Browsing by Author "Alhablain, Eman"

Now showing 1 - 1 of 1

Restricted
Evaluating the Accuracy of Artificial Intelligence in Gingivitis Diagnosis
(Saudi Digital Library, 2026) Alhablain, Eman; Sedky, Nabila; Alsuhaibani, Mohammed
Introduction: Gingivitis is one of the most prevalent oral inflammatory conditions and is reversible when detected and managed at early stages. Conventional diagnosis relies on clinical examination, which may be influenced by examiner experience and inter-examiner variability. In recent years, artificial intelligence has demonstrated promising potential in supporting non-invasive oral disease diagnosis. However, most existing studies have focused on binary classification of gingivitis without adequately addressing disease severity. Objectives: This study aimed to evaluate the accuracy of artificial intelligence models in diagnosing and grading gingivitis severity using intraoral photographs, based on a multi-class classification framework utilizing the Modified Gingival Index (MGI). Materials and Methods: A cross-sectional diagnostic accuracy study was conducted. Frontal intraoral photographs of anterior teeth were used, and gingivitis severity was visually graded according to the MGI by calibrated examiners. Multiple deep learning models were trained to classify gingivitis severity into five classes. Model performance was evaluated using accuracy, recall, precision, macro-averaged performance metrics, and confusion matrix analysis. Results: The AI models demonstrated satisfactory performance in grading gingivitis severity, with higher accuracy observed in moderate and severe categories. Relatively lower performance was noted in early-stage gingivitis classification, reflecting subtle visual differences associated with mild inflammation. Multi-class classification provided greater clinical relevance compared to binary diagnostic approaches. Conclusion: The findings of this study indicate that artificial intelligence models can accurately diagnose and grade gingivitis severity using intraoral photographs. These results support the potential role of AI as a non-invasive screening and monitoring tool in preventive dentistry and dental public health. Further studies are recommended to validate these findings across diverse populations and real-world clinical settings.
13 0