Skip to content
DraconiticResearch
Search Sources API
209,932 studies31 sourcesUpdated 1 min ago
Peer-reviewedCC BY 4.0Frontiers in Oncology

Machine learning to predict osteoradionecrosis in head and neck cancer

  • Vitoria Tavares Castro1
  • Luan Matheus Trindade Dalmazo2
  • Paula Elaine Diniz Dos Reis3
  • Wenzel Castro de Abreu4
  • Ricardo Gomes dos Reis4
  • Ana Carolina Prado-Ribeiro5,6
  • André Ferreira Leite1
  • Fabiana T. Almeida7
  • Eliete Neves Silva Guerra1
7 affiliations
  1. 1University of Brasilia, Laboratory of Oral Histopathology, School of Health Sciences
  2. 2Department of Informatics, Federal University of Parana
  3. 3University of Brasilia, Interdisciplinary Laboratory of Research Applied to Clinical Practice in Oncology, Nursing Department, School of Health Sciences
  4. 4High-Complexity Oncology Care Center, University Hospital of Brasilia
  5. 5Dental Oncology Service, Instituto Do Câncer Do Estado de São Paulo, Faculdade de Medicina da Universidade de São Paulo
  6. 6Bioengineering Graduate Program, Scientific and Technological Institute, Brazil University
  7. 7Mike Petryk School of Dentistry, Faculty of Medicine and Dentistry, University of Alberta
Read full textPDFDOICite

Abstract

Objective: Osteoradionecrosis is one of the most severe late adverse effects of head and neck radiotherapy. Predictive models may help identify patients at increased risk and support personalized treatment. This exploratory study aimed to develop a machine learning–based prediction model for osteoradionecrosis in a real-world cohort, following the TRIPOD+AI statement.

Methods: Clinical variables and mandibular dose–volume histogram parameters were collected from head and neck cancer patients treated with radiotherapy. Forward Stepwise Logistic Regression was used for predictor selection. Six machine learning algorithms, Logistic Regression, Random Forest, Support Vector Machine, CatBoost, XGBoost, and TabPFN, were evaluated under class-weighted baseline modelling, threshold optimization, and data augmentation with repeated stratified 5-fold cross-validation. Performance was primarily assessed using sensitivity. Pairwise comparisons used the Nadeau–Bengio corrected paired t-test with Holm adjustment.

Results: Ninety-seven patients were included, of whom 8 developed osteoradionecrosis (8.25%). Of the 80 available variables, V60Gy and tumor location in the oral cavity were selected as predictors. Without augmentation, Logistic Regression achieved the highest sensitivity (0.74 [95% CI: 0.40–1.00]), while several algorithms showed poor sensitivity. With data augmentation, sensitivity ranged from 0.72 to 0.77 across models, with TabPFN, CatBoost, Logistic Regression and Support Vector Machine reaching the highest sensitivity. Logistic Regression achieved the highest AUC under both strategies (0.87). However, F1 scores remained modest, reaching a maximum of 0.48.

Conclusion: Machine learning models demonstrated exploratory potential for ORN risk assessment within an internal validation framework. Class-imbalance strategies were associated with greater minority-class detection in this cohort, but their generalizability remains uncertain. Larger multicenter cohorts and external validation are needed before clinical implementation.

Related

  1. A Machine Learning Model to Predict Complications Following Head and Neck Microvascular Surgery
    Caroline A. Bonaventure, Adam Bess, Reed Smith +5
    Laryngoscope investigative otolaryngologyPeer-reviewedCC BY-NC-NDFull textGeneral medicineEurope PMC
  2. Clinicopathological Machine Learning Models for Mortality Risk Prediction in Head and Neck Squamous Cell Carcinoma: A Retrospective Analysis of the The Cancer Genome Atlas Head-Neck Squamous Cell Carcinoma (TCGA-HNSC) Cohort
    Sharjeel Chaudhry, Zarmina Ehtesham, Fatimah AL- Ayiedh +2
    CureusPeer-reviewedCC BYFull textGeneral medicineEurope PMC
  3. Genesis—progression of early jaw osteoradionecrosis in head and neck cancer survivors: dynamics of independent but interconnected factors
    Sylvie E. Delanian
    Frontiers in oncologyPeer-reviewedCC BYFull textOncologyEurope PMCCrossref
  4. Osteoradionecrosis in patients with head and neck cancer after reirradiation - a systematic review and meta-analysis.
    Annemarie Fritz, Nils Gleim, Anita Kloss-Brandstätter +5
    Clinical and translational radiation oncologyPeer-reviewedCC BYFull textGeneral medicineEurope PMC
  5. Machine learning models to predict pathological complete response after chemoradiotherapy in muscle-invasive bladder cancer: toward optimized indications for bladder preservation.
    Motohiro Fujiwara, Hajime Tanaka, Daisuke Hirahara +13
    Clinical and translational radiation oncologyPeer-reviewedCC BYFull textSurgery & anaesthesiaEurope PMCPMC

