Resource type
Thesis type
(Thesis) Ph.D.
Date created
2023-09-08
Authors/Contributors
Author: Momtahen, Shadi
Abstract
This Ph.D. thesis investigates the potential of the Diffuse Optical Breast Scanning (DOB-scan) probe, utilizing advanced imaging algorithms, machine learning techniques, and ensemble learning approaches to enhance breast cancer diagnosis, treatment response monitoring, and patient outcomes. The research demonstrates the effectiveness of the DOB-scan probe in breast cancer imaging by developing and evaluating the probe along with the modified diffusion equation (MDE) algorithm for tumor localization. The MDE algorithm shows promise in detecting breast cancer abnormalities, reconstructing functional optical images, identifying cancerous lesions, and evaluating chemotherapy response. These advancements provide valuable insights into tumor characteristics and treatment outcomes, improving diagnostic accuracy and personalized treatment strategies. Furthermore, machine learning techniques are integrated to detect and classify breast cancer based on the extracted features from the DOB-scan images. The utilization of machine learning enables accurate prediction of residual cancer burden (RCB) based on changes in optical properties, facilitating a comprehensive assessment of the disease status and determining appropriate treatment strategies. The integration of an ensemble learning approach further enhances the accuracy of breast cancer detection, combining the strengths of multiple algorithms for improved sensitivity and specificity. This enables earlier interventions and better patient outcomes by enabling timely and targeted treatments. Moreover, integrating the DOB-scan probe into the SIMA (Scanning-based Imaging and Measurement Analysis) approach revolutionizes real-time breast cancer imaging. SIMA provides high-resolution cross-sectional images for accurate localization and characterization of breast cancer-related abnormalities. By leveraging the DOB-scan probe's advanced imaging capabilities and the analytical power of SIMA, valuable insights into tumor morphology and treatment response are gained. This thesis contributes to the advancements in breast cancer imaging, highlighting the promising role of the DOB-scan probe, advanced imaging algorithms, machine learning techniques, and ensemble learning approaches in facilitating early detection, improving patient outcomes, and advancing the field of breast cancer research.
Document
Extent
150 pages.
Identifier
etd22745
Copyright statement
Copyright is held by the author(s).
Supervisor or Senior Supervisor
Thesis advisor: Golnaraghi, Farid
Language
English
Member of collection
Download file | Size |
---|---|
etd22745.pdf | 12.2 MB |