Characterisation of carotid atherosclerotic plaque composition in vivo using 3D B-mode ultrasound

Embolisation from carotid atherosclerotic plaque is a major cause of ischaemic stroke. Studies have shown that it is not only the degree of narrowing of the carotid artery but also the constituents of the plaques which are important factors in determining whether a plaque ruptures and embolises. A reliable and repeatable method has been developed to acquire in vivo data of carotid plaques using 3D B-mode ultrasound. An artificial neural network was enabled using MATLAB software programs to characterise each of the plaque constituents into one of five classes based on statistical and textural parameters of the B-mode images. Comparison of the method with histology processing post surgery found that characterisation of plaque constituents in vivo is not readily achievable. The method, however, was found to be repeatable in serial scans and was used to monitor the effect of two different lipid lowering drug treatments over a 12 week period. Two patients who underwent LDL apheresis treatment showed reduction in softer class types and increases in the harder class types along with reductions in the overall plaque volumes. Nine patients underwent statin treatment with changes found in 4 of the 9 patients over the period of the study. There are several possible reasons why changes were seen in the apheresis group and only in some of the statin group. The apheresis treatment is more aggressive than the statin treatment, the time period of the study was relatively short and changes in the tissue types are dependant on the initial constituents of the plaque