Cancer is the second leading cause of human deaths in the United States. Detection of early-stage cancer has been widely recognized as one of the most critical factors to successfully treat cancer and reduce mortality. Current gold standards for cancer diagnosis includes computed tomography, magnetic resonance imaging, conventional X-ray imaging, and positron emission tomography followed by tissue biopsy or fine-needle aspiration to confirm. However, these techniques are ineffective in sensing early-stage cancer development since very early cancerous alterations happens at macromolecule scales within pre-cancer cells that cannot be detected by the aforementioned techniques. While optical microscopy is an ideal tool to access the information within cellular structures, the diffraction limit prevents resolving length scales under ~200nm.

Facilitated by three-dimensional, full-vector finite-difference time-domain (FDTD) computational solutions of Maxwell's equations, our group has developed a novel optical microscopy technique, Partial Wave Spectroscopic (PWS) microscopy, that can detect nanoarchitectural alterations within biological cells. These alterations are not visible using conventional optical microscopy techniques. With a scale sensitivity as fine as ~20 nm, PWS enables probing nanoscale density variations in macromolecular complexes, such as chromatin remodeling. PWS appears capable of detecting nanoscale changes occurring during the entire carcinogenesis process, thereby permitting early-stage cancer screening.

Despite the success of PWS, it made simple assumptions about the model of cells and instrument. For example, it assumes the cell has flat surfaces and the microscope has an illumination NA of 0. Additionally, the effect of biological staining on spectra is not clear.

In this dissertation, we expand the theory to include more realistic models of cells and instrument and validate rigorously via FDTD simulations. The results from these theoretical developments will provide essential information to optimize PWS for better diagnosis. In this dissertation, we first develop essential modules in our FDTD software package to allow simulations of rough surface, staining, and finite NA (Chapter 2). Facilitated by FDTD simulations, we develop new protocols and biomarkers to enhance the diagnostic power of PWS (Chapter 3). Finally, we perform clinical studies on prostate biopsies to test whether PWS can detect progression in prostate cancer (Chapter 4).