Role of Artificial Intelligence in the Diagnosis of Oral Squamous Cell Carcinoma: A Systematic Review

Oral squamous cell carcinoma (OSCC) is a serious worldwide health issue. Early OSCC identification by the analysis of digital oral photos is possible with the combination of artificial intelligence (AI) and computer vision. The purpose of this systematic review was to evaluate the current evidence on the role of AI in the diagnosis of OSCC, focusing on the diagnostic performance, methodologies employed, and potential limitations of AI applications in this context. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines to search for relevant studies across PubMed, Scopus, Web of Science, and Cumulative Index to Nursing and Allied Health Literature (CINAHL). In these databases, we found 286 studies, which were first screened for duplicates and then assessed on inclusion and exclusion criteria. Only 11 studies were found most relevant and were included in this study. These studies were also assessed for risk of bias using the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool. Numerous studies have shown impressive results for this job, frequently covering about 1000 photos and regularly reaching sensitivity rates above 85% with accuracy rates above 90%. The review examines these research in detail, providing insight into their methods, which include the application of contemporary machine learning and pattern recognition techniques in conjunction with various supervision techniques. However, because various datasets are utilized in different articles, it can be difficult to compare the results. In light of these results, this study emphasizes how urgently the area of OSCC detection needs more solid and trustworthy datasets. Additionally, it emphasizes how sophisticated methods like ensemble learning, multi-task learning, and attention mechanisms can be used as essential instruments to improve the sensitivity and accuracy of OSCC identification in oral photos. Together, these observations highlight how AI-driven methods for early OSCC diagnosis have the potential to greatly enhance patient outcomes and medical procedures.