Late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) is considered the gold standard for myocardial viability assessment due to excellent scar contrast and high spatial resolution [1,2]. Although the 2D multi-slice, breath-hold approach is widely used in clinic, it has a number of shortcomings. Complete left ventricle (LV) coverage takes nearly 10 minutes to acquire, breath-holds are required for each acquisition, mis-registration errors can occur between 2D slices, and sicker patients with breath-holding difficulties may result in poor quality LGE images resulting from motion artifacts [3]. The purpose of this study was to develop a free-breathing, 3D-LGE (FB 3D-LGE) CMR technique and to compare it with clinically used breath-hold 2D-LGE (BH 2D-LGE)

The proposed FB 3D-LGE technique is a cardiac-gated sequence consisting of an adiabatic inversion pulse (INV), an inversion delay (TI), a fat saturation pulse (FS), outer volume suppression (OVS), followed by 1D projection acquisition and a 3D, segmented stack-of-spirals acquisition (Figure 1). The OVS consisted of a 90° tip-down followed by a spiral 2D tip-up pulse with a circular passband of 14 cm [4]. The OVS allows for reduced field of view (FOV) acquisition, and in addition, the OVS based navigators improve the motion estimation accuracy by isolating the heart in the 1D-projection images. One-dimensional projection signals were acquired along three orthogonal directions. Motion along S-I, A-P and R-L directions were estimated for each cardiac cycle, and a corresponding linear phase was applied along each direction to achieve motion correction (Figure 2). A dual density, segmented stack-of-spirals with 10 interleaves was designed that fully sampled the center of k-space and 1.7 fold under sampled the outer regions of k-space. To utilize the magnetization preparation more effectively, centric acquisition along kz was employed.

Following Institutional Review Board (IRB) approval, FB 3D-LGE data were obtained in 29 adult, cardiac patients (19 women, age = 48.10 ± 14.69), who were scheduled for clinically ordered CMR exams that included BH 2D-LGE scans. Contrast agent was administered intravenously (0.1 mmol/kg, MultiHance). The FB 3D-LGE scans consisted of the following imaging parameters: spatial resolution = 1.6x1.6x2 mm³, in-plane FOV = 30 cm², slab thickness = 16 cm, TR = 10.76 ms, flip angle = 25°, number of partition encodes = 80, acquisition window = 55 ms, 1R-R acquisition, total scan time = 200 beats (3 min, 20 s at 60 bpm). The BH-LGE scans had a spatial resolution of 1.6 x 1.4 mm², FOV = 36 cm, slice thickness = 8 mm. The scans consisted of approximately 10 short axis slices to cover the LV, and a few long axis views (2-chamber and 4-chamber). The FB 3D-LGE scan was performed post-contrast either preceeding or following the BH 2D-LGE scan (range = 7-25 min post-contrast). The 3D images were motion corrected, reconstructed offline and reformatted for comparison with 2D images. Two experienced readers scored the images for image quality (5 point scale) and image artifacts (3 point scale). In patients with scar, fractional scar volume and scar-remote myocardium contrast to noise ratio (CNR) was calculated by drawing manual regions of interest (ROI’s).

Figure 3 contains results from the 2D and 3D techniques from three different patients illustrating representative conditions. There were no significant differences between the FB 3D-LGE and BH 2D-LGE datasets in terms of overall image quality score (P = 0.12) and image artifact score (P = 0.63) ( Table 1). Myocardial infarcts were identified in 5 of the 29 patients with both techniques. The average difference in fractional scar volume between the 3D and 2D methods was 1.1 %, with a Pearson correlation coefficient of 0.96 (n=5), suggesting excellent agreement. There were no significant differences with scar-remote myocardium CNR between the 2D and 3D datasets (P = 0.19) (Table 2). Total scan time was significantly shorter for the FB 3D-LGE over BH 2D-LGE by a factor of 2.83 ± 0.77 (P < 0.0001).We have developed and tested the FB 3D-LGE with the BH 2D-LGE sequence in a cardiac patient cohort. The FB 3D-LGE offers near-isotropic resolution and contiguous LV coverage, in significantly shorter imaging time than the clinically used BH 2D-LGE, while delivering similar image quality and diagnostic value. In conclusion, FB 3D-LGE is a viable option for patients, particularly in acute settings or in patients who are unable to comply with breath-hold instructions.