Measurement of bolus arrival time and velocity in Circle of Willis using dynamic MR angiography

Isaac Huen¹, Joanne Beckmann¹, Yuriko Suzuki², Maria A Zuluaga³, Andrew Melbourne³, Matthias JP van Osch⁴, David Atkinson⁵, Sebastien Ourselin³, Neil Marlow¹, and Xavier Golay¹

¹Institute of Neurology, University College London, London, United Kingdom, ²Philips Medical Systems, Philips, Tokyo, Japan, ³Centre for Medical Image Computing, University College London, London, United Kingdom, ⁴C.J. Gorter Center for High Field MRI, Leiden University Medical Center, Leiden, Netherlands, ⁵Centre for Medical Imaging, University College London, London, United Kingdom

Synopsis

The properties of the cerebral vasculature are of clinical relevance and thus of interest for investigation. Here an ASL-based 3D angiographic technique (CINEMA-STAR) is used to measure blood arrival times and blood velocities of a bolus labeled in the neck. Results are shown to be in broad agreement with existing methods.

Background

Vasculature is central to cerebral pathology. Aortic stiffness increases with age[1] and is associated with multiple cardiovascular pathologies[2-6] including stroke[3]. Increased vascular stiffness can result in reduced vascular resistance causing damage to sensitive tissues as in the brain[1]. Parameters indirectly related to vascular stiffness such as pulse wave velocity (PWV) correlate with cerebral small-vessel disease, cognitive decline and Alzheimer’s disease[7]. However, these methods potentially reflect the entire vasculature and are therefore not specific to the brain. Measurement of cerebrovascular blood velocities using phase-contrast angiography is a well established technique[8] that provides blood velocities at particular instants. However, this usually only reports on a short length of vessel, without access to the entire vascular pathway. Dynamic MRA, where blood is labelled (typically in the carotid arteries) and then repeatedly imaged in the target slice, makes characterization of the passage of an individual bolus through the vasculature possible. The CINEMA-STAR[9] sequence acquires a 3D volume at high temporal resolution immediately following labelling. Here we propose to use it to gather novel information about blood flowing through arteries. In this work a bolus is labelled, and its blood arrival time (BAT) at each point of the vasculature measured in order to yield distance travelled and velocity of the bolus along a selected blood vessel. Velocities are shown to be in agreement with existing methods.

Methods

Scan protocol

Four 19 year old healthy subjects were scanned on a 3T Philips Achieva system (Philips Medical Systems, Best, NL) under informed consent (REC: 13/SC/0514). Data was acquired using CINEMA-STAR[10] adapted for 3D volume imaging. A 300mm thick slab 20mm inferior to the CoW was labelled using a STAR[11] labelling. After a delay of 20ms, blood arrival in a parallel volume beginning at the carotid arteries was monitored by segmented 3D TFEPI acquisition of 16 phases spaced by 65ms with FOV: 200x200x70mm³; Acquisition matrix: 116x113; Flip angle: 10°; TR =13ms; SENSE factor: 3; TFE factor: 3; EPI factor: 5; TI/ΔTI/final TI: 20ms/65ms/995ms. ECG triggering was employed to ensure blood arrival was imaged starting from the same point in the cardiac cycle throughout. Scan duration was approximately 12 minutes.

Data processing

Unless otherwise specified, image processing was performed using MATLAB R2012b (The Mathworks, Natick, MA). Image processing was based on previous 2D CoW angiography analysis.[12] Signal from control volumes was voxelwise subtracted from label volumes for all 16 phases, yielding signal profiles against time for each voxel. These profiles were corrected for label T₁ decay (Figure 1) using the method employed by van Osch et al.[12].

Gaussian smoothing was applied to the imaging volume prior to analysis (FWHM=1.56mm). BAT was defined as the phase at which signal increased most rapidly and found using a Canny[13] filter to identify the time of the first peak. A mask of the vasculature (Figure 2A) was defined by thresholding regions of greater inflow, and then skeletonized (Figure 2B) using the Skeletonize plugin in ImageJ 1.48v (Wayne Rasband, National Institutes of Health, USA) to obtain the skeleton of the vasculature. BATs were calculated at each point of the skeleton (Figure 2C) by averaging BATs of adjacent voxels.

Distance travelled from labelling plane was first calculated at each point of the skeleton. Individual vessels were then selected using an in-house MATLAB routine. Along the selected vessel, distance travelled was plotted against BAT (Figure 3B). Distance travelled was obtained as a continuous function by fitting a sigmoid function to BAT of the following form:

$$$ r(t) = \frac{A}{B+e^(-Ct)} $$$

Where r(t) is distance travelled, A, B and C are fitted constants and t is BAT. Velocity was found as the derivative of this distance travelled, dr/dt. Distance travelled and velocity were plotted against BAT for the selected vessel (Figure 3B).

