Experimental radiological study on the readiness for real-world clinical application of an artificial intelligence-based program in dentistry

Abstract

Background. Artificial intelligence (AI) is rapidly evolving and gaining applications in various fields of medicine, including dentistry. In particular, AI systems for analyzing radiological data hold the potential to enhance diagnostic accuracy and efficiency. However, challenges such as compliance with national regulatory standards and the adaptation of AI technologies to clinical practices in the Russian Federation remain significant obstacles. Aims. To evaluate the readiness of AI for real-world clinical application in dentistry in the Russian Federation by comparing the performance of the Diagnocat system with experienced dentists in the interpretation of radiological data. Materials and methods. The study included 100 radiological examinations (CBCT), each corresponding to a unique patient (55 women, 45 men; mean age 42.9±12.1 years). The evaluated parameters included the frequency of diagnostic errors, diagnostic accuracy (e.g., caries, periodontal diseases), and the average time required for data analysis. To assess the statistical significance of differences, the χ² test was applied. The results obtained using Diagnocat were verified and compared with those of clinicians, as confirmed by independent experts. Results. Diagnocat demonstrated no errors in the anterior region and fewer errors in the posterior region compared to dentists (4.86% vs. 7.29%). Diagnocat’s average analysis time was significantly shorter (4.18 minutes) compared to that of dentists (25.05 minutes). While Diagnocat showed high accuracy, some limitations were identified, including difficulties in interpreting complex anatomical structures and pathology severity. Conclusions. Diagnocat shows significant potential for integration into clinical dentistry in the Russian Federation, offering advantages in speed and accuracy under high workload conditions. However, further algorithm refinements and adaptations to Russian medical standards are necessary. The study highlights the value of combining AI capabilities with the clinical expertise of dentists to achieve optimal diagnostic outcomes.

Ähnliche Arbeiten