Predictive modelling of mandibular osteoradionecrosis in head and neck cancer patients: clinical and dosimetric insights.

INTRODUCTION

Mandibular osteoradionecrosis (ORN) is a serious complication of radiotherapy (RT) for head and neck cancer (HNC), with an incidence of 3-15%. ORN results from radiation-induced bone necrosis and may require surgical intervention. This study investigates clinical and dosimetric predictors of ORN risk and develops a predictive model for individualized risk assessment.

MATERIALS AND METHODS

This retrospective case-control study included 298 HNC patients treated with RT or chemoradiotherapy between January 2012 and May 2020. Dosimetric parameters, including mandibular V10-V60, mean dose (Dmean), and maximum dose (Dmax), were analyzed alongside clinical data such as age, tumor site, smoking history, and dental extractions.

RESULTS

Over a median follow-up of 32.4 months, 20 patients (6.7%) developed ORN, with a 5-year cumulative incidence of 7.4%. Multivariate analysis identified mandibular V50 (HR = 1.05, p = 0.0015) and post-RT dental extractions (HR = 2.51, p < 0.0001) as significant ORN risk factors, while age was protective (HR = 0.96, p = 0.047). A V50 cutoff of 25.4 cm³ was most predictive (p = 0.0016). The multivariate model incorporating V50, age, and dental extractions demonstrated strong accuracy (C-index: 0.815, AUC: 0.8).

CONCLUSION

V50 and post-RT dental extractions are key predictors of ORN. The developed nomogram enables personalized risk assessment, supporting treatment optimization. These findings emphasize the need for tailored RT planning and dental care to mitigate ORN risk, warranting validation in multi-institutional cohorts.