Quantification and Visualization of CSF flow in the Cervical Spine using 4D Spiral flow MRI
MJ Negahdar1, Robert Bert2, and Amir Amini1

1Electrical and Computer Engineering, University of Louisville, Louisville, KY, United States, 2Department of Radiology, University of Louisville, Louisville, KY, United States


To determine feasibility of 4D spiral flow in measurement and visualization of CSF flow in the cervical spine, 5 normal volunteers underwent both a 4D spiral flow and a 4D conventional flow. Results indicate that 4D spiral flow achieved highly accurate flow waveforms with a substantial reduction in total scan time.


Phase Contrast MRI is widely used to noninvasively measure blood velocity and flow, in vivo1. 4D flow MRI can derive all velocity components within a 3D imaged volume in a single acquisition resulting in shorter total scan times compared to 3D flow imaging which requires separate 3D scans for each flow direction2. The velocity field can then be used to obtain flow pattern, wall shear stress, vascular compliance, blood pressure, and other hemodynamic information. The translation of 4D phase Contrast MRI has been delayed to clinical practice by its relatively long scan time. Spiral k-pace trajectory has been utilized in 4D flow imaging to reduce scan time3. In this work, we investigated application of 4D spiral flow to imaging of low flow rate CSF flow in the cervical spine of normal volunteers.

Material and Methods

Five normal volunteers were scanned with an axial field of view which spanned C2 to C6 using 4D flow MRI. Two 4D flow MRI techniques were utilized for data acquisition, conventional 4D Cartesian and 4D Spiral (using a stack of spiral trajectories in k-space) [3]. MR imaging was performed on a Philips Achiva 1.5T scanner (Philips Healthcare, Best, NL) using a 16 element NeroVascular coil. Typical scan parameters were as follows: FOV =160*160 (varied with patient size) volume thickness = 8cm, TR = 10 ms, TE = 5.7ms, temporal resolution =79 ms, tip angle = 6o, Venc = 15 cm/s for all 3 directions, in plane resolutions = 1.8*1.8 mm, slice thickness = 4mm, number of cardiac phases = 12, and matrix size=88*88. For 4D spiral flow 36 spiral interleaves has been used to cover the k-space, all parameters were identical other than TR = 12ms, TE = 4.5 ms, number of cardiac phases = 20. All 4D flow results were visualized with GTFlow (GyroTools, Zurich, Swizerland).


Figure 1 shows velocity profiles in axial slices at C2 and C3 levels overlaid on velocity vector plots of a sagittal resliced image at peak systolic and diastole times. Using particle tracking, the motion of each voxel is traceable which is beneficial in patients with CSF flow abnormalities. Figure 2 shows flow waveform averaged in 4 cervical locations (C2-C5) using the two 4D flow techniques in one normal volunteer. Results show good agreement between the two methods however Spiral 4D flow has mild underestimation in flow measurement in comparison with the conventional method. On the other hand, Spiral 4D flow achieves better temporal resolution (20 phases vs. 12 phases) in a ~20% reduced scan time (15:36 minutes vs. 19:06 minutes).


In this abstract, we have reported initial results of using 4D Spiral flow in assessment of CSF flow in normal subjects. Results were compared with 4D Conventional flow and shown to be similar though with substantial savings in scan time.


No acknowledgement found.


1. N. J. Pelc, et al., Magn Reson Q, vol. 10, pp. 125-47, Sep 1994.

2. M. Markl, et al., J Magn Reson Imaging, vol. 36, pp. 1015-36, Nov 2012.

3. MJ Negahdar, et al., Magn Reson, 2015 DOI: 10.1002/mrm.25636.


Figure 1. Flow through the cervical spine in a normal volunteer. Direction of head to foot is bottom to top. Left: Systolic flow showing flow in the subarachnoid space from head towards the foot. Right: Diastolic flow from foot towards the head.

Figure 2. Comparison of 4D conventional flow and 4D spiral flow measured CSF flow waveforms in the cervical spine.

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