Molecular imaging in atherosclerosis452Ultrasound invariant image features can be associated with histological changes in an atherosclerotic rabbit model

    loading  Checking for direct PDF access through Ovid

Abstract

Purpose

As it has been reported that ultrasound (US) backscattering changes according to tissue microstructure, we evaluated whether exists an association between invariant image features of B-mode ultrasound (US) and histological changes in arteries of rabbits with induced Atherosclerotic Vascular Disease (AVD) at early stages.

Methods

AVD was induced to 20 of a total of 40 New Zealand Rabbits that were fed with a 1% cholesterol supplemented diet for 70 days. Vascular US images were acquired during anesthetic stage, obtaining samples from ascendant aorta. Histological study was performed to quantify collagen, elastin and lipids. B-mode US images were manually delineated by an expert and processed by Automated Quantitative Ultrasound Analysis (AQUA) to extract invariant image features. We evaluated association between US invariant image features and histological quantification, firstly, by partial least squares (PLS) regression and secondly, by computing R squared only within the AVD rabbits to evaluate whether subtle histological changes within AVD rabbits corresponded to consistent changes of invariant image features.

Results

Average and standard deviation of intima-media thickness was 0.4 and 0.2 mm for control and 0.6 and 0.2 mm for AVD. Lipids and collagen increased significantly in AVD rabbits. PLS regression resulted in R squared of 0.88 and R squared only within the AVD rabbits resulted in 0.85 and p < 0.0001.

Conclusions

B-mode US invariant image features can be associated with histological changes and these changes are consistent with subtle histological changes that correspond to different degrees of AVD. This suggests that (1) the well differentiated acoustic properties of lipids compared to elastin, collagen and water produce a contrast in the US images of AVD rabbits that can be robustly measured by invariant image features and that (2) US invariant image features can develop predictive models of intima-media microstructure. Further research will be focused on developing systems to estimate specific microstructural aspects of wall vessel and the associated cardiovascular risk.

Related Topics

    loading  Loading Related Articles