<![CDATA[Early identification of malignant transformation in oral leukoplakia is crucial to prevent progression to Oral Squamous Cell Carcinoma (OSCC). However, conventional clinical assessment still faces limitations in terms of accuracy and interpretability, highlighting the need for reliable and transparent predictive approaches. This study aims to evaluate the performance of interpretable machine learning models in predicting malignant transformation of oral leukoplakia and OSCC based on clinical and histopathological data. A retrospective dataset consisting of 237 patient medical records was analyzed using several interpretable models, including Explainable Boosting Machine (EBM), Generalized Additive Models (GAM), and Symbolic Regression, and compared with black-box models such as Random Forest and Deep Neural Network. Model performance was evaluated using accuracy, sensitivity, specificity, and area under the curve (AUC). The results demonstrate that interpretable models achieve competitive predictive performance compared to black-box models while offering superior transparency and interpretability. Feature contribution analysis indicates that histopathological characteristics are the most influential factors in malignancy prediction. These findings suggest that interpretable machine learning models have strong potential as clinical decision support systems for early oral cancer detection.]]>
Copyrights © 2026
