Automated Diagnosis Of OSCC And Related Leukoplakia In Pathological Images

A novel deep learning model enables automated precise auxiliary diagnosis of OSCC and related leukoplakia lesions. We proposed a novel deep learning model RRGNet, which was improved based on the RegNet architecture. By introducing the Ghost module and the residual channel attention module in the final stage of the model, and adopting strategies such as label smoothing, Mixup data augmentation, and SWALR learning rate adjustment, the feature extraction efficiency and computational cost were significantly optimized. Trained on a three-class dataset comprising oral squamous cell carcinoma (OSCC) and related leukoplakia lesions, the proposed model was benchmarked against established architectures including GhostNet, HRNet, RegNet-50, and ViT. Experiments show RRGNet achieves enhanced accuracy and high AUC, which is better than all the comparison models. At the same time, the number of parameters of RRGNet is significantly lower than that of other models, showing excellent computational efficiency and generalization ability. In addition, activation heat map analysis shows that the model can accurately focus on the lesion area and has strong clinical interpretability. The RRGNet model performed well in the three-classification task, providing pathologists with an efficient and reliable auxiliary diagnostic tool. It provides technical support for the early detection and precise treatment of OSCC.