Performance of Free-Running Cardiac and Respiratory Motion-Resolved Whole-Heart 5D MRI in Congenital Heart Disease...

The performance of 5D imaging in CHD has been previously demonstrated using ferumoxytol rather than gadolinium.

To evaluate the performance of gadolinium enhanced 5D MRI relative to 2D and 3D MRI in CHD patients.

Retrospective.

45 consecutive CHD patients referred for clinically indicated MRI who underwent 2D, 3D, and 5D MRI.

2D bSSFP CINE, 3D bSSFP, and 5D FISS at 1.5 T.

Scan time (2D, 3D, and 5D), left and right ventricular (LV, RV) end diastolic and end systolic volumes (LVEDV, RVEDV, LVESV, and RVESV, respectively), ejection fraction (LVEF, RVEF) for 2D versus 5D, and image quality on a 4-point Likert scale (3D vs. 5D).

Paired comparisons were made using a paired t-test or Wilcoxon signed rank test. Agreement between measurements was evaluated with Bland-Altman analysis including the bias and 95% limits of agreement.

Scan time for 5D (6:18 [0]) was significantly shorter than 2D (7:16 [4:39]) and 3D (6:55 [1:30]) MRI. 5D measurements were strongly correlated with 2D measurements of LVEDV (R² = 0.94), LVESV (R² = 0.83), LVEF (R² = 0.51), (R² = 0.85), RVESV (R² = 0.80), and RVEF (R² = 0.66) but with significant biases. Sharpness of the coronary arteries were significantly higher for 3D (RCA: 36.8% ± 10.2%, LAD: 40.5% ± 10.5%) than 5D (RCA: 24.3% ± 7.5%, LAD: 31.7% ± 8.7%), but longer vessel segments were visible using 5D (LAD: 6.9 ± 2.5) than 3D (LAD: 5.3 ± 2.9 mm) acquisitions. The grades for 5D (R1: 3 [1], R2: 2.5 [0.5], R3: 3 [0.6]) were significantly higher than 3D (R1: 3 [1], R2: 2.5 [1], R3: 2 [1]).

Gadolinium-enhanced 5D MRI may provide an efficient and simplified acquisition for CHD patients, with multiple quantitative parameters from 5D being not significantly different from those from reference standard 2D and 3D MRI.

4 STAGE OF TECHNICAL EFFICACY: 1.