RESUMO
PURPOSE: To develop a mobile application (app) for oral cancer screening. METHODS: The app was developed using Android system version 4.4.2, with JAVA language. Information concerning sociodemographic data and risk factors for oral cancer development, e.g., tobacco and alcohol use, sun exposure and other contributing factors, such as unprotected oral sex, oral pain and denture use, were included. We surveyed a population at high risk for oral cancer development and then evaluated the sensitivity/specificity/accuracy and predictive values of clinical oral diagnosis between two blinded trained examiners, who used movies and data from the app, and in loco oral examination as gold-standard. RESULTS: A total of 55 individuals at high risk for oral cancer development were surveyed. Of these, 31% presented homogeneous/heterogeneous white lesions with potential of malignancy. The clinical diagnoses performed by the two examiners using videos were found to have sensitivity of 82%-100% (average 91%), specificity of 81%-100% (average 90.5%), and accuracy of 87.27%-95.54% (average 90.90%), as compared with the gold-standard. The Kappa agreement value between the gold-standard and the examiner with the best agreement was 0.597. CONCLUSION: Mobile apps including videos and data collection interfaces could be an interesting alternative in oral cancer research development.