Improved MRI guidance of pediatric catheterization via autonomous multi-beat data synthesis

点击翻译按钮获取中文摘要

Project Summary Cardiovascular catheterization is a minimally-invasive approach to measure hemodynamics and treat abnormalities, such as vascular stenosis. Conventionally, X-ray fluoroscopy guides catheterizations, but it uses ionizing radiation and suffers from poor soft tissue contrast even with the use of exogenous contrast agents. This is problematic in children as they are particularly susceptible to radiation, often require repeat assessments, and can have complex anatomy that is difficult to navigate. MRI guidance provides a non- ionizing alternative that can improve soft tissue contrast and enhance evaluations. In addition to guiding diagnostic catheterization, first-in-human studies 15 years ago demonstrated that MRI could guide cardiac interventions such as coarctation angioplasty. However, clinical adoption since then has been limited by 1) few MRI-safe interventional devices and 2) the relatively poor real-time MRI image quality available during procedures. The number of MRI-safe devices is increasing, and low-field MRI has emerged as a new way to significantly reduce device heating. Yet, poor real-time image quality remains a significant barrier. Despite parallel imaging and accelerated image reconstructions, image quality is constrained by the limited set of MRI samples available for image reconstruction in a single-shot, real-time acquisition. The solution: This proposal aims to leverage multi-beat information to improve real-time MRI image quality during pediatric interventions. The goal is to improve image quality by increasing data available for image reconstruction and improve device visualization by highlighting motion between beats. The central hypothesis is that, in the interventional setting, multi-beat information will improve real-time MRI image quality and device visualization, leading to improved operator confidence and performance. To test this hypothesis, multi-beat’s impact on interventional image quality (Aim 1), impact on device visualization (Aim 2), and clinical utility in pediatric patients (Aim 3) are evaluated. The proposal is supported by a team that includes pediatric cardiologists at the forefront of MRI-guided interventions and prior research experience developing a closed- loop MRI data collection technique which robustly identifies similar heartbeats. The platform was originally designed for a different application – improving non-interventional MRI imaging of adults with arrythmia. Here, it is adapted to explore a new imaging paradigm (interventional imaging) in a new patient population (pediatric patients) for the PI. The work is expected to establish a new approach for real-time MRI-guided cardiovascular interventions by addressing the areas that currently limit image guided confidence: poor anatomic and device visualization. Success would not only improve MRI guidance of cardiovascular interventions in children but also create new opportunities for MRI guidance of cardiovascular and non-cardiovascular (biopsy, drainage, thermal therapy) procedures for adults.