Systematic Review and Meta-Analysis on the Accuracy of Artificial Intelligence Algorithms in Individuals Gender Detection Using Orthopantomograms

Abstract

The integration of artificial intelligence (AI) into dental imaging has led to significant advancements, particularly in the analysis of panoramic radiographs, also known as orthopantomograms (OPGs). One emerging application of AI is in determining gender from these radiographs, a task traditionally performed by forensic experts using manual methods. This systematic review and meta-analysis aim to evaluate the accuracy of AI algorithms in gender determination using OPGs, focusing on the reliability and potential clinical and forensic applications of these technologies. A systematic review and meta-analysis were conducted according to PRISMA guidelines. The study included research articles that utilised AI algorithms for gender detection based on OPG images. Five major databases were searched, and studies were selected based on strict inclusion and exclusion criteria. The analysis focused on studies that reported accuracy, sensitivity, and specificity of AI models. Statistical analyses were performed using R software, including forest plots and funnel plots, to evaluate the diagnostic performance and potential publication bias. The meta-analysis included 13 studies, yielding a pooled accuracy estimate of 88.66%. The results demonstrated high specificity among the AI models, with some studies achieving accuracy rates as high as 99.20%. However, there was variability in sensitivity across different studies, indicating that some models are more reliable than others depending on the dataset and features used. The funnel plot analysis suggested slight asymmetry, indicating potential publication bias or heterogeneity. AI models show significant potential in accurately determining gender from OPG images, with some models achieving near-perfect accuracy. Continued research is needed to enhance model consistency and expand the applicability of these tools across different demographic groups.

Ähnliche Arbeiten