Insights Into AI-Enabled Early Diagnosis of Oral Cancer: A Scoping Review

Abstract

Oral cancer (OC) remains a significant global health burden, with early detection being critical to improve prognosis and survival rates. Hence, early assessment is the primary challenge in improving OC outcomes due to gaps in specialist referrals and early diagnosis. Recently, artificial intelligence (AI) has emerged as a promising tool to enhance the early detection of oral potentially malignant disorders (OPMDs) and OC. We aimed to assess the various AI techniques for early OC diagnosis by searching PubMed for articles published between January 2016 and May 2025 using terms like "artificial intelligence", "deep learning", "machine learning", and "oral cancer". Following the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines, a comprehensive literature search was conducted. Of 88 articles retrieved, 28 met the inclusion criteria. Most studies originated from Southeast Asia and employed AI methods such as convolutional neural networks (CNNs), deep CNNs, artificial neural networks (ANNs), random forests, and decision trees. Standard data inputs included photographic and mobile images, with cytology and radiographic images also used. Deep CNNs showed the highest performance concerning sensitivity, specificity, and accuracy. Despite variability in techniques and datasets, overall diagnostic performance was promising. The study indicates that AI tools offer a strong potential for enhancing early diagnosis, particularly in low-resource settings.

Keywords: artificial intelligence; early detection; oral cancer; scoping review; early diagnosis.

Figure 1. Summary of search strategy