Spiral SPIRIT Tissue Phase Mapping enables the acquisition of myocardial motion with high temporal and spatial resolution during breath-hold

Marius Menza¹, Daniela Föll², Jürgen Hennig¹, and Bernd Jung³

¹University Medical Center Freiburg, Dept. of Radiology - Medical Physics, Freiburg, Germany, ²University-Heart Center Freiburg, Cardiology und Angiology I, Freiburg, Germany, ³University Hospital Bern, Institute of Diagnostic, Interventional and Pediatric Radiology, Bern, Switzerland

Synopsis

MR Tissue Phase Mapping (TPM) is a powerful approach to assess left ventricular (LV) function. Conventional Cartesian acquisition-strategies with k-t-based parallel imaging acceleration allow the acquisition of a single slice within a breath-hold, but suffer from low spatial resolution. In this work a comparison with undersampled high-resolution spiral SPIRIT TPM for different trajectory designs within one breath-hold and free breathing Cartesian k-t-accelerated PEAK TPM is presented. High image quality, comparable peak velocity values and time to peaks of spiral SPIRIT TPM for high resolution within a breath-hold might enhance myocardial functional analysis.

Introduction

MR Tissue Phase Mapping (TPM) is a powerful approach to assess left ventricular (LV) function. Conventional Cartesian acquisition-strategies with k-t-based parallel imaging acceleration¹ allow the acquisition of a single slice within a breath-hold period with high temporal resolution of about 20 ms. However, despite a k-t-undersampling factor of R=5 the spatial resolution is limited to about 2.2 to 3.4 mm and therefore only acceptable for the evaluation of the LV. Introducing navigator respiration control enables a dramatic increase of spatial resolution allowing the assessment of the right ventricular motion; however, at the cost of much longer scan times².

The aim of this study was to use undersampled spiral imaging in combination with SPIRIT³-reconstruction to enable the acquisition of three-directional velocity encoded TPM in a single slice during breath-hold, while maintaining high temporal and spatial resolution.

Methods

Spiral trajectories with 8 Interleaves were designed and optimized⁴ to support a field of view (FOV) of 32cm (fully sampled calibration area) in a radius of 1/8 for reconstruction purposes for SPIRIT. Outside the calibration area FOV was reduced to 16 cm and 10 cm for dual density (DD) spirals (2-fold, DD 2; 3-fold, DD 3). For variable density (VD) an iterative method was used to linearly decrease the FOV outside the calibration area proportional to the radius in k-space until the readout duration reaches half (VD 2) and one third (VD 3) of a fully sampled readout length (Fig. 1). Measurements in 10 healthy volunteers (age 31±5 years) were performed on a 3T-Prisma-system (Siemens).

For spiral TPM measurements a basal slice was acquired during 16 heartbeats using three-directional velocity-encoded black-blood prepared off-centre spiral gradient echo sequence with prospective ECG gating and 1-1-binomial water excitation in breath-hold (Table 1). For comparison a Cartesian gradient echo sequence was used with navigator respiration control, 1-1-binomial water excitation and k-t-accelerated PEAK-GRAPPA with R=5^1,3. Spiral images were iteratively reconstructed using conjugate gradients SPIRIT with a kernel size of 7x7. The stopping criterion for the algorithm was a residual of less than 10^-³. Data post-processing (Matlab) included semi-automatic segmentation of the LV, eddy current correction and a transformation of the measured in-plane velocities (Vx,Vy) into velocity components perpendicular (Vr) and tangential (Vφ) to the inner heart wall. For segmental analysis the basal LV was divided according to the AHA 16-segment model⁵. Global (averaged over the entire slice) and segmental systolic and diastolic peak velocities and the corresponding time to peak (TTP) values were derived for Vr and Vz. Statistical analysis was performed using an Anova test and multiple comparison with Dunn-Sidak-correction between Cartesian and spiral datasets (*p<0.05;**p<0.01).

