Even though renal blood flow accounts for nearly a quarter of cardiac output, the renal medulla operates in an environment with a scant supply of oxygen. The reason for this apparent discrepancy is thought to be threefold. Blood flow to the outer medulla is less than 50% of that received by the cortex. Secondly, countercurrent shunting of oxygen occurs between the arterial and venous system in the vasa recta, significantly reducing the amount of delivered oxygen. Lastly, oxygen consumption is quite high in this region, primarily due to sodium reabsorption in the medullary thick ascending limbs. Active tubular transport processes account for around 80% of renal oxygen consumption and are highly sodium load-dependent. Operating in this environment subjects the renal medulla to an increased risk of hypoxic damage, making it particularly sensitive to intermittent periods of hypoperfusion or hemodilution. Renal hypoxia has been implicated in acute kidney injury and in the initiation and progression of chronic kidney disease. Starting targeted treatment early can slow disease progression, reduce complications and improve quality of life. Understanding renal oxygenation may be critical in understanding the progression of chronic kidney disease as well. Magnetic Resonance Imaging (MRI) sequences based on the blood oxygen level-dependent (BOLD) contrast mechanism have previously been used for following changes in renal oxygenation. However, these measurements are sensitive to other confounding factors in addition to oxygenation. Additionally, renal BOLD studies are prone to bias due to inherent difficulties in segmenting renal tissue. There is a need for an objective, quantitative, and non-invasive measure of renal oxygenation. In this work, methods to increase the clinical utility of BOLD MRI in assessing renal oxygenation are investigated. We explore the use of BOLD data from the entire parenchyma to reduce the subjectivity in analyzing renal BOLD images. To remove the extra confounding factors influencing BOLD changes, a statistical model is presented that provides an estimate of PO2 derived from BOLD data. An invasive method for refining and calibrating the model is investigated that has the ability to independently measure blood and tissue PO2. Finally an alternative technique for acquiring BOLD data that is not influenced by a subject’s hydration level is presented. These findings provide methods that can estimate renal oxygenation, while controlling confounding factors, along with an objective analysis of the parametric maps, and may provide for widespread utility in pre-clinical and clinical settings.

Last modified

• 02/14/2018

Creator

• Jon Thacker

DOI

• https://doi.org/10.21985/N2FD46

Subject

• Biomedical Engineering

Keyword

• Oxygenation
• Kidney
• MRI

Date created

• 2016-01-01

Resource type

• Dissertation

Rights statement

• In Copyright