Squamous intraepithelial lesions (SIL) are persistent white and white-red lesions that may represent precursor lesions for invasive squamous cell carcinoma (SCC). SILs encompass epithelial hyperplasia/hyperkeratosis, dysplasia, and carcinoma in situ. Specific indicators within SILs would be helpful for risk determination and treatment decisions. The purpose of this study was to determine the risk of SIL transformation to SCC using clinical appearance, histology, and immunohistochemical staining of biomarkers. A retrospective cohort study was conducted of patients diagnosed with oral cavity SILs between 2002 and 2012. Subjects were identified using billing records and pathology reports. Subjects were included if they had a new diagnosis of SIL via biopsy, histologic blocks/slides available for review, and > 5 years of follow-up after biopsy. Subjects with an existing or previous diagnosis of SCC at time of SIL diagnosis were excluded. Subjects were divided into two groups: 1) Those with SILs who did not transform to SCC and 2) those who transformed. Predictor variables included demographics (gender, age) and medical history (history of tobacco, alcohol, betel nut use, immune status, history of radiation). Macroscopic lesion related variables were site, size, and clinical appearance (described as white, red, heterogenous red-white, ulcerated, or nodular). Microscopic lesion variables were severity of dysplasia, location of atypia within epithelium (lower one-third, lower two-third, or full thickness), mitotic index, and amount of apoptosis, nuclear pleomorphism, rete peg complexity, lichenoid chronic inflammation, and hyperkeratosis. Mitotic index was quantified as the number of mitotically active cells recorded per high-powered field (hpf, 400x). Apoptotic abundance, nuclear pleomorphism, and rete-peg complexity were graded on a semi-quantitative scale 1-5. Lichenoid inflammation and hyperkeratosis were assessed as present or absent. Immunohistochemical lesion variables were p53 and Ki-67 immunoreactivity and measured as amount of positively stained cells per hpf. The primary outcome variable was presence of SCC on a subsequent biopsy (i.e., transformation to SCC). All diagnostic slides were reviewed and rated by a head and neck pathologist (WCF). Statistical results were analyzed using a Chi2 test, Fisher exact test, and 2-tailed t-test. Statistical results were considered significant with a 95% confidence interval if they had a P-value less than .05. A total of 60 subjects (45%, 27 female) were included of which 38 (63%, 17 female) did not transform to SCC whereas 22 (37%, 10 female) transformed. Mean ages of groups 1 and 2 were 58.2 ± 12.0 and 63.1 ± 15.1 years, respectively (P = .202). There were no differences in medical history and lesion location. Mean size of the SIL in groups 1 and 2 was 234 mm2 and 228 mm2, respectively, (P = .957). Transformation was significantly higher for heterogeneous red/white lesions: 5 (13%) in group 1, 9 patients (42%) in group 2 (P = .014). Patients with severe dysplasia were more likely to transform with 10 in group 2 (45.5%) and 2 in group 1 (5.3%) [P = .0003]. Atypia limited to the lower one-third was less likely to transform (P = .0173). Increased apoptosis (P = .043) and rete peg complexity (P = .014) were associated with SCC transformation. There was no difference in immunohistochemical staining of Ki-67 between groups (P = .305). In conclusion, SILs that transformed to SCC were more likely to have a red-white appearance, severe dysplasia, atypia throughout the epithelium, greater apoptosis, and rete peg complexity. Ki-67 immunoreactivity was not associated with transformation. Ongoing evaluation of p53 will be reported. The presence of these factors may influence decision to excise a SIL vs surveillance.