Predictive Motion Observation and Tracking for Dynamic and Real-Time Cardiac Magnetic Resonance Imaging

Cardiovascular disease is one of the leading causes of death globally. A key diagnostic tool for medical intervention is Magnetic Resonance Imaging (MRI), which provides excellent soft tissue contrast and spatial resolution without employing ionizing radiation. Unfortunately, at the present time, MRI for cardiac investigations has limited real-time and dynamic capabilities. To address these limitations, we propose a novel signal modelling algorithm called predictive motion observation and tracking (PMOT) that ensures cardiac image reconstruction fidelity without sacrificing spatial resolution, temporal resolution, or volume coverage. In Chapter 1, we define dynamic and real-time rapid cardiac CINE MRI. We subsequently identify the inability of established methods to reconstruct subtle but important cardiac structural details and to accurately track irregular cardiac dynamics. To address these limitations, we formulate a system modelling problem that can be solved via statistics to ensure accurate and exquisite cardiac image reconstruction. In Chapter 2, we provide a review of rapid cardiac CINE MRI. In Chapter 3, we derive a predictive signal model (PMOT) capable of representing the underlying cardiac dynamics that can be used to augment a Kalman filtering or a compressed sensing reconstruction framework. We then perform Kalman filter reconstructions on simulated datasets, including datasets with irregular dynamics, to demonstrate convergence and to establish the statistical validity and superiority of PMOT for high-quality cardiac image reconstruction. Lastly, we perform Kalman filter and compressed sensing reconstructions of a clinical dataset to assess the feasibility of PMOT in a clinical setting. In Chapter 4, to advance towards clinical implementation, we derive a rapid training algorithm for PMOT (mPMOT) and a multi-rate Kalman filter to create a computationally efficient cardiac reconstruction framework. Porcine heart reconstructions with mPMOT exhibit superior image quality with significant statistical improvements compared to the literature. Human and porcine heart reconstructions with mPMOT demonstrate superior reconstruction fidelity at high acceleration factors. In Chapter 5, we outline areas of research that would benefit from the superior image quality and modelling of mPMOT. In Chapter 6, we summarize our key contributions, thereby laying a foundation for achieving robust dynamic, and potentially real-time, rapid cardiac CINE MRI.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.