ECR 2014 / C-2302
Correlation between Perfusion Micro-Bubble Ultrasound and Perfusion Contrast Enhanced MRI for Carotid Plaque Neovascularization
This poster is published under an open license. Please read the disclaimer for further details.
Congress: ECR 2014
Poster No.: C-2302
Type: Scientific Exhibit
Keywords: Tissue characterisation, Image registration, Segmentation, Contrast agent-intravenous, Ultrasound, MR, Vascular, Neuroradiology brain, Contrast agents
Authors: W. Abutaleb1, M. J. Graves2, J. H. Gillard2; 1Riyadh/SA; Cambridge/UK, 2Cambridge/UK

Methods and materials

Twelve patients with more than 30% carotid stenosis identified by ultrasound underwent both DCE-MR and DCE-US. Three subjects eliminated from this sample due to excessive MRI motion artifact in two cases and very low ultrasound images quality due to obesity in one case.


Here I will elaborate the method used for each modality, followed by the correlation method:


1.1: DCE-MRI Protocol


   The magnetic resonance imaging session takes up to an hour. The patients have been in supine, head first position in a 1.5 Tesla GE™ Magnetic Resonance system. To enhance signal-to-noise ratio, a 4-channel phased-array coil was used, by being placed in close approximation to the carotid artery. To minimise the motion artefact a dedicated vacuum-based head restraint system placed around the head and neck area to fix the head and neck in a comfortable position and to allow close apposition of the surface coils.


  Following an initial three-plane localizer, axial 2D TOF MR Angiography sequence was performed to locate the carotid bifurcation and the region of maximum stenosis. Three millimetres thick axial non-gated FS T1W, T2W, PDW and pre and post contrast spoiled gradient-recalled echo (SPGR) images were simultaneously acquired at four locations, centred on the maximum stenosis level. 


  For the perfusion, the SPGR images were obtained at 51 time points separated by a repetition interval of 5.5-seconds intervals. Coincide with the third phase in the sequence 1mmol kg-1 Omniscan™ was injected at a rate of 2 mL s-1 via a power injector. Relevant SPGR imaging parameters were: TR = 162.26 ms, TE = 2.92 ms, imaging frequency = 63.86, thickness = 3 mm, spacing between the slices = 3, number of averaging = 1, percent sampling = 100, percent phase field of view = 100, flip = 75°, field of view = 20 × 20 cm, matrix = 512 × 256, and pixel bandwidth = 488.281.

1.2: DCE-MR Image Analysis


  DCE-MR will be analysed to produce perfusion measurements within the blood vessel wall over time. I am using Kinetic modelling to quantify the neovascularisation present as demonstrated by Kerwin et al (Kerwin, 2006). In this model vasa vasorum image generated by the dynamic contrast enhanced DCE-MRI with kinetic remodel. Mainly I used these steps to get the Ktrans values (Fig. 4): boundaries drawn on a conventional contrast define a region of interest for processing. Within this region, DCE-MRI images are run through the Kalman Filtering Registration and Smoothing algorithm, and then a blood curve is automatically extracted and used for kinetic modelling. Finally, the boundaries are mapped to the resulting vasa vasorum image and Ktrans is measured in the wall (Kerwin, 2008).

Fig. 4: Illustration of the processing steps in measuring Ktrans
References: Kerwin W, Oikawa M, Yuan C, Jarvik G, Hatsukami T: MR imaging of adventitial vasa vasorum in carotid atherosclerosis. Magn Reson Med 2008, 59(3):507-514.


   Ktrans is used as a quantifying value of the rate of the delivery of the contrast agent from plasma into extravascular extracellular space (Kerwin, 2003). These values are proportional to the percent of neovascularisation (Kerwin, 2006). The mean values of the 4 locations plaque Ktrans within each artery wall were calculated. Values higher than 0.05 min-1 was considered “high” and bellow or equal 0.05 min-1 was considered “low”.



2.1: DCE-US protocol


Imaging was performed with Toshiba™ ultrasound system, using an 8MHZ transducer. First, the carotid bifurcation toward the ICA was imaged in the affected side with special care to identify the maximum stenosis areas plaques by B-mode and Doppler if needed. Then the preset real-time, carotid contrast-specific imaging software with a low mechanical index was switched on and image settings adjusted to maximize contrast signal visualization.

The patients then received contrast-enhanced ultrasound imaging, with special attention to the previously identified lesions.


SonoVue™UK(Bracco,UKlimited) was used as a bubble contrast enhancement agent. A 2.5 cc bolus of SonoVue™ was injected and flushed by 5mm normal saline. CHI images have been acquired for about 1-3 minutes coinciding with the bolus. The studies were digitally stored for later analysis either on or off site. 


3.7: DCE-US analysis


  Special software generated to analyse the intensity of microbubble concentration, by obtaining the level of dB-Enhancement power of outlined area of interest versus the time curve. A Toshiba™ ultrasound machine and/or a portable Toshiba™ workstation has been used for that purpose. 


  For each patient, two regions have been drawn; one region is the plaque mapped in the area where the maximum stenosis estimated or the best visualised plaque. A second region of interest indicates controlled area was drawn in the lumen (Fig.5). The stored images have been evaluated categorizing the carotid plaque, delineating the region of interest and in the end performing the software analysis of microbubbles concentration within the regions of interest and recording intensities values versus the time of the scan (Fig.6) .


Fig. 5: DCE-US imaging analysis (a): two regions of interests have been drawn; one region is the plaque mapped (T). A second region indicates controlled area was drawn in the lumen
References: Wafa Abutaleb, Department of Radiology, University of Cambridge - Cambridge/UK
Fig. 6: DCE-US images analysis (b): the automated intensity vs. time curve of the two regions of interest, the plaque in pink and the controlled in blue. X-axis is the time, Y-axis is the Intensity.
References: Wafa Abutaleb, Department of Radiology, University of Cambridge - Cambridge/UK


  Subsequently I calculated (manually) peak intensity (PI), basal intensity (BI) for all the region of interests. The enhancement intensity (EI) measured by subtracting BI from PI. Lastly I normalised the data by dividing the EI of the plaque on the EI of the Lumen to measure the EI ratio. The values of EI ratios above 4.00e-02 considered "High", and bellow it considered "Low".


3: Statistical analysis: 

The correlation between the non-parametric variables (high/low) of the results have been measured using non-parametric correlation (Spearman r) with no Gaussian assumptions. We measured P value for the significance.


2 clicks for more privacy: On the first click the button will be activated and you can then share the poster with a second click.