Machine learning for the prediction of toxicities from head and neck cancer treatment: A systematic review with meta-analysis

Anna Luíza Damaceno Araújo; Matheus Cardoso Moraes; Maria Eduarda Pérez-de-Oliveira; Viviane Mariano da Silva; Cristina Saldivia-Siracusa; Caique Mariano Pedroso; Marcio Ajudarte Lopes; Pablo Agustin Vargas; Sara Kochanny; Alexander Pearson; Syed Ali Khurram; Luiz Paulo Kowalski; Cesar Augusto Migliorati; Alan Roger Santos-Silva

doi:10.1016/j.oraloncology.2023.106386

Machine learning for the prediction of toxicities from head and neck cancer treatment: A systematic review with meta-analysis

Oral Oncol. 2023 May:140:106386. doi: 10.1016/j.oraloncology.2023.106386. Epub 2023 Apr 4.

Authors

Anna Luíza Damaceno Araújo¹, Matheus Cardoso Moraes², Maria Eduarda Pérez-de-Oliveira³, Viviane Mariano da Silva², Cristina Saldivia-Siracusa³, Caique Mariano Pedroso³, Marcio Ajudarte Lopes³, Pablo Agustin Vargas³, Sara Kochanny⁴, Alexander Pearson⁴, Syed Ali Khurram⁵, Luiz Paulo Kowalski⁶, Cesar Augusto Migliorati⁷, Alan Roger Santos-Silva⁸

Affiliations

¹ Oral Diagnosis Department, Piracicaba Dental School, University of Campinas (UNICAMP), Piracicaba, São Paulo, Brazil; Head and Neck Surgery Department, University of São Paulo Medical School (UFMUSP), São Paulo, São Paulo, Brazil.
² Institute of Science and Technology, Federal University of São Paulo (ICT-Unifesp), São José dos Campos, São Paulo, Brazil.
³ Oral Diagnosis Department, Piracicaba Dental School, University of Campinas (UNICAMP), Piracicaba, São Paulo, Brazil.
⁴ Section of Hemathology/Oncology, Department of Medicine, University of Chicago, Chicago, IL, United States; University of Chicago Comprehensive Cancer Center, Chicago, Chicago, IL, United States.
⁵ Unit of Oral and Maxillofacial Pathology, School of Clinical Dentistry, University of Sheffield, S10 2TA Sheffield, United Kingdom.
⁶ Department of Head and Neck Surgery and Otorhinolaryngology, A.C. Camargo Cancer Center, São Paulo, Brazil; Head and Neck Surgery Department and LIM 28, University of São Paulo Medical School, São Paulo, Brazil.
⁷ College of Dentistry, University of Florida, Gainesville, FL, United States.
⁸ Oral Diagnosis Department, Piracicaba Dental School, University of Campinas (UNICAMP), Piracicaba, São Paulo, Brazil. Electronic address: alan@unicamp.br.

PMID: 37023561
DOI: 10.1016/j.oraloncology.2023.106386

Abstract

Introduction: The aim of the present systematic review (SR) is to summarize Machine Learning (ML) models currently used to predict head and neck cancer (HNC) treatment-related toxicities, and to understand the impact of image biomarkers (IBMs) in prediction models (PMs). The present SR was conducted following the guidelines of the PRISMA 2022 and registered in PROSPERO database (CRD42020219304).

Methods: The acronym PICOS was used to develop the focused review question (Can PMs accurately predict HNC treatment toxicities?) and the eligibility criteria. The inclusion criteria enrolled Prediction Model Studies (PMSs) with patient cohorts that were treated for HNC and developed toxicities. Electronic database search encompassed PubMed, EMBASE, Scopus, Cochrane Library, Web of Science, LILACS, and Gray Literature (Google Scholar and ProQuest). Risk of Bias (RoB) was assessed through PROBAST and the results were synthesized based on the data format (with and without IBMs) to allow comparison.

Results: A total of 28 studies and 4,713 patients were included. Xerostomia was the most frequently investigated toxicity (17; 60.71 %). Sixteen (57.14 %) studies reported using radiomics features in combination with clinical or dosimetrics/dosiomics for modelling. High RoB was identified in 23 studies. Meta-analysis (MA) showed an area under the receiver operating characteristics curve (AUROC) of 0.82 for models with IBMs and 0.81 for models without IBMs (p value < 0.001), demonstrating no difference among IBM- and non-IBM-based models.

Discussion: The development of a PM based on sample-specific features represents patient selection bias and may affect a model's performance. Heterogeneity of the studies as well as non-standardized metrics prevent proper comparison of studies, and the absence of an independent/external test does not allow the evaluation of the model's generalization ability.

Conclusion: IBM-featured PMs are not superior to PMs based on non-IBM predictors. The evidence was appraised as of low certainty.

Keywords: Convolutional neural network; Dysphagia; Hearing loss; Hypothyroidism; Machine learning; Mucosistis; Osteoradionecrosis; Prediction model studies; Xerostomia.

Publication types

Meta-Analysis
Systematic Review
Review
Research Support, Non-U.S. Gov't

MeSH terms

Biomarkers
Head and Neck Neoplasms* / drug therapy
Humans
Machine Learning
Xerostomia*

Substances

Biomarkers