| Breast cancer is a leading cause of the cancer-related deaths in women, and early detection is important for the high-risk population. Infrared thermography has distinguished itself as a useful adjunctive tool to conventional X-ray mammography in early breast cancer screening due to its high sensitivity and specificity, passive (non-radiation) nature, and low cost. Nevertheless, one open question for breast thermography is how to quantify the complex relationships between the breast thermal behaviors and the underlying physiological and pathological conditions. Numerical simulations, advancing with the rapidly evolving computer technologies, increasingly have shown the promising potentials of investigating the roles of various properties of breast tissues in relation to thermography. Unfortunately, numerical studies to date have not been designed comprehensively, omitting several critical factors from the thermography modeling procedure.For example, previous thermography modeling generally did not account for either the gravity-induced elastic deformations arising from various body postures, or for the nonlinear elasticity associated with large deformations. Furthermore, no modeling study has been reported for a dynamic thermal procedure, such as cold stress and thermal recovery, which have been commonly implemented in clinical practice. Thus, the simulation-elaborated relationships between breast thermography and the internal and external factors might not be accurate enough. Consequently, the outcomes of those limited modeling studies cannot be relied on for robust inference of the physiological or pathological information from the measured breast thermogram.This dissertation develops more comprehensive breast thermography modeling techniques using the 3-D Finite Element Method (FEM) for both thermal and elastic properties of the breast. It explicitly considers gravity-induced deformation with linear/nonlinear elasticity. Moreover, along with conventional static thermography, dynamic thermal imaging procedures are examined to enhance the scientific and clinical application foundations of breast thermography. Based upon these improvements, the challenging "inverse problem" of breast thermography, i.e., how to estimate the individual thermal properties of non-tumorous breast tissues from the measured surface temperature distribution, can then be explored effectively.Our major findings include: (1) The gravity-induced elastic deformation does have significant impact on the breast temperature distribution according to the body posture and the elastic properties of breast tissues (2) The tumor-induced skin temperature alterations are recognizable for only superficial tumors (at depths less than 20mm) and the tumor size plays a less important role than the tumor depth (3) The tumor-induced surface thermal contrast has a different dynamic pattern from that of deformation-induced thermal contrast in cold-stress and/or thermal recovery processes and (4) Estimation of the tumor-induced thermal contrast can be significantly improved by using the proposed inverse problem-solving techniques to provide the individual-specific thermal background.These new techniques thus will provide a stronger foundation for, and greater specificity and precision in, thermographic diagnosis, and treatment, of breast cancer.Key Words: Breast cancer, Dynamic thermal imaging, Elastic-thermal modeling, Finite element method, Forward problem, Inverse problem, Non-linear elasticity, Thermogram, Thermography. |
