Abstract Oral cancers remain both a challenge and an opportunity for clinicians and scientists working with this disease. Worldwide, it represents a significant global challenge, with close to 300,000 new cases diagnosed each year. The disease is potentially preventable: we have knowledge of its risk factors; it is also at an easily accessible site and is often preceded by oral premalignant lesions (OPLs). The challenge has been to develop a framework that would let us both detect and better manage such lesions. Unfortunately, only a small proportion of OPLs will progress to a cancer and histology alone has not allowed clinicians to differentiate high- from low-risk lesions. This uncertainty has also meant that, even when OPLs are detected, there is no consensus on who to treat and how. In 1999, we established the Oral Cancer Prevention Program in British Columbia to develop an integrated knowledge translation approach to facilitate systematic change for prevention, detection and management of this disease across the continuum of care throughout the province. This effort has involved optimization of screening in community dental practices, creation of a triage pathway from a centralized oral biopsy service to oral dysplasia clinics where patients are assessed, and formation of linkages to the cancer agency and local hospitals to facilitate both treatment and follow-up. A new Canadian Partnership Against Cancer-funded Oral Dysplasia Surveillance System has been developed in BC to monitor natural history of the disease. This province-wide structure has facilitated the formation of the Oral Cancer Prediction Longitudinal (OCPL) study, a unique resource for development and validation of biomarkers of risk of malignant progression. To date, ∼ 450 low-grade (mild and moderate) dysplasia cases have been accrued to this study and are in follow-up. In September of this year, we published results from the first 296 patients accrued to the OCPL study (Zhang et al., Cancer Prevention Research). We focused on the validation of a LOH risk model proposed by our group in 2000 that utilized a set of microsatellite markers at key chromosomal loci to predict progression of low-grade dysplasia in a retrospective cohort (Rosin et al., Clin Cancer Res 2000; 6:357–62). In the new study, cases were classified into high- or low-risk profiles to validate the 2000 model. Risk models were further refined using recursive partitioning and Cox regression analyses. The study showed that the high-risk lesions (3p &/or 9p LOH) had a 22•6 -fold increase in risk (P = 0•002) compared to low-risk lesions (3p & 9p retention). Addition of another two markers (loci on 4q/17p) further improved the risk prediction, with five-year progression rates of 3•1%, 16•3%, and 63•1% for the low-, intermediate-, and high-risk lesions, respectively. Compared to the low-risk group, intermediate- and high-risk groups had 11•6-fold and 52•1-fold increase in risk (P < 0•001). The new LOH profiles in the refined model were validated as risk predictors by using the initial retrospective cohort from 2000. Multi-covariate analysis with clinical features showed LOH models to be the most significant predictors of progression. The importance of this study is that it has direct implication for standard of care in the community setting, providing a tool by which patients could be triaged to different levels of intervention. Patients with elevated risk would be guided towards increased surveillance with this risk providing a rationale for the targeting of intervention regimes, even if these were associated with some morbidity. Two of three patients with such high-risk profiles progressed to cancer in 5 years. In contrast, patients with low-risk profiles could be spared in from aggressive monitoring and intervention. In our study, low-risk patients represented close to half of those individuals that were accrued to the study. Future research should be aimed towards exploring such possibilities, as an initial step towards personalizing oral cancer prevention strategies. LOH is currently being used in two chemoprevention trials of patients with OPLs to guide patient accrual: the phase III Erlotinib Prevention of Oral Cancer study and the Phase II Cetuximab for Treatment of High-Risk Pre- Malignant Upper Aerodigestive Lesions trial. The results of these studies should provide even greater knowledge upon which to build new strategies for patient triage that could better focus risk reduction programs for this disease. (Supported by grants from the NIH and the National Institute of Dental and Craniofacial Research (R01DE13124 and R01DE17013). Citation Format: Miriam P. Rosin, Lewei Zhang, Catherine Poh, Michele Williams, Denise M. Laronde, Ken Berean, Pamela J. Gardner, Huijun Jiang, Lang Wu, J. Jack Lee. Malignant risk prediction for patients with oral premalignant lesions. [abstract]. In: Proceedings of the Eleventh Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2012 Oct 16-19; Anaheim, CA. Philadelphia (PA): AACR; Cancer Prev Res 2012;5(11 Suppl):Abstract nr CN02-02.