AI-Enhanced Deep Learning Framework for Prognostic Oral Cancer Detection and Hierarchical Classification using Multi-Spectral Imaging and Genomic Markers

Oral cancer is a widespread cancer all over the world. Mostly men are affected by oral cancer and main reasons for oral cancer are tobacco chewing, alcohol consumption and HPV. Early detection leads to an increase in the recurrence rate of the patient and is helpful for prognosis of oral cancer. Currently various screening procedures are made for diagnosis of oral cancer. Hence Imaging techniques are utilized for the screening of oral cancer as part of the diagnostic process. These techniques encompass Magnetic Resonance Imaging (MRI), Computed Tomography (CT), and Positron Emission Tomography (PET). MRI is mostly preferable for diagnosis of squamous cell carcinoma in oral cavity.In proposed method, squamous cell carcinoma is automatically detected and classified using Digital Image Processing technique and Machine Learning algorithm. This process consists of three main steps, initially enhancement of image is performed. Then tumor region is segmented using K-means segmentation algorithm. Secondly, features from segmented image are extracted based on TNM classification of squamous cell cerci no main oral cavity. Finally, with the help of extracted features oral cancer image is classified into various stages. In this proposed method, cancer location is accurately identified and small lesions which present in depth of the tissue are also classified. Time consumption for this process is lesser than the manual method of classification of oral cancer.