Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article

Degree

Master of Engineering Science

Program

Biomedical Engineering

Supervisor

Goldman, Daniel

2nd Supervisor

Frisbee, Jefferson C.

Co-Supervisor

Abstract

The microcirculation plays a critical role in tissue blood flow distribution and is thus a topic of importance for understanding organ pathophysiology. As an alternative to experimental investigations of microvasculature, this thesis introduces a computational algorithm based on constrained constructive optimization (CCO) which aims to generate visually and statistically realistic branching arteriolar network architecture in healthy skeletal muscle tissue. The algorithm includes a list of user-specified adjustable model parameters to generate networks characteristic of different skeletal muscle tissues. Geometric (including mean vessel diameters, lengths, and numbers of bifurcations per order, Horton’s Law ratios, and fractal dimension) and hemodynamic (Murray’s Law exponent and hematocrit) properties of the generated networks matched experimental values from literature when compared for validation. The resulting algorithm is a valuable tool for investigating network architecture and blood flow in various skeletal muscles.

Summary for Lay Audience

Microvasculature refers to the smallest vessels of the cardiovascular system which embed and distribute blood flow to tissues. As the direct point of contact between the cardiovascular system and tissues, the microvasculature is a topic of focus in research on tissue pathophysiology. Experimental investigations into microvascular networks are often labor-intensive, due to the large number and varying sizes of microvascular vessels, and difficult to collect due to limitations in modern-day technology. Additionally, individual and tissue-specific differences in microvasculature make experimental findings difficult to generalize. Therefore, computational modeling of vessel networks has been explored as an alternative in microvascular research which may circumvent experimental limitations. In this thesis, we present a computational algorithm which was produced with two main goals: 1) the algorithm generates networks comparable to real life physiology, and 2) the algorithm may be adjusted to generate networks applicable to different tissues and experimental conditions. To achieve the first goal, the algorithm was written based on constrained constructive optimization (CCO), a popular algorithm for computationally generating visually and statistically realistic vessel network architecture. To achieve the second goal, the algorithm was made to rely on user-adjustable geometric and hemodynamic parameters which will alter the generated networks. In order to validate that algorithm accomplishes these two goals, first, networks were generated with user-adjustable parameters set to fit properties from different experimental datasets of healthy skeletal muscle arteriolar networks. Then, vessels within networks were labelled using the Strahler’s and centrifugal ordering schemes, two common methods for grouping vessels of similar geometric/topological properties. Using these two ordering schemes, geometric and hemodynamic properties of vessels may be calculated over multiple levels within the generated networks. To evaluate how statistically realistic and adaptable the generated networks are, their geometric and hemodynamic properties were matched and compared, based on their user-adjustable parameters, to different experimental datasets of healthy skeletal muscle arteriolar networks. Results demonstrated that geometric and hemodynamic properties of the generated networks were similar to experimental values from different healthy skeletal muscle arteriolar network datasets. Notably, though it is commonly conceived in literature that both the Strahler’s and centrifugal ordering schemes group vessels of similar properties, and are hence interchangeable, these ordering schemes provided different vessel groupings when applied to identical networks. In conclusion, the resulting CCO-based algorithm has been proven to generate realistic and adaptable networks based on application to various experimental skeletal muscle datasets, making it a valuable tool for investigations into skeletal muscle microvascular architecture and blood flow. It may also act as a good reference for future work in developing adaptable algorithms that can generate realistic networks for different vascular territories.

Creative Commons License

Creative Commons Attribution-Noncommercial 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License

Download

Share

COinS