BackgroundThere is variability in the literature on the role of the depth of invasion (DOI) for recommending an elective neck dissection (END). PurposeThe purpose of the study is to estimate the DOI threshold for recommending an END. Study Design/Setting/SampleA retrospective cohort study was performed at McGill University Health Centre from 2008-2018 with 5 years of follow-up. The sample was subjects with clinical T1/T2 oral squamous cell carcinoma (OSCC) and clinically negative neck. Subjects with previous head and neck cancer were excluded. Predictor VariableThe primary predictor variable was DOI measured from the basement membrane of the adjacent normal mucosa on final pathology, coded as <4mm or ≥4mm. DOI is a continuous variable converted to a binary variable. Main Outcome VariableThe main outcome variable was time to development of neck disease (RD+) defined as the time from surgery to development of pathologic nodes. Time to RD+ for pathologic nodes discovered from the END was considered 0 months. The secondary outcome variable was overall survival. CovariatesDemographics (age, sex, smoking/alcohol history) and tumor characteristics (tumor location, clinical T, tumor differentiation, perineural invasion, lymphovascular invasion) were analyzed. AnalysesTime to RD+ and survival were analyzed using Cox hazard ratio, Kaplan-Meier curves, and log-rank test. Student T-test and Chi-square test were used for bivariate analyses; p≤0.05 was statistically significant. ResultsThe final sample were 64 subjects (average age 65.25 (SD 13.06) years and 36 (56.2%) males). Twenty-nine subjects had DOI<4mm, and the 5-year RD+ was 3.4% (the 1 occurrence of RD+ was at 5.3 months). Thirty-five subjects had DOI≥4mm, and the 5-year RD+ was 45.7% (15 subjects had RD+ discovered from the END, and 1 subject had RD+ at 7.6 months). DOI≥4mm had significantly higher risk of RD+ than DOI<4mm (HR 17.91; 95% CI 2.37-135.3; p=0.01), which remained significant after adjusting for clinical T, tumor differentiation, perineural invasion, and lymphovascular invasion (HR 9.53; 95% CI 1.12-81.44; p<0.05). The shallowest DOI with >20% risk of RD+ was in the DOI 4-4.9mm group. Conclusions/Relevance: Among patients with OSCC of T1 or T2 and clinically negative necks, END should be considered with DOI≥4mm.