Synopsis

Cardiac T₁ mapping provides valuable information on fibrosis in various cardiomyopathies. Commonly, data acquisition is restricted to a small percentage (mid-diastole) of the cardiac cycle to prevent motion artefacts. This leads to low scan efficiency and limits the achievable resolution. We present an 8s T₁ mapping approach, which employs a large acquisition window (80% of the cardiac cycle) and corrects for cardiac motion using the same data. The approach was evaluated in native T₁ mapping in healthy volunteers and post-contrast T₁ mapping in patients. It significantly improved the precision of the obtained T₁ maps while successfully minimizing cardiac motion artefacts.

INTRODUCTION

Myocardial T₁ is an important marker for the detection of fibrosis in various cardiomyopathies.¹ Cardiac triggering is necessary in T₁ mapping techniques to avoid cardiac motion artefacts. Usually a relatively short acquisition window in mid-diastole of only 15-30% of each cardiac cycle is used, limiting the spatial resolution of the acquired qualitative images (TI images).²

Here, we present cardiac motion-corrected T₁ mapping in a short breath-hold of 8s with an in-plane resolution of 1.3x1.3mm². The same raw data is used both for cardiac motion estimation from cine images and high-resolution T₁ mapping.³ Due to cardiac motion correction, images obtained during 80% of each cardiac cycle could be used for T₁ mapping, improving the acquisition efficiency by a factor of 5 to standard T₁ mapping approaches.

The proposed method was evaluated using native T₁ mapping in 10 healthy volunteers and its clinical feasibility was shown in two patients with focal myocardial fibrosis by comparing post-contrast T₁ maps to late gadolinium enhancement (LGE) images.

METHODS

Data acquisition: 2D slices were acquired continuously with golden-angle radial sampling (Figure 1). Data were acquired within a breath-hold on 3 tesla (Verio, Siemens Healthineers): flip angle: 5°, TE/TR: 2.03/4.93ms, FOV: 320x320 mm², resolution: 1.3x1.3x8.0 mm³, acquisition time: 8s and inversion pulses were applied every 2276ms.

Motion estimation: Data was retrospectively binned into 15 cardiac motion states and cine images were reconstructed iteratively using spatial and temporal total variation regularisation.⁴ Non-rigid cardiac motion estimation was performed on these cine images using NiftyReg.⁵

T₁ mapping: In each cardiac cycle, 15 TI images were reconstructed of exactly the same cardiac phases as used for cine reconstruction. Images were reconstructed using non-Cartesian iterative SENSE.⁶ Estimated inverse deformation fields were applied to TI images to correct for cardiac motion. Three systolic cardiac phases were excluded to minimize through-plane motion artefacts, resulting in a total T₁ mapping window of 80% per cardiac cycle. T₁, M₀ and the flip angle were voxel-wise obtained by a 3-parameter fit to the signal model of the acquisition.^3,7

In vivo imaging: The approach was applied without contrast agent in 10 healthy subjects (6 males, aged 31.3±8.4years) in a mid-ventricular short axis orientation (T1_moco). For comparison, T₁ maps were also obtained without motion correction of the same TI images (T1_uncorr) and using the standard cardiac-gated approach (T1_1phase, using only data from a window of 168ms per cardiac cycle). Post-contrast T₁ maps were obtained in two patients with focal fibrosis (1 female/1 male, aged 48/66 years) after contrast administration (0.15 mmol/(kg body weight) Gadoteridol). To detect focal fibrosis, LGE images were acquired 15 minutes after contrast administration.

Evaluation: T₁ values were assessed in six myocardial segments and in the left-ventricular blood pool.⁸ For precision analysis, the standard deviation (STD) of native T₁ in each segment was determined and differences were assessed by a comparison between T1_1phase and T1_moco by a Wilcoxon signed-rank test in each segment across healthy volunteers. In both post-contrast T1_moco maps, T₁ values of the fibrotic segment were compared against T₁ of all healthy segments using ANOVA (GraphPad Prism).

RESULTS AND DISCUSSION

Cine images had a consistent contrast over all cardiac phases and motion estimation succeeded in all subjects, with only small residual motion (Figure 2). The window length of each cardiac bin was between 55 and 87 ms, depending on the heart rate (46-73 bpm).

T₁ maps are shown in figure 3. T1_moco and T1_1phase times over all myocardial segments and healthy volunteers were 1287.2±46.7ms and 1258.7±26.0ms, respectively (Figure 4a).

T1_moco maps had a 40% lower STD compared to T1_1phase maps in all myocardial segments and in blood (averaged over all myocardial segments: 57.0±12.5 and 94.8±15.4 ms, respectively, P<0.004, Figure 4b). The total amount of data used for T1_moco mapping was on average 5.2 (4.3–6.8, depending on heart rate) times higher compared to T1_1phase mapping, leading to the improvement in precision.

In post-contrast T₁ maps, the border between myocardium and blood was better defined in the T1_moco maps compared to the T1_uncorr maps (Figure 5). Focal fibrosis could be detected in post-contrast T1_moco maps at the same location as in the LGE images by a decreased T₁ value in the fibrotic segment compared to all healthy segments (-69.5±11.9ms and -51.8±12.0ms, P<0.001).

CONCLUSION

Acknowledgements

References

1. Schelbert EB, Messroghli DR. State of the Art: Clinical Applications of Cardiac T1 Mapping. Radiology 2016;278(3):658-676.

2. Piechnik SK, Ferreira VM, Dall’Armellina E, Cochlin LE, Greiser A, Neubauer S, Robson MD. Shortened Modified Look-Locker Inversion recovery (ShMOLLI) for clinical myocardial T1-mapping at 1.5 and 3 T within a 9 heartbeat breathhold. J Cardiovasc Magn Reson 2010;12(1):69.

3. Becker KM, Schulz‐Menger J, Schaeffter T, Kolbitsch C. Simultaneous high‐resolution cardiac T1 mapping and cine imaging using model‐based iterative image reconstruction. Magn Reson Med 2018;00:1–12.

4. Block KT, Uecker M, Frahm J. Undersampled radial MRI with multiple coils. Iterative image reconstruction using a total variation constraint. Magnetic Resonance Imaging 2007;57(6),1086–1098.

5. Modat M, Ridgway GR, Taylor ZA, et al. Fast free-form deformation using graphics processing units. Comput Methods And Programs Biomed.2010;98(3),278–284.

6. Pruessmann KP, Weiger M, Boernert P, Boesiger P. Advances in sensitivity encoding with arbitrary k-space trajectories. Magn Reson Med 2001;46(4),638–651.

7. Deichmann R, Haase A. Quantification of T1 Values by SNAPSHOT-FLASH NMR Imaging. J Magn Reson. 1992;96:608-612.

8. Cerqueira MD, Weissman NJ, Dilsizian V, Jacobs AK, Kaul S, Laskey WK, et al. Standardized myocardial segmentation and nomenclature for tomographic imaging of the heart: a statement for healthcare professionals from the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. Circulation 2002;105(4):539-42.