Scoping review evaluates AI and machine learning in predicting malignant transformation of oral disorders, suggesting clinical implications.
Key Points
This review aims to assess the use of AI and machine learning with biomarkers for predicting malignant transformation in oral potentially malignant disorders.
Analyzed ten retrospective studies utilizing AI/ML algorithms.
Incorporated biomarkers including gene expression panels and histomorphometric features.
Evaluated predictive accuracy and feasibility of various biomarker integration methodologies.
Histology-derived features showed the greatest clinical feasibility for predicting malignant transformation.
Variable methodologies limited the generalizability of findings.
The need for prospective validation and multimodal biomarker integration was highlighted.