Estimation of Water Exchange across the Blood Brain Barrier using Contrast-enhanced ASL
Helen Beaumont1, Aimee Pearson2, Matthias J van Osch3, and Laura M Parkes1

1Centre for Imaging Sciences, University of Manchester, Manchester, United Kingdom, 2Dept of Physics, University of Manchester, Manchester, United Kingdom, 3C.J. Gorter Center for High Field MRI, Dept of Radiology, Leiden University Medical Centre, Leiden, Netherlands

Synopsis

This study investigates the possibility of estimating water exchange across the blood-brain barrier by manipulating the T1 of blood using a gadolinium-based contrast agent, together with pre- and post-contrast Arterial Spin Labelling measurements. Gadolinium lowers the T1 of blood, but not of tissue, allowing the proportions of label in intra- and extra-vascular tissue to be estimated. A Look-Locker readout was used to measure the temporal evolution of the ASL signal at four doses of contrast agent. Even with T1 of approximately 500ms, an ASL subtraction signal was still detected at an inversion time of 2s, indicating that labelled blood water has exchanged with tissue water.

Purpose

Breakdown of the Blood-Brain Barrier (BBB) is implicated in many brain diseases such as MS, stroke and neurodegeneration1. Gadolinium contrast agents and dynamic T1-weighted imaging can be used to measure leakage of Gadolinium across the BBB. However, subtle breakdown is hard to detect2, probably due to the relatively large molecular weight (~500 Da) of the agent. Measurements of endothelial water exchange may demonstrate greater and more reliable BBB differences between healthy brain and early disease. Several MRI sequences have been proposed to measure water exchange3,4 but sensitivity is generally poor5. Here we propose an approach to increase the sensitivity of ASL to water exchange through the introduction of very small doses of Gadolinium contrast agent which alter the T1 of blood. This will increase the sensitivity of the measurement to water exchange, but will reduce the signal to noise ratio due to signal decay during transit. This study investigates the feasibility of this approach using simulations and imaging.

Methods - Simulation

As a simple proof of principle, the Buxton model6 was adapted to allow the T1 of the labelled water within the voxel to vary linearly between that of tissue and that of blood, as a function of the mean extraction fraction E (E=1 for complete exchange for which T1 = T1tissue; E=0 for no exchange for which T1 = T1blood). Model parameters were CBF=60 ml/min/100ml, arrival time=500ms, bolus duration=1000ms, T1tissue=1.3s, T1blood=1.6s, blood-brain partition coefficient=0.9.

Methods - Imaging

One person was scanned on a Philips 3T Achieva system using a Look-Locker ASL sequence7 with 17 inversion times (TI) 120ms apart. STAR labeling was employed with 15cm labelling slab and zero distance between label and imaging slab. Other parameters were: 80 control/label pairs, TR=2500 ms; TE=22ms; flip angle 40 degrees; 3.5 x3.5 x 7 mm voxels with 1mm gap between 3 axial slices. Bipolar ‘vascular crusher’ gradients were added to dephase fast flowing spins and so remove large vessel signal, increasing the sensitivity to exchanged water. A measurement of T1blood was made using the ‘Varela’ technique8 in a single slice through the sagittal sinus. Parameters were as described in the reference, with the collection of 20 averages to improve the precision of the estimate. Measurements were made at baseline and following consecutive contrast agent injections of 0.8ml, 1.6ml and 2.1ml of Dotarem. These were chosen to give a cumulative reduction of T1blood from 1.60s at baseline to 1.06s, 0.63s and 0.41s assuming relaxivity of 3.5 mM-1s-1 and total blood volume of 4.4L. One minute was left following injection before measurement to allow equilibration. Estimation of CBF and arrival time was made by fitting a single compartment model, adapted for Look-Locker readout9, to the global signal from the pre-contrast data, with bolus duration of 1.1s and T1blood of 1.6s. The post-contrast ASL data were fitted for T1 using the values for CBF and arrival time estimated from the baseline data.

