Evaluation of the Ability to Predict Subsequent Metastasis of Early Oral Squamous Cell Carcinoma Using PET Radiomics...

Area of Science:

• Oncology
• Radiology
• Artificial Intelligence

Background:

• Oral squamous cell carcinoma (OSCC) presents a risk of late metastasis, even in early stages.
• Accurate prediction of metastasis is crucial for timely intervention and improved patient outcomes.

Purpose of the Study:

• To develop and validate radiomics-based machine learning (ML) models for predicting the risk of late metastasis in early-stage OSCC.
• To assess the efficacy of 18F-FDG PET-CT derived radiomics features in metastasis prediction.

Main Methods:

• Retrospective analysis of 109 patients with T1 or T2 OSCC who underwent preoperative 18F-FDG PET-CT.
• Extraction and selection of radiomics features from PET images.
• Development and evaluation of ML models to predict late cervical lymph node metastasis.

Main Results:

• The most effective ML model achieved an area under the curve (AUC) of 0.977 and an accuracy of 87.5%.
• Radiomics features from PET images were successfully utilized to build predictive models.

Conclusions:

• Machine learning models incorporating radiomics features from PET images demonstrate significant utility in predicting late metastasis in early-stage OSCC.
• This approach offers a promising non-invasive tool for risk stratification and personalized treatment planning.