| Biophotonics represents a major new area of hope and growth for high-tech medical and surgical intervention, and advanced optical imaging has much to offer and potentially a major role to play in this regard. In order to take advantage of recent advances in genomics, proteomics and cell biology in clinical practice, our level of understanding at tissue and pre-clinical levels needs to be enhanced using novel imaging-based technologies for quantification and visualization. In this dissertation, Multi-Mode Optical Spectral Signature (MMOSS) imaging for biomedical applications is introduced, and used to correlate histopathology with intra- and extraoperative imaging of normal and abnormal tissue specimens. Moreover, practical solutions to the challenges of applying MMOSS imaging technology to the real pre-clinical and clinical environment are proposed and implemented.Among a number of optical imaging modes available to us, hyperspectral imaging and fluorescence lifetime imaging were more thoroughly tested in our animal models. All of the concepts and designs for multi-mode optical spectral signature were implemented and demonstrated using Java programming language, in our custom-developed software program SpectralJ. It is implemented as a plugin for ImageJ, an NIH supported open-source image processing suite, and it is equipped with many functional features such as spectral-signature management (signature selection, registration, load and search), spectral signature similarity measure algorithms, enhanced color representation methods, dimensionality reduction algorithms and intelligent signature imaging capabilities for efficient real-time in vivo imaging.The usefulness of our new image analysis and visualization methods is demonstrated for selected preclinical research cases, including rat neck dissection, early detection of breast cancer, and the surgical management of Hirschprung's disease. |
