Electronic Thesis and Dissertation Repository

Thesis Format

Monograph

Degree

Master of Engineering Science

Program

Electrical and Computer Engineering

Supervisor

Samani, Abbas

Abstract

Current prostate cancer detection methods can be costly to obtain, such as magnetic resonance imaging, or lack specificity, such as a digital rectal exam. Ultrasound elastography, a method that can be used to develop and test algorithms that output stiffness, strain, and displacement data captured by ultrasound radio frequency readings, offers a potential solution to these challenges. An initial algorithm utilizing dynamic programming and analytic minimization estimates the radial and angular displacements from a pre- and post-compression data set to determine the required material properties. This estimate of displacements is then refined through an algorithm where incompressibility, Laplacian smoothing, and strain compatibility are enforced. The refined displacement field can generate a strain image of the prostate. Material properties such as Young’s modulus, Poisson’s ratio, and shear modulus can be iteratively reconstructed using finite element analysis to enhance the output images further. The material property calculation process, known for its accuracy, yields informative results to clinicians when diagnosing prostate malignancies.

Summary for Lay Audience

Detecting prostate cancer early is one of the best ways of ensuring an individual’s survival, and current detection methods, such as an antigen test or MRI, may be non-conclusive or too costly to undergo. To avoid these issues, medical imaging techniques with an ultrasound system can be used to determine the material properties of human prostate tissue by collecting a pre- and post-compression image of the prostate. The initial estimate algorithm is responsible for collecting and analyzing data, while the second algorithm refines the estimates by enforcing material physics onto them. Enforcing these known properties in this process allows researchers to capture accurate and quantifiable stiffness measurements of the human prostate in real-time. This allows for rapid analysis of the prostate tissue for multiple regions of the prostate. Simulation and tissue-mimicking materials can be used as data sources to analyze the algorithms used to compute the stiffness and measure a ground truth. The simulated data can be obtained and modified using MATLAB software packages to fit the needs of the research goal and provide data sets that are both accurate and moderately lifelike.

Our research process has been meticulous and thorough, ensuring the reliability and validity of our findings. Tissue-mimicking material, created using gelatin and foam, provides a cheap alternative to human tissue. This material can be scanned with a natural ultrasound system to simulate real clinical cases, where the collected data may not be of good quality or contain image regions where artifacts exist. In addition, the calculated stiffness can be used to further enhance the imaging capabilities of ultrasound data by additional algorithm processes, allowing for additional information to be learned from the output images. Overall, this thesis outlines the steps taken to develop, quantify, and evaluate the effectiveness of three different algorithms in processing ultrasound data to quantify the material properties better. This will provide clinicians with a tool that is specific and easily accessible when screening for prostate cancer, offering a promising avenue for future research and application.

Creative Commons License

Creative Commons Attribution-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-No Derivative Works 4.0 License.

Download

Share

COinS