volume (ESV), ejection fraction (EF), and scan duration were computed using established methods, and compared between BH-cine and RT-cine acquisitions using paired Student’s t-test.
Results:
Both BH-cine and RT-cine images were deemed of diagnostic quality by two expert readers. The EDV, ESV and EF showed good agreement (Mean EDV (BH-cine first): 144 ± 27 vs 137 ± 27 ml, p = 0.03; mean ESV: 58.8 ± 10.4 ml vs. 54.9 ± 6.4, p = 0.29; and mean EF = 58.8 ± 5% vs 54.9 ± 6.4 % ;p = 0.03 ); see figure 1. The scan times were significantly shorter for RT-cine (2 heart beats/slice; ~45 seconds for entire 14 slice acquisition) compared with BH-cine acquisition (6-8 beats/slice; ~5:38 minutes for 14-slice acquisition, p< 0.01). Figure 2 shows representative images acquired.
Conclusions:
CS-based free-breathing RT-cine technique is feasible and provides images of comparable quality and diagnostic value when compared to conventional BH-cine SSFP, while significantly reducing scan time. It offers a potential solution to avoid the need for sedation for functional evaluation, improve diagnostic utility of MRI in arrhythmia, and improve CMR efficiency and workflow. Future reduction in reconstruction times, is needed prior to implementation of CS-based RT techniques into clinical practice.