Results

Spiral images for all trajectory designs exhibit excellent image quality (Fig. 2). Despite longer spiral readout durations no significant signal voids caused by susceptibility artefacts can be observed. Undersampling artefacts are removed by SPIRIT reconstruction.
Global velocity time courses of Vz and Vr agree well for all spirals and PEAK (Fig. 3) with no significant differences for global and segmental as well as systolic and diastolic Vz peak velocities (Table 2). However, systolic global and lateral segmental Vr peak velocities (DD 2: 2, 3, 6 of 6 segs; DD 3, VD 2, VD 3: 2, 3 of 6 segs) show significant differences between PEAK and all spirals, whereas no significant differences could be observed in diastole. All global and segmental TTP in systole and diastole for both velocities correlate well.

Discussion

Using spiral SPIRIT TPM with undersampling factors of 2 and 3 for DD and VD provide excellent image quality and agreement of peak velocities and TTP for Vz and Vr with k-t-accelerated Cartesian acquisition.
Spirals show more pronounced systolic and diastolic peak velocities for Vr, which might be caused by temporal blurring due to k-t-acceleration in Cartesian data. With a much smaller acquisition matrix (and thus lower spatial resolution) the recently introduced study with the same acceleration factor of R=5 did not show temporal blurring effects¹. Simpson et al.⁶ reported also an increase in systolic and diastolic of peak velocities (Vz,Vr,Vφ) for undersampled 8 interleaved spiral acquisition, reconstructed with nonlinear SENSE. In comparison to Simpson delineation and structural details of the LV and RV seem to be clearly improved due to a somewhat higher spatial resolution, SPIRIT reconstruction and reduced artefacts due to shorter spiral readouts in our study. Assessment of peak velocities and TTP revealed no significant differences between different spirals acquisition strategies, so that for further studies VD 3 can be used to achieve high temporal resolution.
Future work will comprise a comprehensive biventricular analysis of myocardial motion.

Acknowledgements

No acknowledgement found.

References

1. Bauer et al. K-t GRAPPA Accelerated Phase Contrast MRI: Improved Assessment of Blood Flow and 3-Directional Myocardial Motion During Breath-hold, J Magn Reson Imaging. 2013 Nov; 38(5):1054-62.
2. Menza et al. Segmental Tissue Phase Mapping analysis of biventricular heart function, Proceedings ISMRM 2013, No. 4483
3. Lustig et al. SPIRiT: Iterative Self-consistent Parallel Imaging Reconstruction From Arbitrary k-Space, Magn Reson Med. 2010 Aug; 64(2):457-71.
4. Hargreaves et al. Time-optimal multidimensional gradient waveform design for rapid imaging, Magn Reson Med. 2004 Jan; 51(1):81-92.
5. Cerqueira 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, Int J Cardiovasc Imaging. 2002 Feb;18(1):539-42
6. Simpson et al. Spiral Tissue Phase Velocity Mapping in a Breath-hold with Non-Cartesian SENSE, Magn Reson Med. 2014 Sep; 72(3):659-68

Figures

Figure 1: First spiral trajectories for dual density (left) and variable density (right) for undersampling factors of 2 (top) and 3 (bottom). 1/8 around the k-space centre is fully sampled as calibration area for SPIRIT reconstruction.

Figure 2: Comparison of magnitude images for PEAK and spiral TPM data for systole (top) and diastole (bottom); Breath-hold spiral TPM provide comparable image quality to free breathing Cartesian PEAK-Grappa TPM.

Figure 3: Comparison of mean time courses over all volunters between PEAK and spiral SPIRIT TPM for Vz (left) and Vr (right). Global velocity time courses of Vz and Vr agree well, whereas peak velocities of Vr in systole and diastole for PEAK data seem to be slightly underestimated.

Table 1: Overview of sequence parameters for Cartesian PEAK TPM and undersampled spiral SPIRIT TPM.

Table 2: Overview of systolic and diastolic peak velocities and TTP for Cartesian PEAK and spiral SPIRIT TPM data. Significant differences between PEAK and spiral are indicated by * p<0.05; ** p<0.01).

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)

3131