Introduction

The cardiac conduction system (CCS) is responsible for propagating electrical impulses throughout the heart, which results in the synchronized beating of the heart. The architecture of the CCS is implicated in arrhythmogenic pathologies such as ventricular fibrillation ¹, which often results in sudden cardiac death. Characterizing the CCS is therefore of high importance for understanding these cardiac arrhythmias.
It has been demonstrated that Magnetization Transfer (MT) MRI provides a contrast between the conducting Purkinje fibers (PF) and the myocardium by targeting the collagen sheath surrounding the fibers ². This exploratory study at 9.4 T aims at evaluating a new contrast between the PF and myocardium sensitive to macromolecular content using inhomogeneous magnetization transfer (ihMT) ^3,4 on different cardiac samples from sheep’s left ventricle.

Materials & Methods

Samples
Three samples from the anterior left ventricle of three different sheep’s hearts (all female, details in table 1), which include several free-running PF (red arrows in figure 1.a-c) were chosen. After euthanasia, samples were fixed in 4% formaldehyde containing 0.1% of GD DOTA (Guerbet, France) of total volume. For data acquisition, each sample was immersed in perfluoropolyether (figure 1-d) and the temperature was kept at 36±1 °C monitored by a temperature probe (SA Instruments, NY) connected to an external computer.
MR acquisition
All experiments were performed at 9.4T/30cm (Bruker Biospin MRI, Ettlingen Germany) with a cylindrical transmit coil (87 mm intern) coupled with a 4 arrays reception coil. A 2D ihMT-RARE sequence with frequency-alternated pulses for the dual-offset saturation ^4,6 was used to acquire single offset and dual-offset MT weighted images with the saturation parameters listed in table 2.
IhMT ratios (ihMTR in %) were calculated as:
$$ihMTR=100\times\frac{(M_{sing}-M_{dual})}{M_{0}}$$
Where M_sing=MT₊+MT_- and Mdual=MT_±+MT_∓.
MT₊ and MT_- refer to the MT weighted images acquired with positive offset saturation and negative offset saturation, respectively and MT_± and MT_∓ refer to MT weighted images acquired with dual offset saturation. For dual offset saturation, the frequency of the pulses was alternated between positive and negative offsets every τ_switch, where τ_switch is defined by τ_switch=N×$$$\triangle$$$t. N is the number of pulses at the same frequency and $$$\triangle$$$t the saturation pulse repetition time. Allowing a large value for N increases the τ_switch value and hence sensitizes the ihMT signal to longer T_1D relaxation ^5,7. IhMTR maps were generated using the script: https://github.com/lsoustelle/ihmt_proc and a custom-made python script was used to define two ROIs manually in PF (Fig. 2, red) and in myocardium (Fig. 2, green). Mean ihMTR values were calculated in each ROI of all samples.

Results & Discussion

For each different frequency offset and τ_switch value, mean ihMTR and standard deviation in myocardium and PF were calculated from the three samples. The decay of the mean ihMTR with increasing τ_switch is shown in Figure 3 for the three different frequency offsets. Interestingly, for the saturation parameters used here, the ihMTR value is consistently higher in the PF than in myocardium for all different offset frequencies. Figures 3.e-f show the contrast between PF and myocardium calculated as contrast=ihMTR(PF)-ihMTR(myocardium), for which the maximum contrast of 1.36% (absolute scale) is observed for offset-frequency of 20 kHz and τ_switch at 4 ms. Note however, the relatively low ihMTR values (~3-5%) and the overlapping of the myocardium and PF standard deviations (shaded areas), indicative of a relatively low SNR.
By increasing the τ_switch, ihMTR decreases in different conductive fibers (at 20 kHz, ihMTR(τ_switch=0.8 ms)=4.6% vs ihMTR(τ_switch=8 ms)=2.8%) and myocardium (at 20 kHz, ihMTR(τ_switch=0.8 ms)=3.8% vs ihMTR(τ_switch=8 ms)=1.7 %). This loss of signal suggests close and fast T_1D relaxation properties for fibers and myocardium.

Conclusion

This study highlighted the potential of generating an MRI contrast between CCS fibers and the cardiac tissue, based on ihMT. A good reproducibility between the three samples has been demonstrated on three different sheep hearts at very high field (9.4T). Future studies will focus on improving the SNR and further optimizing the contrast between the conducting fibers and the myocardium for 3D imaging of free-running and intramural fibers and their segmentation.

Acknowledgements

This study received financial support from the French Government by the National Research Agency (ANR; SYNATRA ANR-21-CE19-0014-01.