A Comperative Study of Use Of Artificial Intelligence in Oral Radiology Education

Abstract

Purpose: The aim of this study is to compare the efficacy of artificial intelligence use in oral radiology learning in the undergraduate dental students. Materials&Methods: Fifty third-year students in the University of Lokman Hekim were detected images with the artificial intelligence method (AI) and standard lecture method (SL) for anatomical landmarks in panoramic radiographs. SL consisted of a frontal lecture through a standardized presentation. CranioCatch model (Eskisehir, Turkey) was used as deep learning-based artificial intelligence model. One panoramic image was loaded to the application and anatomic landmarks were detected by teacher, students were asked to mark. AI recorded and scored students answers. A questionnaire study was conducted for the perception of students in terms of validity and reliability regarding assessment and evaluation for each methods. Results: 50 undergraduate students (26 female,24 male) answered 7questions, 5-point Likert type. The conformity to the normal distribution was evaluated with the Shapiro-Wilk test and the graphical approach (Normal Q-Q Plot). The values did not conform to the normal distribution. As a result of the reliability analysis performed for the measurement tool, the Cronbach’s Alpha coefficient was found 0.828. Wilcoxon Test was used to test the significance of the difference between each methods. There is a statistically significant difference between the mean values of evaluation measurements(p=0.014). AI was higher than the mean of evaluation measurement values compared to SL. Conclusion: AI models have performed very well in measurement and evaluation in oral radiology learning.