High field MR carotid vessel wall imaging: reproducibility of five different MR-weightings

Introduction

Magnetic Resonance Imaging (MRI) has emerged as a promising noninvasive imaging modality for the serial assessment of vessel wall thickness in the carotid artery as an early marker of atherosclerosis. For clinical application of this technique, Scan-Rescan reproducibility is paramount. Currently, a multicontrast protocol, including a combination of MR-weightings is used as reference standard for quantitative and morphologic measurements.

Purpose

To investigate Scan-Rescan reproducibility for each of the commonly used weightings analyzed separately. To investigate which of the MR-weightings approximates best the combined multicontrast protocol (reference standard).

Methods

5 healthy volunteers (60% male, mean age=28years) underwent repeated MRI examinations of the left carotid artery with five contrast-weighted scans to image lumen and vessel wall (Table 1). The scan and a rescan were acquired using a 3T (Philips) MRI scanner. A standard phased-array coil with two flexible elements of 14×17cm was used to obtain nine transverse imaging sections of the left carotid artery with identical in-plane resolution (0.46×0.46mm²). Scan-Rescan analysis was performed in the third slice of the imaging stack, representing a slice in the common carotid artery. An example is provided in Figure 1. Manual contour segmentation of the lumen and vessel wall was performed using in-house developed software (VesselMASS). Vessel wall area (mm²) and lumen area (mm²) were assessed by one blinded observer for the different contrast weightings and compared with the rescan acquisition. Furthermore, vessel wall- and lumen areas from the different contrast weightings were compared with the reference standard.

Table 1 Carotid Imaging Protocol at 3T: Scan Parameters

Figure 1

An example of the co-registered contrast-weightings and 3D time-of-flight. The red line on the sagittal view of the carotid bifurcation indicates position of the analyzed slices.

Results

Reproducibility of the repeated assessment was high for all MR-weightings, for both lumen area and vessel wall areas. (Table 2). The Bland-Altman plot for vessel wall area, is shown in Figure 2 for Scan-Rescan reproducibility. Highest reproducibility was found for the T1-TFE and T2-TSE sequences.

Table 2

Figure 2

Bland-Altman Plot: Scan-Rescan Reproducibility (Vessel Wall Area)

T1-TFE showed highest correlation for lumen (r=0.97) and vessel wall area (r=0.90) assessment when compared with the reference standard.

Conclusion

This pilot Scan-Rescan study showed best reproducibility of lumen and vessel wall area assessment for the T1-TFE and T2-TSE weightings. T1-TFE showed highest correlation to the reference standard.