Oral cancer is the sixth most common cancer worldwide and has been marked by high morbidity and poor survival rates that have changed little over the past few decades. Beyond prevention, early detection is the most crucial determinant for successful treatment and survival of cancer. Yet current methodologies for cancer diagnosis based upon pathological examination alone are insufficient for detecting early tumor progression and molecular transformation. Development of new diagnostic tools incorporating tumor biomarkers could enhance early detection by providing molecular-level insight into the biochemical and cellular changes associated with oral carcinogenesis. The work presented in this doctoral dissertation aims to address this clinical need through the development of new automated cellular analysis methods, incorporating lab-on-a-chip sensor techniques, for examination of molecular and morphological biomarkers associated with oral carcinogenesis. Using the epidermal growth factor receptor (EGFR) as a proof-of-principle biomarker, the sensor system demonstrated capacity to support rapid biomarker analysis in less than one-tenth the time of traditional methods and effectively characterized EGFR biomarker over-expression in oral tumor-derived cell lines. Successful extension from in vitro tumor cell lines to clinically relevant exfoliative brush cytology was demonstrated, providing a non-invasive method for sampling abnormal oral epithelium. Incorporation of exfoliative cytology further helped to define the important assay and imaging parameters necessary for dual molecular and morphological analysis in adherent epithelium. Next, this new sensor assay and method was applied in a small pilot study in order to secure an initial understanding of the diagnostic utility of such biosensor systems in clinical settings. Four cellular features were identified as useful indicators of cancerous or pre-cancerous conditions including, the nuclear area and diameter, nuclear-to-cytoplasm ratio, and EGFR biomarker expression. Further examination using linear regression and ROC curve analysis identified the morphological features as the best predictors of disease while a combination of all features may be ideal for classification of OSCC and pre-malignancy with high sensitivity and specificity. Further testing in a larger sample size is necessary to validate this regression model and the LOC sensor technique, but shows strong promise as a new diagnostic tool for early detection of oral cancer. / text