Multi-institutional Normal Tissue Complication Probability (NTCP) Prediction Model for Mandibular Osteoradionecrosis: Results from the PREDMORN Study
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Mandibular osteoradionecrosis (ORN) is a severe late complication affecting patients with head and neck cancer (HNC) treated with radiotherapy that significantly impacts patients’ quality of life and can require costly interventions. While radiation dose is a key factor, other clinical and demographic risk factors influence ORN development. Previous predictive models have primarily been single-institutional, limiting their generalizability. The PREDMORN Consortium was established to address these limitations. In this first analysis, we have aimed to reproduce existing statistical association and modelling analyses on the largest and most diverse mandibular ORN cohort worldwide to allow comparison with previous studies. As such, we have developed, tested and externally validated a multi-institutional normal tissue complication probability (NTCP) model for mandibular ORN.
Methods
This retrospective multi-institutional study included 1,184 HNC patients (389 ORN cases) from seven institutions. Clinical, demographic, and dosimetric (DVH) variables were analysed to develop a prediction model (any grade of ORN vs. no ORN) using forward stepwise logistic regression with correlation-based variable pre-selection. The ORN NTCP model was developed on 80% of data from six institutions, tested on the remaining unseen 20%, and externally validated on the seventh institution’s dataset.
Results
Key predictors of ORN were D30%, V70Gy, pre-RT dental extractions, and smoking status. The ORN NTCP model demonstrated good calibration and predictive performance, with AUCs of 0.69 for internal testing and external validation, which improved when tested on a sub-cohort of oropharyngeal and locally advanced larynx/hypopharynx cancer cases (AUCs of 0.75).
Conclusion
The PREDMORN NTCP model is the largest multi-institutional effort to predict ORN risk in HNC patients. We provide guidance on how to adjust the NTCP predicted probabilities for differences in target population baseline ORN risk to facilitate application of the model. Future research will focus on incorporating imaging-based spatial data and further external validation to enhance clinical applicability.
