Non-invasive imaging techniques to detect and characterize infarcted myocardium are critical for myocardial infarction (MI) prognosis and post-MI risk assessment. Many existing imaging techniques (such as T1-mapping) for MI detection require the use of contrast agents, which are not recommended for patients with renal failure. Cardiac diffusion tensor magnetic resonance imaging (cDTI), a non-contrast imaging technique, allows estimating the aggregate cardiomyocyte microstructure in healthy subjects and its remodeling as a result of cardiac diseases. Two main questions are investigated in this thesis: 1) how cDTI based biomarkers can characterize microstructural changes occurring post-MI; and 2) if cDTI based biomarkers can be used to train Machine Learning models to accurately detect, localize, and quantify infarcted tissue. In this study, both \emph{ex vivo} and \emph{in vivo} MRI datasets (e.g., cDTI, T1-weighted, T2-weighted images) were used. These datasets were previously acquired in seven Yorkshire swine subjects with chronic myocardial infarction. To address the first question, microstructural changes were quantified in the border zone and infarct region by comparing diffusion tensor invariants – fractional anisotropy (FA) and mean diffusivity (MD) – radial diffusivity, and diffusion tensor eigenvalues with the corresponding values in the remote myocardium. Furthermore, the Extracellular Volume (ECV) fraction was computed and compared with corresponding native T1 and diffusion quantities values in remote, border, and infarcted regions. In order to understand the increase in diffusion tensor eigenvalues observed experimentally, the diffusion of water molecules in the extracellular space was simulated numerically as a function of ECV. To address the second question, a U-Net based on the Convolutional Neural Network (CNN) architecture was trained using high-resolution \emph{ex vivo} T1-weighted MRI data and \emph{ex vivo} diffusion tensor invariant maps. The infarct segmentation outcomes from the T1-weighted trained model and the diffusion tensor invariants trained model were compared against each other to assess the potential of using cDTI for myocardial infarct detection. The findings of this thesis suggest that diffusion quantities - especially radial diffusivity - can be potential biomarkers to detect myocardial infarction and quantify the microstructural changes occurring post-MI.


If this is your thesis or dissertation, and want to learn how to access it or for more information about readership statistics, contact us at STARS@ucf.edu

Graduation Date





Perotti, Luigi


Master of Science in Mechanical Engineering (M.S.M.E.)


College of Engineering and Computer Science


Mechanical and Aerospace Engineering

Degree Program

Mechanical Engineering; Mechanical Systems


CFE0009312; DP0026916





Release Date

June 2023

Length of Campus-only Access

1 year

Access Status

Masters Thesis (Open Access)