<h3>Purpose/Objective(s)</h3> Late dysphagia after head and neck radiotherapy is an important driver of poor quality-of-life. We aimed to identify risk factors associated with long-term gastrostomy tube (g-tube) dependence in patients with head and neck squamous cell carcinoma treated with definitive or postoperative radiotherapy. <h3>Materials/Methods</h3> A retrospective cohort analysis of patients with head and neck squamous cell carcinoma treated with definitive or adjuvant intensity-modulated radiotherapy was performed. Patients with locoregional or distance recurrence were excluded from the analysis. G-tube dependence was used as a surrogate marker for late dysphagia, and its placement and removal from initial presentation to last follow-up was tracked. The risk of g-tube dependence from years one through five was calculated, and factors associated with 1- and 2-year g-tube use were analyzed with multivariable logistic regression. <h3>Results</h3> Our analytic cohort consisted of 421 survivors with head and neck squamous cell cancer treated from 2001 to 2018. A total of 210 patients required a g-tube with 80, 124, and 13 g-tubes placed before, during, and after radiotherapy, respectively. The median follow-up of the entire cohort was 4 years (IQR 3-5), with 314, 222, and 124 patients at-risk in years 3, 4 and 5. The Kaplan-Meier g-tube dependence rate at 1 and 2 years after radiotherapy were 11% and 6%, respectively, and 5%, 3% and 2% at three, four and five years, respectively. On multivariable regression analysis on 1-year g-tube dependence, history of depression or anxiety (odds ratio, OR, 95% confidence interval, CI, 3.86 [1.47, 10.3]), clinical T3/T4 stage (3.75 [1.60, 9.30]), clinical N2 or N3 stage (3.66 [1.03, 14.4]), baseline swallowing impairment to both solids and liquids (11.1 [3.22, 41.8]), and intra-treatment weight loss greater than 10% (3.75 [1.51, 9.81]) were significant predictors. A total of 390 patients had adequate follow-up data for the 2-year analysis. On multivariable analysis, age (1.06 [1.00, 1.14]), clinical N2/N3 stage (15.2 [1.72, 393]), and baseline swallowing impairment to both solids and liquids (95% confidence interval: 41.5 [5.59, 551]) were significantly associated with increased odds of 2-year g-tube dependence. Definitive vs. adjuvant treatment was not significantly associated with 1 or 2-year g-tube dependence. Out of 22 patients with g-tube dependence at two years after radiotherapy, only six (27%) patients had a known tube removal in the years to follow. <h3>Conclusion</h3> The overall incidence of prolonged g-tube dependence is low but not trivial, and the vast majority of patients with dependence at two years remained dependent permanently. We have identified a cohort at highest risk for long-term use, for whom early interventions should be targeted since late removal becomes progressively more unlikely.