Date of Award

Spring 5-15-2020

Author's School

Graduate School of Arts and Sciences

Author's Department


Degree Name

Doctor of Philosophy (PhD)

Degree Type



Molecular oxygen (O2) is vital for efficient energy production and improper oxygenation is a hallmark of disease or metabolic dysfunction. In many pathologies, knowledge of tissue oxygen levels (pO2) could aid in diagnosis and treatment planning. The gold standard for pO2 measures in tissue are implantable probes, which are invasive, require surgery for placement, and are inaccessible to certain regions of the body. Methods for determining pO2 both non-invasively and quantitatively are lacking. The slight paramagnetic nature of O2 provides opportunities to non-invasively characterize pO2 in tissue via magnetic resonance (MR) techniques. As such, O2 can be treated as a weak endogenous contrast agent for longitudinal relaxation and, therefore, the measured longitudinal relaxation rate constant (R1) is directly proportional to pO2. Precise characterization of R1 in the absence of oxygen (R1,0) and the relaxivity of O2 (r1) would allow for an R1-based pO2 measurement. Additionally, the effective transverse relaxation rate constant (R2*) in tissue is strongly affected by the magnetic susceptibility effects of deoxyhemoglobin within the vasculature. Many forms of placental dysfunction, e. g. , pre-eclampsia and intrauterine growth restriction, are proposed to be caused by altered vasculature development within the placenta, potentially leading to adverse outcomes for both mother and fetus. Improved biomarkers of placental function would aide in optimal timing for early delivery once the placenta can no longer support fetal development. The objectives of this dissertation were to: 1) investigate the efficacy of an R1-based method of pO2 quantification in a tissue surrogate; and 2) apply MR methods of monitoring pO2 in tissue and O2 within the vasculature in mouse models of disease and insufficiency to assess placental development and function. For the first goal, Bayesian probability theory-based model selection was used to evaluate potential models of longitudinal relaxation in in vivo tissue and an in vitro tissue surrogate, crosslinked bovine serum albumin (xBSA). xBSA was then used to investigate physiologic confounds to an R1-based method of pO2 quantification, including temperature, pH, and protein concentration, and R1,0 and r1 were determined. For the second goal, mouse models of both placental insufficiency and Zika virus infection during pregnancy were monitored in late gestation for changes in volume, R1, and R2* at baseline and with a gas challenge to assess the placental response to an altered environment. It was found that 1) both in vivo and xBSA relaxation data are best fit with a biexponential model and, therefore, xBSA is a good surrogate for tissue, in terms of longitudinal relaxation; 2) physiologic confounds to an R1-based method of pO2 quantification exert considerable affects upon measured R1 and must, therefore, be precisely controlled or accounted for; 3) placental volume, R2*, and change in R2* due to a breathing gas challenge hold promise as biomarkers of placental development and dysfunction. These findings suggest that an R1-based method for pO2 quantification in vivo is likely not feasible on a routine basis due to the small water relaxivity of pO2 and confounds to the analysis due to relaxation effects of tissue pH, temperature, and protein concentration, but MR methods could provide much needed information regarding placental function in high risk pregnancies and warrants further investigation.


English (en)

Chair and Committee

Anders E. Carlsson

Committee Members

Joseph J. Ackerman, Joel R. Garbow, Shankar Mukherji, Jeffrey J. Neil,