Validation of the Efficacy of Artificial Intelligence in Detecting the Common Dental Diseases Prevailing in Patients

Abstract

Aim: To access the efficiency, reliability and accuracy of detecting dental ailments by Artificial intelligence(AI) machines

Material and Methods: In this study, 454 volunteers provided informed consent, resulting in 227 Orthopantomograms (OPGs) and 681 intraoral images. An AI kiosk captured images of their anterior teeth, upper arch, and lower arch to assess dental parameters such as caries, edentulous spaces, malocclusions, root stumps, bone loss, furcation defects, impactions, nerve and sinus involvement, spacing and crowding, The AI's identification of dental issues was validated by a dentist for accuracy, and the collected data was analysed to evaluate its significance. The area under the receiver operating characteristic (ROC) curve allowed a comparison of efficacy between network and examiner diagnosis.

Results: Crosstab statistics were computed to determine the test's sensitivity and specificity. Various aspects and variables were recorded and are presented in the following tables. The positive predictive value (PPV) and negative predictive value (NPV) were calculated to assess the accuracy of the AI method for identifying dental findings. IBM SPSS version 25 was used for these calculations.

Conclusion: The machine learning algorithms developed in this study exhibit strong performance and enable effective implementation by dental and non-dental professionals. Clinicians are encouraged to utilize the algorithms from this study for early intervention and treatment strategies.