Mobile-Based AI Decision Support System For Early Oral Cancer Risk Screening

Abstract

Oral cancer is a serious global public health concern, especially in developing nations where poor oral hygiene, alcohol intake, and tobacco use greatly raise the risk of disease development. Improving survival rates requires early detection, but traditional screening techniques frequently rely on clinical infrastructure and qualified specialists, which may not be available in remote or resource-constrained areas.This study suggests a mobile-based artificial intelligence decision support system that combines risk factor assessment and medical picture analysis for early oral cancer risk screening. While a structured questionnaire module assesses behavioral and clinical risk factors, the system uses a Convolutional Neural Network (CNN) to analyze intraoral pictures and detect worrisome lesions. Grad-CAM, an Explainable Artificial Intelligence (XAI) approach, is used to highlight important areas in the oral picture in order to visually explain the model's predictions. The overall cancer risk level is calculated by combining the results of the questionnaire-based assessment with the picture classification module using a weighted risk fusion technique. The system classifies users as low, moderate, or high risk based on the final risk score and offers suitable healthcare advice. The goal of the proposed mobile decision support system is to improve awareness, enable early screening, and help medical personnel identify patients who need more clinical investigation.