Cite

BibTeXRIS
BibTeX
@article{castro2026machine,
  title = {{Machine learning to predict osteoradionecrosis in head and neck cancer}},
  author = {Vitoria Tavares Castro and Luan Matheus Trindade Dalmazo and Paula Elaine Diniz Dos Reis and Wenzel Castro de Abreu and Ricardo Gomes dos Reis and Ana Carolina Prado-Ribeiro and André Ferreira Leite and Fabiana T. Almeida and Eliete Neves Silva Guerra},
  journal = {Frontiers in Oncology},
  year = {2026},
  month = oct,
  doi = {10.3389/fonc.2026.1955800},
  url = {https://www.frontiersin.org/articles/10.3389/fonc.2026.1955800/full},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Machine learning to predict osteoradionecrosis in head and neck cancer
AU  - Vitoria Tavares Castro
AU  - Luan Matheus Trindade Dalmazo
AU  - Paula Elaine Diniz Dos Reis
AU  - Wenzel Castro de Abreu
AU  - Ricardo Gomes dos Reis
AU  - Ana Carolina Prado-Ribeiro
AU  - André Ferreira Leite
AU  - Fabiana T. Almeida
AU  - Eliete Neves Silva Guerra
PY  - 2026
DA  - 2026/10/21
JO  - Frontiers in Oncology
DO  - 10.3389/fonc.2026.1955800
UR  - https://www.frontiersin.org/articles/10.3389/fonc.2026.1955800/full
AB  - Objective: Osteoradionecrosis is one of the most severe late adverse effects of head and neck radiotherapy. Predictive models may help identify patients at increased risk and support personalized treatment. This exploratory study aimed to develop a machine learning–based prediction model for osteoradionecrosis in a real-world cohort, following the TRIPOD+AI statement.  Methods: Clinical variables and mandibular dose–volume histogram parameters were collected from head and neck cancer patients treated with radiotherapy. Forward Stepwise Logistic Regression was used for predictor selection. Six machine learning algorithms, Logistic Regression, Random Forest, Support Vector Machine, CatBoost, XGBoost, and TabPFN, were evaluated under class-weighted baseline modelling, threshold optimization, and data augmentation with repeated stratified 5-fold cross-validation. Performance was primarily assessed using sensitivity. Pairwise comparisons used the Nadeau–Bengio corrected paired t-test with Holm adjustment.  Results: Ninety-seven patients were included, of whom 8 developed osteoradionecrosis (8.25%). Of the 80 available variables, V60Gy and tumor location in the oral cavity were selected as predictors. Without augmentation, Logistic Regression achieved the highest sensitivity (0.74 [95% CI: 0.40–1.00]), while several algorithms showed poor sensitivity. With data augmentation, sensitivity ranged from 0.72 to 0.77 across models, with TabPFN, CatBoost, Logistic Regression and Support Vector Machine reaching the highest sensitivity. Logistic Regression achieved the highest AUC under both strategies (0.87). However, F1 scores remained modest, reaching a maximum of 0.48.  Conclusion: Machine learning models demonstrated exploratory potential for ORN risk assessment within an internal validation framework. Class-imbalance strategies were associated with greater minority-class detection in this cohort, but their generalizability remains uncertain. Larger multicenter cohorts and external validation are needed before clinical implementation.
SN  - 2234-943X
ER  -