Oral squamous cell carcinoma (SCC) is the 6th most common malignancy in the US and is often associated with a history of tobacco and alcohol use. Five year survival and prognosis remain poor at 60-80% with the outcome becoming bleaker upon metastasis to the cervical (neck) lymph nodes (25-40%). Oral SCCs are preceded by oral dysplasias (precancers), but not all oral dysplasias turn malignant. Thus, challenges facing oral cancer surgeons and oncologists include prediction of which dysplasias will undergo malignant transformation and which tumors will metastasize to the neck. Therefore there is an urgent need to identify accurate biomarkers to aid in the treatment and management of patients with oral cancers or pre-cancers. This thesis research addressed these problems by assessing whether genomic DNA copy number measurements could provide such markers. As described below, the work has culminated in the identification of two oral SCC subtypes that differ in clinical behavior, specifically the risk of metastasis. The genomic copy number biomarker is suitable for stratifying patients for risk of metastasis prior to surgery and so can guide surgical planning for one of the most challenging treatment decisions - how to treat patients with clinically node-negative (N0) necks. Current clinical practice recommends removal of the cervical lymph nodes at the time of surgery to resect the tumor, because salvage of patients who subsequently develop lymph node metastasis has a poor success rate. Identification of patients with low risk for metastasis would spare them additional major surgery with its risks and morbidity, as well as reduce medical costs. A molecular marker for metastasis risk will also inform follow up management of oral cancer patients.