Abstract

Breast cancer is one of the most common cancers affecting the lives of many women and taking over 683,000 lives worldwide. Early detection of breast cancer through screening has been instrumental in reducing the mortality rate. However, current screening methods based on mammography have issues with patient discomfort, invasiveness, cost and low accuracy, mainly due to dense breast tissue in approximately 40% of women population. Infrared imaging (IRI) is shown to be noninvasive, cost effective, comfortable, and not affected by breast density. The present work describes a novel patient-specific IRI inverse heat transfer algorithm based on the digital model of the breast generated from MRI images and the IRI-Numerical Engine (IRI-NE). Validation of the IRI-NE is conducted with clinical IR images of 23 biopsy-proven breast cancer patients (24 breasts with cancer) through patient-specific inverse heat transfer modeling. This is a continuation of a collaborative study between RIT and Rochester General Hospital. The IRI-NE was able to accurately detect the presence and absence of breast cancer in all patients regardless of breast density, cancer type, tumor size, and tumor depth. The tumor size prediction was compared with actual size obtained from MRI and patient reports and showed a predicted size within 2.4 mm of the actual size. This shows the potential of IRI as an effective adjunct to mammography. Additionally, this work evaluates the effect breast size, shape, tissue density, vascularity, tumor size and location on the detectability with the IRI-NE. Surface temperature data for various scenarios related to these factors are generated through numerical simulation on the actual digital model of the breast. The IRI-NE was able to predict the presence or absence of cancer for all the cases studied. The detectability limit of the IRI-NE is shown to be dependent on the thermal sensitivity of the IR camera, and a higher sensitivity camera is needed to detect deep tumors in large breasts. Finally, the thermal effects of vascularity and tumor angiogenesis on the detection accuracy are investigated. The blood perfusion rate was shown to significantly influence the heat transfer in small growing tumors due to angiogenesis. This shows the ability to further study cancer through a thermal perspective.

Publication Date

6-5-2024

Document Type

Dissertation

Student Type

Graduate

Degree Name

Engineering (Ph.D.)

Department, Program, or Center

Mechanical Engineering

College

Kate Gleason College of Engineering

Advisor

Satish G. Kandlikar

Advisor/Committee Member

Isaac Bernabe Perez-Raya

Advisor/Committee Member

Cristian Linte

Campus

RIT – Main Campus

Download

Share

COinS