Improving early detection of cancer represents the most promising opportunity to reduce the mortality and morbidity of cancer. This thesis presents model-based analysis of tissue reflectance and autofluorescence which is used to improve our understanding of the biological basis of tissue spectra in the uterine cervix and oral cavity. An analytical model to extract biologically and diagnostically relevant parameters from clinical measurements of cervical tissue is developed, validated, and implemented. In addition, Monte Carlo based models are used to investigate the effect of dysplastic progression and benign inflammation on oral cavity spectra. Results show depth-sensitive spectroscopy may be able to improve cancer detection and be used as an adjunct to wide field imaging where inflammation causes decreased specificity.