Carotid plaques are one of the commonest causes of neurological symptoms due to embolization of plaque components or flow reduction. The classification of plaques vulnerability is then a relevant clinical issue, and a technical challenge. Recently, several atherosclerotic plaque characterization methods were proposed based on plaque morphology assessed through 2D ultrasound. One of these methods, proposed by Seabra et al [1] presents a measure with clinical significance, known as enhanced activity index (EAI), that the clinician then uses to classify the plaque. The present paper builds upon that work and by using machine learning, proposes an ensemble classifier that shows promising results outperforming both the gold medical standard degree of stenosis and the EAI score. Results are obtained on a real clinical database of 146 plaques. Future work will investigate the predictive performance of the proposed classifier, i.e., how well does the classifier identify stable lesions at high risk of becoming symptomatic.
展开▼
