.png){kind=icon}
.png){kind=icon}
Return
Personalized Prevention of Osteoradionecrosis of the Jaws: From Clinical Risk Profiling and Site-Specific Dosimetry to Artificial Intelligence and Individualized Treatment-Effect Estimation
.png){kind=icon}
OA
AI
L
H
F
G
G
G
G
DOI：10.3390/jpm16080413.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Background/Objectives: Osteoradionecrosis of the jaws remains a severe late complication of head and neck radiotherapy. Current preventive strategies are still frequently based on population-level dose thresholds and broadly standardized dental-clearance protocols, despite substantial heterogeneity in individual risk. This comprehensive review aimed to integrate established preventive guidance with emerging tooth-level, spatial-dosimetric, prognostic, and causal approaches within an integrative conceptual framework for personalized osteoradionecrosis prevention. Methods: Structured searches of five databases were completed on 30 May 2026. Evidence sources were selected and mapped across six a priori thematic domains, with the selection process summarized in a simplified flow diagram and an evidence map. Priority was given to guidelines, systematic reviews, large cohorts, externally validated prediction models, and clinically actionable studies. Results: Osteoradionecrosis risk is determined by interactions among tumor site, surgical anatomy, mandibular dose distribution, dental and periodontal disease, smoking, diabetes, nutritional status, biological susceptibility, and expected survival. Tooth extraction is not uniformly protective, and its potential benefit depends on tooth prognosis, local radiation dose, oncological urgency, and patient-level vulnerability. Dose–volume parameters and site-specific mapping provide more clinically relevant information than prescribed dose alone. Contemporary normal tissue complication probability and machine-learning models increasingly support individualized risk estimation, although calibration, external validation, data quality, and workflow integration remain limiting. Competing-risk and causal-inference approaches may further distinguish baseline risk from the expected benefit of specific preventive interventions. Conclusions: The evidence reviewed can be organized within an integrative conceptual framework combining tooth-level prognosis, spatial dosimetry, systemic susceptibility, competing mortality, and intervention burden. This framework synthesizes and reorganizes existing evidence; it is not a validated clinical model and requires prospective validation and clinical-impact evaluation before routine implementation. Until then, available prediction models should support, rather than replace, multidisciplinary clinical judgement.
Keywords:
osteoradionecrosis
head and neck radiotherapy
personalized medicine
dental clearance
mandibular dosimetry
normal tissue complication probability
machine learning
risk prediction
causal inference
tooth preservation
Journal
IF:
3
Papers:
7.4K
Citations:
1.3W
Organization
Cited Papers