1356502 - Towards Cardiopulmonary ASL perfusion imaging at 0.55T using VS-ASL bSTAR

Contemporary 0.55T MRI systems provides a great promise for improved cardiopulmonary imaging largely due to improved B0 and B1+ homogeneity and reduced SAR. Recently, a 3D balanced steady-state free precession sequence, referred to as bSTAR, has been proposed for free-breathing thoracic imaging¹. This sequence maximizes signal from lung parenchyma and myocardium with different FAs and has provided high-resolution respiratory-resolved volumes with reduced banding artifacts and is expected to provide further reduced banding artifacts at low field due to improved B0 homogeneity.

 Arterial spin labeling (ASL) is a non-contrast perfusion imaging technique that uses inverted blood spins as an endogenous tracer and can be a risk-free alternative to gadolinium-based perfusion imaging². Cardiac ASL in humans has been demonstrated with spatially selective labeling at higher fields³ but spatially selective labeling is limited at low field because a significant amount of perfusion signal is lost while traversing from the tagging location to the target tissue due to faster T1 relaxation. Therefore, velocity selective (VS) labeling⁴ is ideally suited.

In this work, we explore bSTAR imaging in conjunction with VS labeling for whole-heart cardiac and pulmonary ASL at 0.55T. As an initial step, we demonstrate ECG-gated, free-breathing bSTAR in a healthy volunteer, with settings suited to maximizing myocardial and lung parenchyma signal.

Methods:

Experimental Methods: All experiments were performed on a whole-body 0.55T scanner (prototype Aera XQ, Siemens Healthineers). An ECG-triggered, respiratory self-navigated sequence with VS labeling, and bSTAR imaging, was implemented with the Pulseq framework.

Labeling: BIR-4 based VS labeling was implemented⁵. To avoid incident labeling of moving myocardium, velocity labeling was applied at diastole when the movement of heart is minimal and speed of intracoronary blood is highest. A steady-pulsed labeling scheme was used to derive the perfusion signal into steady state.

Imaging: A spiral phyllotaxis pattern for radial half-spokes was used. Respiratory self-gating was achieved with superior-to-inferior (SI) projections, each acquired every RR interval in the beginning of an imaging module. Image reconstruction was performed with XD-GRASP implemented in BART.

Results:

Figure 2 shows a simulated steady state signal as function of bSSFP flip angles for myocardium and lung parenchyma.

Figure 3 shows a representative axial slice of a reconstructed volume at end-expiratory state. No visible banding artifacts appear within the lung and heart.

Conclusion:

We have demonstrated feasibility of cardiopulmonary imaging with bSTAR and the potential for integration in an ASL experiment using VS-labeling. Additionally, this was implemented using the vendor-neutral and open-source Pulseq framework. Several aspects require further exploration, including optimization of VS labeling to minimize eddy-current effects, and to test this approach in larger numbers of subjects.