Results

Results are shown for an example blood vessel (right MCA) in Figure 3.

Average peak velocity across all 4 subjects was 44.3± 14.3cm/s.

Discussion

Multiphase 3D angiography was used to derive distance travelled and velocity as a continuous function along a selected blood vessel. Blood velocity in the right MCA of 44.3± 14.3cm/s was in broad agreement with previous velocities measured using both phase contrast angiography (59.4±4.1cm/s) [14] and transcranial Doppler (58.7cm/s) [15]. Blood velocity peaks, as expected, around the straight M1 segment (distance 5-8cm in this case), followed by a decrease thereafter as blood undergoes further vascular resistance due to vascular branching and the Windkessel effect.

Acknowledgements

Many thanks to Joanne Beckmann and Helen O’Reilly for providing clinical support for EPICure scans; Matthias van Osch, Yuriko Suzuki and M. Nakamura for development of the initial CINEMA-STAR sequence; University College Hospital radiographers for carrying out scans, especially Alex and Jagadish; and Magda and James at 8-11 Queen Square.

References

1. Mitchell, G.F., et al., Arterial stiffness, pressure and flow pulsatility and brain structure and function: the Age, Gene/Environment Susceptibility - Reykjavik Study. Brain, 2011. 134: p. 3398-3407.

2. Laurent, S., et al., Aortic stiffness is an independent predictor of all-cause and cardiovascular mortality in hypertensive patients. Hypertension, 2001. 37(5): p. 1236-1241.

3. Laurent, S., et al., Aortic stiffness is an independent predictor of fatal stroke in essential hypertension. Stroke, 2003. 34(5): p. 1203-1206.

4. Boutouyrie, P., et al., Aortic stiffness is an independent predictor of primary coronary events in hypertensive patients - A longitudinal study. Hypertension, 2002. 39(1): p. 10-15.

5. Mitchell, G.F., et al., Hemodynamic Correlates of Blood Pressure Across the Adult Age Spectrum Noninvasive Evaluation in the Framingham Heart Study. Circulation, 2010. 122(14): p. 1379-+.

6. Meaume, S., et al., Aortic pulse wave velocity predicts cardiovascular mortality in subjects > 70 years of age. Arteriosclerosis Thrombosis and Vascular Biology, 2001. 21(12): p. 2046-2050.

7. Rabkin, S.W., Arterial stiffness: detection and consequences in cognitive impairment and dementia of the elderly. J Alzheimers Dis, 2012. 32(3): p. 541-9.

8. Enzmann, D.R., et al., BLOOD-FLOW IN MAJOR CEREBRAL-ARTERIES MEASURED BY PHASE-CONTRAST CINE MR. American Journal of Neuroradiology, 1994. 15(1): p. 123-129.

9. Nakamura M, et al. Non Contrast Time-Resolved MRA combining High Resolution Multiple Phase EPISTAR (CINEMA-STAR) in Proc. Intl. Soc. Mag. Reson. Med. 19 (2011). 2011.

10. Nakamura, M., et al., Non Contrast Time-Resolved MRA combining High Resolution Multiple Phase EPISTAR (CINEMA-STAR).

11. Edelman, R.R., et al., Signal targeting with alternating radiofrequency (STAR) sequences: application to MR angiography. Magn Reson Med, 1994. 31(2): p. 233-8.

12. van Osch, M.J.P., et al., Non-invasive visualization of collateral blood flow patterns of the circle of Willis by dynamic MR angiography. Medical Image Analysis, 2006. 10(1): p. 59-70.

13. Canny, J., A computational approach to edge detection. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 1986(6): p. 679-698.

14. Valdueza, J.M., et al., Changes in blood flow velocity and diameter of the middle cerebral artery during hyperventilation: assessment with MR and transcranial Doppler sonography. American Journal of Neuroradiology, 1997. 18(10): p. 1929-1934.

15. Utku, U., M. Gokce, and M. Özkaya, Changes in cerebral blood flow velocity in patients with hypothyroidism. European Journal of Endocrinology, 2011. 165(3): p. 465-468.

Figures

Figure 1: Original signal against time (blue) corrected for T1 decay (green).

Figure 2: Volume mask of vasculature (A); skeletonized (B); blood arrival time calculated along skeleton by averaging of neighbouring arrival times (C).

Figure 3A: Example individual vessel (right MCA) selected in vasculature (A).

Figure 3B: In (B) distance travelled (blue line) and velocity (green line) of blood were plotted against BAT for the selected vessel in 4 subjects (B). Peak velocities were (top left, clockwise): 64.6, 31.1, 42.2, 39.3cm/s.

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

3311