EVALUATION OF EARLY TUMOR ANGIOGENESIS USING ULTRASOUND ACOUSTIC ANGIOGRAPHY

Public Deposited

Downloadable Content

Download PDF
Request Version for Screen Reader
Last Modified
  • March 20, 2019
Creator
  • Shelton, Sarah
    • Affiliation: School of Medicine, UNC/NCSU Joint Department of Biomedical Engineering
Abstract
  • Cancer angiogenesis is a feature of tumor growth that produces disorganized and dysfunctional vascular networks. Acoustic angiography is a unique implementation of contrast-enhanced ultrasound that allows us to visualize microvasculature with high resolution and contrast, including blood vessels as small as 100 to 150 micrometers. These angiography images can be analyzed to evaluate the morphology of the blood vessels for the purpose of detecting and diagnosing tumors. This thesis describes the implementation, advantages, and disadvantages of acoustic angiography and evaluates tumor vasculature in a pre-clinical cancer model. Measurements of tortuosity and vascular density in tumor regions were significantly higher than those of control regions, including in the smallest palpable tumors (2-3 mm). Additionally, abnormal tortuosity extended beyond the margin of tumors, as distal tissue separated from the tumor by at least 4 mm exhibited higher tortuosity than healthy individuals. Vascular tortuosity was negatively correlated to distance from the tumor margin using linear regression. Analysis of full images to detect tumors was performed using a reader study approach to assess visual interpretations, and quantitative analysis combined tortuosity with spatial relationships between vessels using a density-based clustering approach. Visual assessment using a reader study design resulted in an area under the receiver operating characteristic (ROC) curve of approximately 0.8, and the ROC curve was significantly correlated with tumor diameter, indicating that larger tumors were detected more accurately using this approach. Quantitative analysis of the same images used a density-based clustering algorithm to combine vessels in an image into clusters based on their tortuosity (using 2 metrics), radius, and proximity to one another. In tumors, highly tortuous vessels were closely packed, forming large clusters in the analysis, while control images lacked such patterns and formed much smaller clusters. Therefore, maximum cluster size was used to detect tumors, achieving an area under the ROC curve of 0.96. Finally, superharmonic molecular imaging was used to image targeted microbubbles with higher contrast to tissue ratios than conventional molecular imaging. These molecular images were combined with vascular acoustic angiography images to begin to relate the expression of endothelial markers of angiogenesis with vascular features such as tortuosity.
Date of publication
Keyword
DOI
Resource type
Rights statement
  • In Copyright
Advisor
  • Dudley, Andrew
  • Aylward, Stephen
  • Lee, Yueh
  • Gallippi, Caterina
  • Dayton, Paul
Degree
  • Doctor of Philosophy
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2017
Language

Relations

Parents:

This work has no parents.

Items

Thumbnail Title Date Uploaded Visibility Actions
file details Shelton_unc_0153D_17408.pdf 2019-04-11 Public Download