Multi-institutional Normal Tissue Complication Probability (NTCP) Prediction Model for Mandibular Osteoradionecrosis:...

Area of Science:

• Oncology
• Radiation Oncology
• Oral and Maxillofacial Surgery

Background:

• Osteoradionecrosis (ORN) is a severe complication following radiation therapy (RT) for head and neck cancer (HNC).
• ORN significantly impacts patient quality of life and necessitates costly treatments.
• Existing predictive models for ORN have limitations in generalizability due to their single-institutional nature.

Purpose of the Study:

• To develop and validate the largest multi-institutional normal tissue complication probability (NTCP) model for predicting mandibular ORN risk.
• To compare findings with previous single-institution studies using the most diverse mandibular ORN cohort globally.
• To identify key clinical, demographic, and dosimetric predictors of ORN.

Main Methods:

• A retrospective analysis of 3928 HNC patients (622 ORN cases) from 8 institutions was conducted.
• A prediction model for any-grade ORN was developed using forward stepwise logistic regression.
• The model was internally validated and externally tested on matched and population-based cohorts.

Main Results:

• Key predictors identified for ORN include D30%, V70Gy, pre-RT dental extractions, and smoking status.
• The developed ORN NTCP model demonstrated strong calibration and improved discrimination upon validation.
• External validation on a large population-based cohort confirmed the model's good generalizability.

Conclusions:

• The PREDMORN NTCP model represents a significant multi-institutional effort in predicting ORN risk in HNC patients.
• The model's generalizability was confirmed through external validation on a large population-based cohort.
• Findings align with current clinical guidelines and support previous single-institution study results.