2751 - Spurn the Scorecard: Personalized Dose Constraints for Head and Neck Radiotherapy to Minimize Treatment Toxicity

Purpose/Objective(s): Toxicity after radiotherapy (RT) for head & neck cancer (HNC) is associated with dose to normal tissues (i.e. organs-at-risk [OARs]). Modern RT planning utilizes OAR dose constraints recommended by national guidelines/protocols. We hypothesized that overreliance on standardized constraints/scorecards may result in suboptimal tissue sparing. The purpose of this study was to investigate if a personalized dose constraint strategy that pushed RT planning to achieve OAR doses that are as low as reasonably achievable (ALARA) could meaningfully reduce OAR doses & decrease toxicity.

Materials/Methods: Patients with primary HNCs treated definitively (Nov 2019 - Dec 2023) were included in this single-institution retrospective study. Personalized ALARA OAR dose constraints were determined for each patient by the treating MD based on qualitative assessment of the target volume & adjacent normal tissues (which were compared to standardized constraints recommended by cooperative group trials, i.e. NRG-HN009). Dosimetric association with toxicity (including G-tube insertion, weight change, and osteoradionecrosis [ORN]) was analyzed using logistic regression.

Results: 168 patients received RT with personalized OAR constraints (75% concurrent chemotherapy, 27% post-operative). Primary sites included oropharynx (38%), larynx (24%), & oral cavity (23%). Delivered doses to parotid and submandibular glands, pharyngeal constrictors, esophagus, mandible, oral cavity, & spinal cord were all significantly lower than NRG-HN009 constraints (Table 1; all P < 0.01). Among patients taking oral nutrition at the start of RT, 16% required temporary G-tube placement (median duration 4.7 months, long-term rate 4.3%). Median weight loss was 5.4 kg and ORN occurred in 5.4% (median follow-up 29.8 months). G-tube placement was associated with every successive Gy increase in mean dose to pharyngeal constrictors (OR 1.07 / Gy, P = 0.009), contralateral parotid (OR = 1.08 / Gy, P = 0.006), & contralateral submandibular gland (OR 1.04 / Gy, P = 0.02) on logistic regression, including age, sex, primary site, concurrent chemotherapy, & prior surgery as covariates. Mean oral cavity dose (OR 1.06 / Gy, P = 0.048) but not max mandible dose was correlated with ORN.

Conclusion: Acknowledging the limitations of this study design, personalized OAR dose constraints for HNC RT are feasible & may decrease toxicity burden. Future work will integrate patient-reported outcomes & develop automated methods to derive personalized OAR constraints

Abstract 2751 - Table 1