Date of Award

Summer 8-15-2022

Author's School

McKelvey School of Engineering

Author's Department

Biomedical Engineering

Degree Name

Doctor of Philosophy (PhD)

Degree Type



Fifty million people are affected by dementia worldwide. Cerebral small vessel disease (CSVD) often leads to vascular contributions to cognitive impairment and dementia (VCID), the second leading cause of dementia after Alzheimer’s disease (AD), accounting for up to 45% of dementia. White matter hyperintensities (WMH) on fluid-attenuated inversion recovery (FLAIR) imaging and disrupted white matter microstructure on diffusion tensor imaging (DTI) have been widely used as biomarkers to evaluate CSVD.

The pathogenesis underlying CSVD is not well understood. It has been postulated that exposure to risk factors, including hypertension, diabetes, and smoking over years injures small arterioles, impairs autoregulation and leads to chronic hypoxia-ischemia. Elevated oxygen extraction fraction (OEF) is a signature of tissue oxygenation stress. A magnetic resonance imaging (MRI) measured OEF (MR-OEF) will provide direct evidence of tissue oxygenation stress in the development of CSVD. Neuroinflammation may also be a pathophysiological mechanism for CSVD. Systemic serum and cerebrospinal fluid markers of inflammation, including C-reactive protein and matrix metalloproteases, are elevated in patients with CSVD. While systemic biomarkers are indirect measures of neuroinflammation, positron emission tomography (PET) imaging using 11C-PK11195, a radioligand that binds to the 18 kDa translocator protein located on the membrane of microglia and astrocytes, allows measuring regional neuroinflammation directly.

The overarching goal of this dissertation is to develop combined PET and MR imaging methods to investigate cerebral tissue oxygenation stress and neuroinflammation, two critical pathogenic factors for CSVD. Specific method developments and their applications include 1) PET/MR attenuation correction (AC) and PET/MR AC repeatability; 2) repeatability of regional MR-OEF, cerebral blood flow (CBF) measurements in healthy young participants and elderly participants with CSVD; 3) MR-OEF validation using 15O PET imaging; 4) association between tissue hypoxia-ischemia and white matter structural and micro-structural impairments; 5) roles of neuroinflammation and amyloid deposition in the progression of the mixed AD and VCID pathology.

In combined PET/MR imaging, AC is crucial for accurate quantitative PET. However, MR signal does not provide information on electron density needed for PET AC. In this dissertation, we first developed and evaluated a deep learning-based approach to convert MR to pseudo computed tomography (pCT) for accurate PET/MR AC. This approach achieves highly accurate and repeatable PET/MR AC at both regional and voxel levels while taking a short processing time and is well-suited for longitudinal PET/MR clinical studies.

Our lab previously developed a noninvasive MRI approach that can provide voxel-wise OEF maps using standard MRI scanners without ionizing radiation. Thus far, this approach has not been validated directly against the 15O PET measurement in human subjects, and its test-retest repeatability remains unclear. To establish MR-OEF as a fully validated and robust imaging biomarker for CSVD, we aimed to validate this approach on a hybrid PET/MR system that allows simultaneous PET and MR image acquisition and evaluate its test-retest repeatability in healthy young participants and elderly participants with CSVD. Moreover, we compared MR-measured CBF to 15O-H2O PET measurements.

We then investigated the association between chronic hypoxia-ischemia, measured by regional OEF and CBF and white matter injury in elderly participants with and without vascular risk factors using MRI in a cross-sectional study. We found that elevated MR-OEF was associated with greater WMH lesion burden and microstructural disruption (measured by DTI metrics). Our results suggested that chronic hypoxia-ischemia may contribute to CSVD pathogenesis and may be an imaging biomarker which helps identify individuals at risk for CSVD progression.

Finally, we explored the roles of neuroinflammation and amyloid beta (Aβ) deposition (a key pathologic biomarker for AD) in WMH progression and cognitive decline over a decade in participants with mixed AD and VCID pathologies using PET and MRI in a longitudinal study. Our results suggested that neuroinflammation and Aβ deposition may represent two distinct pathophysiological pathways, both of which independently contributed to the progression of cognitive impairment in mixed AD and VCID pathologies. Neuroinflammation, but not Aβ deposition, contributed to WMH volume and progression.


English (en)


Hongyu An

Committee Members

Hongyu An, Andria L. Ford, Abhinav K. Jha, Richard Laforest,

Available for download on Monday, August 26, 2024