Related Experiment Videos
Voronoi polyhedra analysis of optimized arterial tree models
Rudolf Karch1, Friederike Neumann, Martin Neumann
1Department of Medical Computer Sciences, University of Vienna, Spitalgasse 23, A-1090 Wien, Austria. rudolf.karch@univie.ac.at
Annals of Biomedical Engineering
|May 22, 2003
Summary
This study analyzed Voronoi polyhedra in arterial trees, finding their spatial distribution differs significantly from random models. These findings offer insights into blood flow and vascular network optimization.
Area of Science:
- * Computational modeling
- * Biophysics
- * Vascular biology
Background:
- * Arterial tree models are crucial for understanding blood flow and tissue perfusion.
- * Voronoi polyhedra (VP) can represent the space supplied by terminal arterial segments.
- * Constrained constructive optimization (CCO) is a method for growing these models.
Purpose of the Study:
- * To analyze topological and metric properties of VPs in CCO-generated arterial trees.
- * To compare VP spatial distributions with random point distributions.
- * To investigate scaling relationships between vascular structures and supplied tissue volume.
Main Methods:
- * Generation of arterial tree models using the constrained constructive optimization (CCO) method.
- * Analysis of Voronoi polyhedra (VP) properties: number of faces (Nf), volume (V), surface area (S), face area (A), and asphericity (alpha).
- * Comparison of VP distributions against random point distributions and between models optimized for different criteria (intravascular volume vs. segment length).
- * Study of scaling laws for intravascular blood volume and arterial cross-sectional area with supplied tissue volume (VP volume).
Main Results:
- * VP distributions (Nf, V, S, A, alpha) in CCO models significantly differ from random distributions.
- * Distributions of V, S, and alpha also differ significantly between CCO models optimized for minimum intravascular volume and minimum segment length (p < 0.0001).
- * VP distributions of Nf, V, and S are well-approximated by two-parameter gamma distributions.
- * Observed scaling exponents for intravascular blood volume range from 1.08 to 1.20, and for arterial cross-sectional area is 0.77.
- * Setting terminal flows proportional to VP volumes resulted in a relative flow dispersion of 37% and skewness of 1.12.
Conclusions:
- * The spatial distribution of supply sites in CCO-generated arterial trees is non-random and distinct from random models.
- * The CCO method produces arterial networks with specific topological and metric properties that can be modeled using gamma distributions.
- * Scaling relationships reveal how vascular volume and cross-sectional area adapt to the volume of supplied tissue.
- * Proportional terminal flow assignment based on VP volume provides a realistic representation of flow distribution in arterial networks.