"Early Breast Cancer Detection with Ultrasound Data using NMF" by Rutvi Khamar

Date of Award

5-2025

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Electrical Engineering and Computer Science

First Advisor

Debasis Mitra

Second Advisor

Venkat Keshav Chivukula

Third Advisor

Siddhartha Bhattacharyya

Fourth Advisor

Brian A. Lail

Abstract

Early detection of breast cancer significantly influences patient outcomes. Dynamic Contrast-Enhanced Ultrasound (DCE-US) has shown promise in early detection by visualizing tumor vascularity and perfusion dynamics in real-time. This study evaluates the efficacy of DCE-US in distinguishing four stages of cancer progression: normal, hyperplasia, ductal carcinoma in situ (DCIS), and invasive cancer, using a transgenic mouse model that mimics human breast cancer. Ultrasound burst pulses, while commonly used to remove unbound contrast agents, can potentially damage human tissues. Using the pre-pulse data helps mitigate this risk, ensuring safer and more reliable measurements. A VEGFR2-targeted microbubble contrast agent was injected, and the analysis was conducted using Non-negative Matrix Factorization (NMF), focusing on pre-pulse data. Our results revealed significant p-values that statistically distinguished all combinations of cancer stages except for normal versus hyperplasia and DCIS versus invasive cancer. Additionally, by considering normal and hyperplasia as benign, and DCIS and invasive as malignant, we applied a non-parametric regression model to define a threshold differentiating benign from malignant tissues, achieving robust specificity and sensitivity metrics. These findings underscore the potential of NMF-enhanced DCE-US for accurate tumor characterization and early breast cancer detection, all the while avoiding the use of harmful ultrasound burst pulses to remove unbound tracers. This approach holds significant promise for clinical applications

Available for download on Sunday, May 10, 2026

Share

COinS