Results - Simulation

Figure 1 shows the simulated ASL signal difference in the case of no contrast agent (a) and contrast agent that reduces T1blood to 0.7 s(b). In the latter case it can be seen that there is a large dependence on E, however the signal is more than halved compared to the no-contrast signal.

Results -Imaging

T1 measurements (the mean from a small region over the sagittal sinus) were 1.33s, 1.18s, 0.51s and 0.45s, in approximate agreement with the predicted signal change according to dose. Figure 2 shows the ASL signal difference for the 4 measurements along with the fitted curves. T1 was estimated as 1.6s (fixed), 1.39s, 1.20s and 1.08s. It is clear that signal has not decayed as much as would be expected according to the known blood T1, suggesting that there is substantial exchange of water into the extravascular space where T1 is higher. This is supported by the subtraction images (Figure 3) which show reasonably high signal at TI of 2s after even the highest contrast agent dose.

Discussion

Despite strong shortening of T1 (from dose calculation and confirmed on T1 measurements), ASL signal is maintained at a relatively high TI of 2s, suggesting that water is exchanging into tissue quickly after labelling. Future work will develop a 2-compartment model to estimate exchange.

Conclusion

This approach could provide estimates of water exchange in vivo in a clinically acceptable time.

Acknowledgements

This work is supported by the EPSRC Sensing and Imaging for Diagnosis of Dementias grant (EP/M005909/1).

References

1. B. V. Zlokovic (2008). `The blood-brain barrier in health and chronic neurodegenerative disorders.'. Neuron 57(2):178-201.

2. P. A. Armitage, et al. (2011). `Use of dynamic contrast-enhanced MRI to measure subtle blood-brain barrier abnormalities.'. Magnetic resonance imaging 29(3):305-314.

3. J. A. Wells, et al. (2013). `Measuring biexponential transverse relaxation of the ASL signal at 9.4 T to estimate arterial oxygen saturation and the time of exchange of labeled blood water into cortical brain tissue.'. Journal of cerebral blood flow and metabolism 33(2):215-224.

4. L. M. Parkes & P. S. Tofts (2002). `Improved accuracy of human cerebral blood perfusion measurements using arterial spin labeling: accounting for capillary water permeability.'. Magnetic resonance in medicine 48(1):27-41.

5. J. P. Carr, et al. (2007). `What levels of precision are achievable for quantification of perfusion and capillary permeability surface area product using ASL?'. Magnetic resonance in medicine 58(2):281-289.

6. R. B. Buxton, et al. (1998). `A general kinetic model for quantitative perfusion imaging with arterial spin labeling.'. Magnetic resonance in medicine 40(3):383-396.

7. M. Günther, et al. (2001). `Arterial spin labeling in combination with a look-locker sampling strategy: inflow turbo-sampling EPI-FAIR (ITS-FAIR).'. Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine. 46(5):974-984.

8. M. Varela, et al. (2011). `A method for rapid in vivo measurement of blood T1.'. NMR in biomedicine 24(1):80-88.

9. S. Al-Bachari, et al. (2014). `Arterial spin labelling reveals prolonged arterial arrival time in idiopathic Parkinson's disease.'. NeuroImage. Clinical 6:1-8.

Figures

Figure 1: Simulated ASL subtraction signal with no contrast agent (a) and contrast agent dose that reduces T1blood to approx 0.7s (b). With the contrast agent, there is a large dependence of the signal on the exchange rate, but the overall signal is much reduced from the case with no contrast agent.

Figure 2: Evolution of the ASL subtraction signal at 4 doses of contrast agent. The points show the subtraction signal, and the lines the associated fit using a modified Buxton model.

Figure 3: Images showing the subtraction signal 2s after labelling at 4 doses of contrast agent. It can be seen that even at the hghest dose of contrast agent there is still label remaining.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
1270