AI-based Analysis of Oral Lesions Using Novel Deep Convolutional Neural Networks for Early Detection of Oral Cancer

Abstract Background: Artificial intelligence (AI) applications in oncology have been developed rapidly with reported successes in recent years. This work aims to evaluate the performance of deep convolutional neural network (CNN) algorithms for the classification and detection of oral potentially malignant disorders (OPMDs) and oral squamous cell carcinoma (OSCC) in oral photographic images. Methods: A dataset comprising 980 oral photographic images was divided into 365 images of OSCC, 315 images of OPMDs and 300 images of non-pathological images. Multiclass image classification models were created by using DenseNet-169, ResNet-101, SqueezeNet and Swin-S. And multiclass object detection models were fabricated by using faster R-CNN, YOLOv5, RetinaNet and CenterNet2. Results: The AUC of multiclass image classification of CNN models was 0.71-1.00 and 0.80- 0.98 on OSCC and OPMDs, respectively. The AUC of multiclass CNN-base object detection models was 0.81-0.91 and 0.34-0.64 on OSCC and OPMDs, respectively. In comparison, CNN- based classification models exhibited a sensitivity and specificity of 0.72-0.99, 0.83-0.99 on OSCC and 0.74-0.95, 0.88-0.97 on OPMDs, respectively. These values were inline with the performance of oral and maxillofacial surgeons and superior to those of general practictioners (GPs). Conclusions: CNN-based models have potential for the identification of OSCC and OPMDs in oral photographic images and are expected to be a diagnostic tool to assist GPs for the early detection of oral cancer.