3651 - Clinical and Dosimetric Factors Associated with Reactive Gastrostomy Tube Placement in Patients Treated with Radiotherapy for Head and Neck Cancer

N. Gallardo¹, C. M. Lanier², N. Razavian¹, R. F. Shenker³, J. H. Lee¹, A. Iyer¹, and R. T. Hughes¹; ¹Department of Radiation Oncology, Wake Forest University School of Medicine, Winston Salem, NC, ²Department of Radiation Oncology, Wake Forest University School of Medicine, Winston-Salem, NC, ³Department of Radiation Oncology, Duke University Medical Center, Durham, NC

Purpose/Objective(s): Patients who receive radiotherapy (RT) for head and neck cancer (HNC) often require a gastrostomy tube (GT) to maintain adequate nutrition. While GTs were historically placed prophylactically (ppx), this practice has been replaced with reactive (rx) GT placement only if needed during or after RT. It remains unclear how best to predict rx GT placement in patients managed with this approach. We assessed the relationship between clinical, functional swallowing, and dosimetric factors to predict rx GT placement. Materials/

Methods: Patients with HNC treated with RT with or without concurrent chemotherapy from 2018-2021 at our institution were analyzed. Patients with baseline GT dependence or lacking follow-up were excluded. Swallowing function at baseline was characterized using the Functional Oral Intake Scale (FOIS), which ranged from 1 (no oral intake) to 7 (total oral intake without restrictions). Outcomes included rx GT placement and weight loss during and after treatment. Mean RT dose (Gy) to the pharyngeal constrictor, RTOG/NRG Pharynx, and glottic/supraglottic larynx were abstracted from the treatment planning system. Multivariable analysis identified predictors of rx GT and adjusted logistic regression models were used to develop a predictive model of rx GT placement.
Results: 212 patients were included for analysis: 105 in the ppx group and 107 in the rx group, 10% of which received rx GT. Patients in the ppx GT group were more likely to have worse performance status, more extensive stage, lower baseline FOIS, bilateral neck RT and concurrent chemotherapy. There were no significant differences in on-treatment weight loss between groups at end RT, 1,3,6, or 12-months post-RT. In the rx GT group, baseline FOIS <6, nodal stage, and concurrent chemotherapy were significantly associated with tube placement. Multivariable models identified Pharynx mean dose, baseline FOIS, and baseline weight as significant predictors of rx GT placement. A logistic regression model of rx GT adjusted for baseline FOIS status (6-7 v. <6) was found to predict risk of GT placement. Among patients with a baseline FOIS <6, the predicted probability of rx GT placement increased with mean pharynx dose: 45 Gy, 38.5%; 60 Gy, 69.9%. In patients with a baseline FOIS 6-7, the probability of GT was: 45 Gy, 8%; 60 Gy, 23.4%.
Conclusion: Placement of rx GTs is associated with pre-treatment swallowing function and dose to swallowing structures. Adjusted for baseline swallowing function, risk of GT placement increases as a function of mean pharynx dose. These data can be used to identify patients at highest risk for rx GT placement and inform ppx GT practice and on-treatment monitoring.