Introduction

Arterial Spin Labelling (ASL) is a non-invasive perfusion technique that by labelling the arterial blood can be used to probe cerebral blood flow (CBF) changes. By applying diffusion gradients, we can probe signal from either the arteries, capillaries or tissue dependent on gradient strength or inversion delay. This information can be useful in Blood Brain Barrier (BBB) permeability studies, where a breakdown in the barrier is associated with increase water exchange into the tissue. We use the high SNR mbASL sequence, which uses adiabatic pulses to label the arterial water. This sequence has shown significantly higher SNR against the standard FAIR sequence in recent studies¹. At current, bipolar gradients are applied around a 180-degree pulse and when the DWASL signal is measured at multiple b-values, a steep drop is seen from the suppressed signal before tending to tissue diffusion². As the signal is suppressed with a small gradient value, these values tend to be noisy & unstable, losing valuable information on the exchange of water in the brain. We propose moving the bipolar gradients to reduce the time between the gradients and increases the gradient strength needed, providing a more stable sequence with the expected drop in vascular signal shifting to a higher b-value. Figure 1 demonstrates the current DWASL sequence and comparison with the modified sequence. The theoretical signal at each b value is demonstrated.

Methods

Scans were performed using a Bruker PharmaScan 7T MRI scanner using a mouse surface coil. A DWASL sequence with gradients around the 180-degree pulse (Sequence A) was compared with the gradients moved to a single side (Sequence B). All other parameters were kept the same as follows: b = 0:100:500 s/mm2. NA = 10. FOV = 25x25mm. CI = 5000ms. TR = 7000ms. TI = 50ms. Np = 20. For Sequence A, Δ = 29.9ms, and Sequence B, Δ = 9.4ms. Small delta = 3.2ms. n=6 CD1 female mice were used for scanning.Data was exported in DICOM format and analysed using in-house MATLAB code. The cortex was used as the ROI region.

Results & Discussion

Figure 2 is a comparison of diffusion signal from a water phantom using both DWASL sequences. This shows that both sequences produce the expected diffusion coefficient of water (2.27 and 2.3x10⁹ s/mm² respectively). Figure 3 compares the ΔM signal from Sequence A and Sequence B. The ASL control image signal is shown which confirms the signal from Sequence A and Sequence B are from flowing blood. The rate of signal suppression has changed for sequence B with the change from intravascular to extravascular signal occurring at b = 300 in Sequence B and at b = 50 in Sequence A, confirming our theory. This is further reflected in the fitting to the bi-exponential model. Comparing the weighting factors for the two sequences, there is a large difference seen between the intravascular and extravascular signals.

Conclusion

In conclusion we have developed a modified DWASL sequence that allows us to probe the intravascular signal further than previously. By changing the time delay between the bipolar gradients and changing their position in the sequence we have been able to create a more stable sequence with the change in signal between the intravascular and extravascular compartments occurring at larger b-values. This information will be vital in probing BBB permeability changes more accurately.

Acknowledgements

Many thanks to Jim Mullin for his technical help in the MRI scanning and to the University of Glasgow and EPSRC for funding this project.