Reading workstation for clinical contrast echocardiography

点击翻译按钮获取中文摘要

Proposal Summary There is increasing appreciation of a syndrome in which patients female patients, present with chest pain due to myocardial ischemia and have a normal or near normal coronary angiogram. Termed coronary microvascular dysfunction (MVD) this disorder is not benign with cardiovascular event rates similar to those with established coronary artery disease. Clinical tools are therefore needed to both identify MVD patients and better understand the mechanisms causing myocardial ischemia. There is evidence that myocardial contrast echocardiography (MCE) provides incremental information in the evaluation of patients with coronary artery disease, myocardial viability, or diseases of the microvasculature. Despite data demonstrating the diagnostic and prognostic benefit of MCE in evaluating patients with MVD, its clinical use has been limited to only a handful of experts in the field, because there are currently no widely available clinical tools to support MCE quantitative analysis and interpretation. The overall aim of this Phase I proposal is to provide clinicians with a new tool to evaluate the myocardial flow-function relationship that is critical to identifying patients with MVD by using echocardiography. We will develop clinical software that can rapidly process MCE data into a standardized, quantitative and easy- to- interpret format. In Aim 1, the power of image averaging and computer aided assessment of radial wall thickening will be used to enhance the current standard of care which relies solely on readers' visual estimation of segmental function. An algorithm will be developed to rearrange the order of images so that images representing the same phase of the cardiac cycle are grouped together. Functional analysis will then be developed using computer-aided tracings of epicardial and endocardial borders. In Aim 2, a software module for quantitative analysis of real-time MCE perfusion will be developed that will incorporate statistical confidence, derived from the performance of image processing algorithms to inform the interpreter about the data strength. Machine learning will be utilized to train and deploy a neural network for the pixel-by-pixel assessment of myocardial perfusion. In Aim 3, we will combine myocardial perfusion and function modules into a novel, perfusion-function mode of imaging (PF-mode). This new mode will be applied to an archival sample of clinically diagnosed MVD cases to demonstrate the feasibility to detect abnormalities in the myocardial flow-function relationship. The composite PF-mode will include a cine-loop rendered for one cardiac cycle where parametric images (perfusion) are superimposed over averaged ultrasound images with an overlay of graphic representation of wall thickness (function). This novel mode of imaging provides the means to diagnose MVD in a single clinical study.