Combined High Resolution, Four Dimensional, CE-MRA and DCE-MRI of Carotid Atherosclerotic Plaque
Jianmin Yuan1, Andrew J Patterson2, Ammara Usman1, Gregory C Makris1, Zhongzhao Teng1, Jonathan H Gillard1, and Martin J Graves1,2

1Department of Radiology, University of Cambridge, Cambridge, United Kingdom, 2Department of Radiology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, United Kingdom


Carotid neovascularization and lumen surface morphology both play important roles in the progression of atherosclerotic plaque. This study uses a high-spatial and temporal resolution dynamic 3D contrast-enhanced (CE) imaging technique to simultaneously acquire and explore the relationship between the two factors. Results demonstrate that carotid luminal stenosis and appearance of ulceration are correlated with the plaque neovascularization. The study shows the ability to perform high-resolution carotid plaque morphology and functional assessment within a reasonable scanning time.


Carotid neovascularization plays an important role in the progression of atherosclerotic plaque, which is strongly associated with plaque inflammation, hemorrhage and instability[1,2]. Dynamic contrast enhanced (DCE) MRI has demonstrated the ability to quantify plaque neovascularization and inflammation[3]. Plaque ulceration is also believed to be a sensitive marker of plaque instability, which showed strong association with plaque rupture, intraplaque hemorrhage and large lipid core[4]. This study uses a high-spatial and temporal resolution dynamic 3D contrast-enhanced (CE) imaging technique to provide simultaneous 4D CE-MRA and Ktrans mapping.


Nine patients (seven males, mean age 67.7 years [range: 58 to 79 years]), eight symptomatic with a recent history of transient ischemic attack (less than 6 month), one asymptomatic who had at least 50% carotid stenosis on screening Doppler ultrasound, were recruited for DCE and multi contrast MRI scanning. Imaging was performed on a 3T system (MR750, GE Healthcare, Waukesha, WI), using a 4 channel phased-array neck coil (PACC, MachNet, The Netherlands). The study protocol was reviewed and approved by the local ethics committee and written informed consent was obtained.

Scan protocol

3D time-resolved imaging of contrast kinetics (TRICKS) was performed to obtain both DCE and CE-MRA at the same time with the following parameters: TR/TE: 3.9/1.5ms, flip angle= 20°, FOV = 140*140mm, matrix = 224*224, 44 coronal slices with slice thickness = 1.4mm. The imaging slab was centered on the carotid bifurcation. Acquisition time was 6min23s, to obtain 30 temporal phases with a time interval of 10.6s. Coincident with the third phase, a bolus of 0.1mmol/kg Gd-DPTA (Gadovist, Bayer Schering, Berlin, Germany) was administered using a power injector at the rate of 3mL/s followed by a 20ml saline flush. The CE-MRA data was obtained by subtracting the baseline scan from the multi-phase acquisition. In addition, for plaque component determination the following 3D sequences were performed: pre- and post-contrast T1w DANTE-prepared fast spin echo, variable flip angle T1 mapping, time of flight and direct thrombus imaging (DTI). Total scanning time was ~37mins.

Image analysis

DCE images were reformatted into the axial plane resulting in a 0.7mm slice thickness. Plaque wall and lumen boundary at each slice were manually drawn on the pre-contrast T1w image. Quantitative DCE analysis was performed using an established pharmaco-kinetic (PK) modeling approach[1], which uses a two-compartment Patlak model to generate a “vasa vasorum image (VVI)” showing partial plasma volume (vp) in red and transfer constant (Ktrans) in green. Adventitial measurements were calculated by averaging all the pixels along the wall boundary. Plaque measurements were calculated by averaging all the pixels between the wall and lumen boundary. The PK parameters of each plaque were calculated as the mean value across the slices. Plaque surface morphology was classified as either ulcerated or smooth using the CE-MRA images[4]. Luminal stenosis was measured for each plaque according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria[5]. Statistical analysis was performed using R. A Student’s t test was used to compare the difference between smooth versus ulcerated plaques. Pearson’s correlation was used to study the correlation between PK parameters and luminal stenosis.


Two out of 18 arteries were excluded due to poor image quality. Mean luminal stenosis was 44% [range 15 to 65%]. Nine of the plaques were classified as ulcerated and seven were classified as smooth. Both adventitial and plaque Ktrans in ulcerated plaques were significantly higher compared with smooth ones (Figure 2, adventitial Ktrans: 0.085±0.014 vs. 0.063±0.014, p = 0.009, plaque Ktrans: 0.070±0.014 vs. 0.054±0.014, p = 0.049). A positive correlation between adventitial Ktrans/vp and degree of stenosis were observed (Figure 3, r = 0.53, p = 0.03 for adventitial Ktrans, r = 0.60, p = 0.01 for adventitial vp, respectively).

Discussion and conclusion

This study for the first time uses a 4D CE acquisition sequence to acquire high spatial and temporal resolution DCE and CE-MRA of the carotid artery allowing for simultaneous morphological and functional assessment of carotid atherosclerosis. The results show that adventitial and plaque Ktrans were higher in ulcerated plaques compare with smooth ones. Adventitial Ktrans and vp were also positively correlated with the luminal stenosis. These results demonstrate that there is a relationship between lumen surface morphology and plaque neovascularization. This combination of sequences illustrates the ability to perform 3D high resolution carotid plaque morphology and functional assessment in a clinically feasible scan time.


The Vasa Vasorum Image tool was kindly provided by Dr William Kerwin of the Vascular Imaging Lab of the University of Seattle. We acknowledge funding from NIHR BRC.


[1] Kerwin W et al., Magn Reson Med, 2008, 59: 507-514. [2] Qiao Y et al., Am J Neuroradiol, 2012, 33:755-760. [3] Chen H et al., Quant Imaging Med Surg, 2013,3(6):298-301. [4] Lovett J et al., Circulation, 2004, 110:2190-2197. [5] Barnett H et al, N Eng J Med, 1991, 325:445-453.


Figure1. (A) shows the CE-MRA, 3D DANTE-prepared T1w FSE and VVI for an ulcerated (white arrow) plaque. T1w image shows a large intraplaque haemorrhage (blue arrow). The VVI shows the high Ktrans region around the ulceration. (B) shows a plaque with smooth stenosis. No high Ktrans region is observed within the plaque. The Ktrans (green channel) in VVI ranges from 0 to 0.5min-1.

Figure2. (A) and (B) shows the adventitial and plaque Ktrans in smooth and ulcerated plaques.

Figure3. (A) shows the correlation between stenosis and adventitial Ktrans. (B) shows the correlation between stenosis and adventitial vp.

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