Deep Learning in the Diagnosis and Prognosis of Oral Potentially Malignant Disorders

Abstract

Oral potentially malignant disorders (OPMDs), characterized by a wide variety of types and diverse clinical manifestations, have always been difficult to diagnose and differentiate. All of them carry a risk of malignant transformation. In addition to pathological examination, which remains the gold standard, various auxiliary diagnostic tests are used in clinical practice. Deep learning, a branch of artificial intelligence, has been applied to medical image analysis. Among deep learning techniques, convolutional neural networks are commonly used for image segmentation, detection, classification, and computer-aided diagnosis. We reviewed several image analysis methods based on deep learning neural networks for the diagnosis and prognosis of OPMDs, including photographic images, autofluorescence images, exfoliative cytology images, histopathological images, and optical coherence tomography images. Additionally, we assessed the current limitations and challenges in applying deep learning to the diagnosis of OPMDs.