Synopsis

Keywords: Non-Proton, Heart, phosphorus

Cardiac 31P-MRS provides insight into metabolism of the failing heart. The commonly employed 3D-MRSI acquisitions are rather slow, even at 7T. Fast readout trajectories, such as concentric ring (CRT) have been recently suggested to substitute the slow Cartesian sampling (CSI), however their repeatability is yet unknown. Our preliminary data suggest comparable intra-session repeatability, but somewhat lower inter-session repeatability for mid septal voxels using 2.5 min CRT in comparison to 6.5 min CSI. Similar trends were observed for higher resolution CRT.

Introduction

Phosphorus magnetic resonance spectroscopy (³¹P-MRS) can probe the energy metabolism of the human heart in vivo by measuring the phosphocreatine to adenosine triphosphate concentration ratio (PCr/ATP), an indicator of heart failure (1), or the potentially more sensitive chemical kinetics of the oxidative phosphorylation system. In both cases fast scanning is desirable, to either fit ³¹P-MRS into a clinical protocol, or to mitigate the multiple acquisitions required for the latter metrics.

Employing fast MRSI readout trajectories, it could be possible to leverage the full 2.8 times increase in SNR(2) to achieve close to the theoretical 7.8-times (2.8²) speed increase when moving from 3 to 7 T. This is not possible using the most common approach, point-by-point Cartesian sampling (CSI). A new method employing a fast MRSI readout trajectory, concentric ring trajectory (CRT), has recently been described (3).

CRT can measure PCr/ATP maps as CSI sampling in a fraction of the time (with matched resolution) or with higher spatial resolution (in a matched time). While the repeatability of cardiac 3D 31P-MRSI sequences using Cartesian sampling at 7T is known (4), the repeatability of the CRT sequences is yet to be determined. Hence, the aim of this study was to evaluate the intra- and inter-session repeatability of three different CRT protocols.

Methods

Four sequences were evaluated: Cartesian acquisition-weighted MRSI, termed CSI henceforth, (matrix size 10x10x10, 4 averages at k=0, 6:31 mins) and the CRT sequence with a matched matrix size employing 12 (2:31 mins) and 10 rings (1:37 mins), and with a matrix size of 12x12x12 and 19 rings (6:55 mins, labelled as high-res CRT [HRCRT]). These four sequences were run twice in a session and on two different days, 72 hours apart, to evaluate intra- and inter-session repeatability (Figure 1 ).

Three healthy, lean, young volunteers (1 female) were scanned supine in a Siemens Magnetom 7T scanner (Siemens, Erlangen) equipped with a square surface transmit and 16 channel receive array coil (Rapid Biomedical) positioned over the heart (4). CSI was run with 240x240x200 mm³ FoV, 1 s TR, 8 kHz bandwidth, and 2048 time-samples. CRT had matched FoV and TR, but 2778 Hz bandwidth with 720 time-samples.

CSI data were reconstructed online and CONCEPT data offline using the non-uniform FFT (NUFFT) toolbox with min-max Kaiser-Bessel kernel interpolation and twofold oversampling (5) in MATLAB (MathWorks, Natick, USA). No density compensation was required. Individual coil data was combined using the WSVD algorithm (6). CSI/MRSI data fitting was performed using the OXSA toolbox (7).
Sequences were compared using PCr/γ-ATP ratios, corrected for partial saturation and blood contamination, in mid-septal voxels from apical, mid, and basal slices of the heart. All corrections were calculated per-subject and per-session (2).

The intra-session variability was assessed through the mean and difference between PCr/ATP ratios from equivalent datasets within the same session, e.g., by comparison of datasets A1 with E1 and A2 with E2 of each volunteer for the CSI acquisitions. And the inter-session variability was assessed through the mean and difference between PCr/ATP ratios from equivalent datasets in both protocols for each subject, e.g, by comparing Dataset A1 with A2, and E1 with E2 in case of CSI (see Figure 1 for detailed description).

The coefficient of repeatability (CoR) was calculated from SD of the signed differences in PCr/ATP between two scans for each subject according to CoR = SD_intrasubject × $$$\sqrt2$$$ × 1.96. A lower CoR reflects higher repeatability.

Results & Discussion

Figure 2 depicts the typical position of the analysed voxels as well as representative spectra from the mid septal voxel of one subject demonstrating good agreement between the acquired data for all sequences. There were no significant differences (paired t-test) in the calculated cardiac PCr/ATP between repeated measurements for all voxels, sequences, and repetitions (Figure 3). The Bland-Altman plots showing the repeatability of all sequences in apical and mid septal voxels are depicted in Figure 4.

High repeatability was observed for the long CSI scan with CoR for mid septal voxel 0.40 for intra-session and 1.17 for inter-session comparison. While our subject numbers are currently low, this falls into the same range as previously reported repeatability of 8x16x8 CSI matrix by Ellis et al. with CoR of 0.67 and 0.75 for intra- and inter-session comparisons, respectively. To achieve the increase in speed, the CRT protocols sacrifice some signal-to-noise (SNR), which leads to higher CoR, i.e., 1.37 for the intra-session and 2.52 for the inter-session comparison of mid septal voxel of the 2.5 minute CRT sequence. For full report of the observed CoRs see Table 1. The high CoR values for basal voxels reflect the very low SNR suggesting that with the current setup these voxels are effectively not usable. This could be potentially mitigated by the use of a whole-body transmit coil (8).

Conclusion

Our preliminary results suggest comparable intra-session repeatability, but lower inter-session repeatability, of highly accelerated 3D-MRSI acquisition using CRT trajectory, providing either improved speed (2.5 minutes in comparison to 6.5 minutes CSI) and increased spatial resolution (6.66 mL in comparison to 11.5 mL nominal resolution).