Theses Doctoral

High frame-rate pulse wave imaging for non-invasive characterization of arterial stiffness in vivo

Kemper, Paul

Recent studies have indicated that vascular stiffness is an important predictor of future cardiovascular disease. Hence, assessment of vascular stiffness would be of interest. Ultrasound is a good modality for assessment of vascular stiffness, due to its hight temporal resolution and non-invasive nature. Using ultrasound, various techniques have been proposedto estimate vascular stiffness, one of them being Pulse Wave Imaging. The ultimate goal of Pulse Wave Imaging is to provide a robust, qualitative and quantitative method to estimate and visualize clinically important parameters and phenomenonfor cardiovascular disease. The objective of this thesis limits itself to 1) expand Pulse Wave Imaging by going beyond just the diastolic stiffness, 2) utilize Pulse Wave Imaging in an atherosclerotic swine model to monitor plaque initiation and progression and 3) improve non-linear stiffness estimation at or near sites of reflections using Pulse Wave Imaging for clinical applications.

In Aim 1, the question pursued was whether Pulse Wave Imaging can be utilized to monitor this non-linear behavior in-vivo. It was observed that in this mouse model, the compliance at diastolic pressure did not change significantly, whereas the compliance at end-systole did. Thus suggesting that Pulse Wave Imaging was able to monitor a change in non-linear stiffness, and that considering this, might be of importance.

In Aim 2, the ability of Pulse Wave Imaging to monitor disease progression for atherosclerotic disease progression was assessed. Since human studies involve various compounding factors, animal models provide the opportunity to study the ability of methods in a more controlled manner. Swine is a good candidate due to its similarity with humans. To doso, first, the feasibility of Pulse Wave Imaging in swine needed to be assessed. While the cardiovascular system might be similar, various other factors, such as the location and depth of the carotid differs. It was revealed that PWI was feasible in swine and that we were able to generate atherosclerotic lesions within 9-months. Subsequently the ability of Pulse Wave Imaging and Vector Flow Imaging to monitor atherosclerotic progression leading to different type of lesions was assessed. The in-vivo findings were compared with histology and nanoidentation. The results indicated that Pulse Wave Imaging was shown to be able to separate to different disease progression pathways leading to different type of lesions.

Finally in Aim 3, lessons learned from the animal models were attempted to be addressed by developing a more reflection robust approach for localized non-linear stiffness estimation for clinical application. First, improvements were proposed to a previously developed inverse problem approach that can resolve reflections within the field of view by including information from the flow velocity. To expand the approach to include non-linearity and reflections that occur outside the field of view, a physics-based neural network approach was considered. This might be of importance since most plaques are located at sites of significant reflections, such as the bifurcation. Chapter 6 revealed that artificial significant sources of reflections hindered its ability for sub-cm scale localized compliance measurements as indicated by an immediate increase in the number of detected segments after the ligation was induced. The approaches was validated using simulated data and feasibility was shown in in-vivo examples. With new progress, new issues tend to arise.

Finally, the purpose of this sub-aim is to utilize the technique and investigate whether or not it can in fact better differentiate between different clinically relevant groups. The findings revealed no significant improvement concerning the mean compliance estimated, but appeared more robust against outliers when only the plaque segment was assessed.

In conclusion, the results shown in this dissertation indicate that Pulse Wave Imaging is a promising approach to assess non-linear stiffness estimation for monitoring of vascular disease. Furthermore, an new methodology was proposed and feasibility was shown, which could further improve localized and non-linear stiffness estimation at or near sources of significant reflections, and which can be used as a starting point for further development.


  • thumnail for Kemper_columbia_0054D_17723.pdf Kemper_columbia_0054D_17723.pdf application/pdf 5.33 MB Download File

More About This Work

Academic Units
Biomedical Engineering
Thesis Advisors
Konofagou, Elisa E.
Ph.D., Columbia University
Published Here
April 5, 2023