Electronic Theses and Dissertations



Document Type


Degree Name

Master of Science


Biomedical Engineering

Committee Chair

Aaryani Tipirneni-Sajja

Committee Member

Amy Dr. Curry

Committee Member

Carl Dr. Herickhoff

Committee Member

Eddie Dr. Jacobs


Iron overload is an excessive accumulation of iron in the body and can be either inherited or acquired through chronic blood transfusions. Assessment of hepatic iron concentration (HIC) is important in the management and monitoring of iron overload. Despite liver biopsy being the gold standard method for assessing HIC, it is invasive, painful, unsuitable for repeated measurements, and carries the risk of bleeding and infection. Magnetic Resonance Imaging (MRI) methods based on transverse relaxation rate (R2*) have emerged as a non-invasive alternative to liver biopsy for assessing HIC. Multispectral fat-water-R2* modeling techniques, such as the non-linear square (NLSQ) fitting and autoregressive moving average (ARMA) models, have been proposed to provide more accurate assessments of iron overload by accounting for the presence of fat, which can otherwise confound R2*-based HIC measurements in conditions of co-existing iron overload and steatosis. However, the R2* estimation by these multispectral models has not been systematically investigated for various acquisition methods like the multiecho gradient echo (GRE) and ultrashort echo time (UTE) across the full clinically relevant range of HICs. To address this challenge, a Monte Carlo-based iron overload model based on true iron morphometry and histological data was constructed, and MRI signals were synthesized at 1.5 T and 3 T field strengths. This study compared the accuracy and precision of multispectral NLSQ and ARMA models against the monoexponential model and published in vivo R2*-HIC calibrations in estimating R2*. The results showed that, for GRE acquisitions, ARMA and NLSQ models produced higher slopes compared to the monoexponential model and published in vivo R2*-HIC calibrations. However, for UTE acquisitions for shorter echo spacing (≤ 0.5 ms) and longer maximum echo time, TEmax (≥ 6 ms), both multispectral and monoexponential signal models produced similar R2*-HIC slopes and precision values across the full clinical spectrum of HICs at both 1.5 T and 3 T. The results from the simulation studies were validated using phantoms and patient data. Future work should investigate the performance of multispectral models by simulating liver models in coexisting conditions of iron overload and steatosis to investigate simultaneous and accurate quantification of both R2* and fat.


Data is provided by the student.

Library Comment

Dissertation or thesis originally submitted to ProQuest


Open Access