Purpose or Case Report: Introduction: Conventional pediatric MRI acquisitions of a short axis (SA) stack for ventricular volumes typically requires one breath-hold per slice, resulting in multiple coordinated breath-holds under anesthesia to acquire the entire stack. Here, we aim to validate a compressed sensing approach to reduce breath-holding during SA balanced steady state free precession (bSSFP) cine imaging. Equivalency is tested via qualitative image scoring and global volumes compared to conventional bSSFP cine imaging.
Methods & Materials: Method: 22 patients undergoing clinically indicated cardiac MRI were enrolled in this IRB-approved study (age: 16±8y [range: 3-36]; 12 females; HR: 71±15bpm [range: 50-117]). All imaging was performed on a commercial 3T scanner (Ingenia, Philips Healthcare) using retrospective ECG-gated cine bSSFP. Cine SA images covering both ventricles were acquired with conventional parallel imaging (SENSE = 1.5-2) and a vendor-optimized parallel imaging/compressed sensing approach (compressed SENSE [CS] = 2.5 < 40 kg ≤ 3.5). Identical image acquisition parameters were used (TR/TE/FA: 3 ms/1.5 ms/40-45°; in-plane voxel resolution: 0.9-1.4 mm; 10-14 slices; slice thickness: 6-10 mm; temporal resolution: 30±5 ms [17-40 ms]). Image quality was scored for blood-myocardial contrast (BMC), edge definition (ED), and presence of artifact (PA) by an experienced radiologist. Quantitative comparisons included LV and RV end systolic volume (ESV), end diastolic volume (EDV), ejection fraction (EF) (Circle CVI42, Calgary).
Results: Results: No differences were found with image scores for BMC, ED, and PA (p = 0.67, 0.27, 0.14). No differences were found between LV EDV, ESV, and EF for the two sequences (p = 0.92, 0.77, 0.33) or RV EDV, ESV, and EF (p = 0.18, 0.13, 0.58). Scan duration (not including rest between breath-holds) was 8±2 s/slice for conventional imaging (1min36s±24s total) vs. 4±2 s/slice for CS (54 s±18 s total). With CS, 2-4 slices were acquired per breath-hold, further speeding up acquisition 2-4 times (including rest between breath-holds).
Conclusions: Conclusion: Equivalent image quality and cardiac volumes were found between acceleration approaches. The time savings per slice acquisition with CS allowed for a 75% reduction in number of breath-holds (12 vs. 3) without differences in qualitative or quantitative measurements as compared to the conventional technique. These time savings will reduce the number of breath holds and result in and shorter and simpler anesthesia protocols.