Characterization of coronary plaque in intravascular ultrasound using histological correlation

Atherosclerosis is a degenerative arterial disease that leads to the gradual blockage of vessels due to plaque formation or acute ischaemic events such as plaque rupture. A thorough understanding of plaque morphology is necessary in the determination of factors underlying coronary artery disease. Intravascular ultrasound (IVUS) represents a diagnostic technique that provides tomographic visualization of coronary arteries. Ultrasound reflects sound at the interfaces between tissue of different acoustic refractive indices theoretically implying that various components within an ultrasound image should be distinguishable. The aim of this study is to classify plaque lesions using advanced digital image processing. Diseased samples were obtained from excised human coronary arteries at autopsy. Intravascular ultrasound images were acquired using an HP Sonos clinical IVUS imaging system and 3.5 F, 30 MHz catheters. Regions of interest were selected within the IVUS images with plaque types determined from the corresponding histological sections. Second order statistical texture features were computed for each region of interest after which tissue classification was performed using discriminant analysis. Preliminary results show that classification and recognition of plaque type is possible for IVUS imaging in vitro. Various plaque types respond differently to clinical interventions and proper classification will therefore be invaluable in tracking the progression, regression, and treatment of coronary artery disease.