Article Text

076 Dual source computed tomography for characterisation of coronary atherosclerotic plaques. A comparison with virtual histology intravascular ultrasound
  1. D R Obaid1,
  2. P A Calvert1,
  3. J H F Rudd1,
  4. D Gopalan2,
  5. M R Bennett1
  1. 1University of Cambridge, Cambridge, UK
  2. 2Papworth NHS Foundation Trust Hospital, Cambridge, UK


Background Non invasive plaque characterisation is important for risk stratification. Dual Source CT (DSCT) offers a significant improvement in temporal resolution over single source CT and may be a non invasive method to classify plaque components. Although CT plaque images have previously been compared with greyscale IVUS they have not been validated against virtual histology intravascular ultrasound (VH-IVUS), a technology which involves spectral analysis of IVUS radiofrequency signals. VH-IVUS is the gold standard for in vivo plaque classification, with proven validation against histology. Our aim was to show that DSCT could distinguish between plaque components as defined by VH-IVUS on the basis of their attenuation values in Hounsfield units (HU).

Methods Twenty-one patients underwent DSCT and VH-IVUS prior to coronary stenting. Twenty-six plaques with VH-IVUS plaque burden >40% were chosen. Plaques were either homogenous for fibrous plaque or contained large confluent areas of calcified plaque or a necrotic core. The VH- IVUS images were co-registered with the corresponding CT image by an unblinded observer. Accurate matching of lesions was performed by using measurements of distance from the coronary ostia. The images were orientated using fiduciary markers (side branches or characteristic calcifications). Multiple regions of interest (ROI) were placed within the plaque components and lumen on the cross sectional CT images in areas that had been pre-classified on the VH-IVUS images (Abstract 76 Figure 1). The density of these ROIs were measured (expressed in HU) and assigned to the relative plaque type or lumen.

Abstract 76 Figure 1

Co-registered VH IVUS (left) and DSCT (right) demonstrating ROI sampling of necrotic core (top) and calcified plaque (bottom).

Results Attenuation values for 171 ROIs (necrotic core 28, fibrous plaque 38, calcified plaque 29 and lumen 76) were measured from the 26 plaques. Mean and standard deviation attenuation values were 35+/−17 HU, 128+/−49 HU, 719+/−237 HU and 362+/−51 respectively (Abstract 76 Table 1). There were statistically highly significant differences in the attenuation values among the four groups (necrotic core, fibrous plaque, calcified plaque and lumen) by nonparametric Kruskal–Wallis test (p<0.0001) (Abstract 76 Figure 2).

Abstract 76 Table 1

Attenuation values derived from the ROIs placed in the plaque components and lumen expressed as means with ranges and standard deviation

Abstract 76 Figure 2

Box-and-whiskers plot of the CT density of the IVUS defined plaque types and lumen showing statistically significant differences between 4 groups (P<0.0001 Kruskal-Wallis).

Conclusions VH-IVUS-defined plaque components have characteristic attenuation values that differ significantly enough to be differentiated by DSCT. This classification may facilitate non-invasive coronary plaque characterisation using DSCT for both therapeutic drug trials and assessment of vascular risk.

  • dual source computed tomography
  • virtual histology
  • intravascular ultrasound

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.