The role of artificial intelligence in intraoral scanning for complete-arch digital impressions: An in vitro study

Artificial intelligence (AI) is increasingly being integrated into intraoral scanners (IOS) to improve the quality of digital impressions. However, information on the accuracy of AI-assisted virtual models is limited. This study aimed to evaluate the effect of AI, specifically its application in filling mesh defects, by evaluating the accuracy of digital impressions of complete arches in an in vitro setting.

Three types of articulated maxillary models and antagonist models made from Type IV gypsum were used: an intact model (Model 1), a model with prepared teeth (FDI 13, 16) and missing teeth (FDI 14, 15) (Model 2), and a model with prepared teeth (FDI 13, 16) and missing teeth (FDI 14, 15, 17) (Model 3). A total of 120 scans were obtained using an IOS (3Shape Trios 5) with 20 digital impressions with AI assistance (AI-ON) and 20 without (AI-OFF). The files were compared with reference files generated by a desktop scanner (3Shape E3) and were analyzed for accuracy using a reverse engineering and metrology software (Geomagic Control X).

While AI-OFF shows better trueness in most relations, AI-ON improves precision and demonstrates significant advantages in accuracy, particularly in arch distortion and whole arch deviation measurements.

This study demonstrates that incorporating AI into digital impressions significantly improves the precision and accuracy of scans, particularly by enhancing consistency and reducing errors. Further studies are needed to explore the full potential of AI across different clinical scenarios and scanner types.

This study highlights the role of AI in improving the accuracy of intraoral scans by addressing mesh defects, contributing to more precise digital workflows in dentistry.