Investigation of Radiation-Induced Toxicity in Head and Neck Cancer Patients through Radiomics and Machine Learning: A Systematic Review

Table 1

Study selection criteria and research keywords according to the PICO model.

Selection criteria	Inclusion criteria	Exclusion criteria	EMBASE search via PICO	MEDLINE/PubMed search via PICO
P: population	Adults (age >18 years) affected by nonmetastatic HNSCC (nasopharynx; oral cavity; oropharynx; hypopharynx, larynx; nasal cavity; and paranasal sinus); salivary gland cancer	Pediatric patients (age < 18); non-HNSCC primary tumors; metastatic HNSCC cancer; and diagnosis of cutaneous squamous cell carcinoma or basal cell carcinoma of HNSCC	“Head and neck tumor”/exp OR “head and neck cancer”/exp	Head and neck tumor
I: intervention	Radiomics with artificial intelligence; radiomics-based machine-learning methods; and quantative radiographic phenotype analysis	Exclusion of radiomic analysis from the machine-learning method (exclusive analysis of biomarkers, genetic profiles, clinical data, etc.)	“Radiomics”/exp OR “machine learning”/exp	Radiomics OR machine learning
C: comparison	(Not explored)
O: outcome	Radiation-induced toxicity; radiation-induced toxicity risk	Prediction of survival outcomes; local disease response; prediction of HPV-status or nodal status; and automatic contouring implementation	“Radiation toxicity”/exp OR “radiation tolerance”/exp OR’radiation injury’/exp	Radiation toxicity/radiation tolerance/radiation injury