Osteoradionecrosis After Intensity-Modulated Radiation Therapy or Proton Therapy in Oropharyngeal Carcinoma.

Abstract

Osteoradionecrosis (ORN) is a potentially debilitating late complication of radiotherapy (RT) for head and neck cancer. While proton therapy offers superior dose conformality, its impact on ORN risk remains uncertain, particularly in patients with oropharyngeal squamous cell carcinoma (OPSCC). To characterize the incidence, severity, and predictors of ORN in a large institutional cohort of patients with OPSCC treated with curative-intent RT, and to compare outcomes between proton therapy and intensity-modulated radiation therapy (IMRT). This retrospective cohort study included consecutive patients with OPSCC treated from January 2013 to December 2023 at a single high-volume academic institution. Patients received either IMRT or proton therapy (uniform scanning or pencil beam scanning). ORN diagnosis and grading were determined through standardized multidisciplinary review. The primary outcome was the 3-year rate of ORN. Cox regressions identified predictors of ORN. Data were analyzed from December 2024 to April 2025. Radiotherapy modality (proton vs IMRT), patient demographic characteristics, smoking status, human papillomavirus status, tumor and/or node stage, chemotherapy, and radiation dosage. The analysis included 1564 patients (mean [SD] age, 61.5 [9.6] years; 208 females [13.3%] and 1356 males [86.7%]) of whom 1389 patients (88.8%) had undergone IMRT, and 175 patients (11.2%) had undergone proton-based treatment. The 3-year incidence of any-grade ORN was 3.02% (95% CI, 2.22%-4.09%). ORN rates were significantly higher after proton therapy compared with IMRT (6.36% vs 2.69% at 3 years; hazard ratio [HR], 2.62; 95% CI, 1.39-4.93). Of 1344 definitive patients, 47 of 1210 patients in the IMRT group (3-year rate, 2.38%; 95% CI, 1.61-3.51%) developed ORN vs 11 of 134 patients in the proton group (3-year rate 7.47%; 95% CI, 3.40-16.02% [HR, 3.62; 95% CI, 1.85-7.09]). On multivariable analysis, proton therapy (HR, 2.92; 95% CI, 1.55-5.50), concurrent chemotherapy (HR, 3.29; 95% CI, 1.03-10.50), and smoking (HR, 2.33; 95% CI, 1.38-3.92) were independently associated with ORN. Grade 3 or greater ORN occurred in 0.67% of patients and did not appear to differ by RT modality. In this large, relatively homogeneous cohort of patients with OPSCC, proton therapy was associated with a higher rate of ORN compared with IMRT, particularly in the definitive setting, although high-grade ORN remained uncommon across both modalities. These retrospective findings should be considered exploratory and underscore the need for future hypothesis-driven studies to refine dose constraints and optimize treatment planning to mitigate ORN risk among patients with OPSCC.