Accuracy of four intraoral scanners in oral implantology: a comparative in vitro study

Background: Until now, only a few studies have compared the ability of different intraoral scanners (IOS) to capture high-quality impressions in patients with dental implants. Hence, the aim of this study was to compare the trueness and precision of four IOS in a partially edentulous model (PEM) with three implants and in a fully edentulous model (FEM) with six implants. Methods: Two gypsum models were prepared with respectively three and six implant analogues, and polyether-ether-ketone cylinders screwed on. These models were scanned with a reference scanner (ScanRider((R))), and with four IOS (CS3600((R)), Trios3((R)), Omnicam((R)), TrueDefinition((R))); five scans were taken for each model, using each IOS. All IOS datasets were loaded into reverse-engineering software, where they were superimposed on the reference model, to evaluate trueness, and superimposed on each other within groups, to determine precision. A detailed statistical analysis was carried out. Results: In the PEM, CS3600((R)) had the best trueness (45.8 +/- 1.6 mu m), followed by Trios3((R)) (50.2 +/- 2.5 mu m), Omnicam((R)) (58.8 +/- 1.6 mu m) and TrueDefinition((R)) (61.4 +/- 3.0 mu m). Significant differences were found between CS3600((R)) and Trios3((R)), CS3600((R)) and Omnicam((R)), CS3600((R)) and TrueDefinition((R)), Trios3((R)) and Omnicam((R)), Trios3((R)) and TrueDefinition((R)). In the FEM, CS3600((R)) had the best trueness (60.6 +/- 11.7 mu m), followed by Omnicam((R)) (66.4 +/- 3.9 mu m), Trios3((R)) (67.2 +/- 6.9 mu m) and TrueDefinition((R)) (106.4 +/- 23.1 mu m). Significant differences were found between CS3600((R)) and TrueDefinition((R)), Trios3((R)) and TrueDefinition((R)), Omnicam((R)) and TrueDefinition((R)). For all scanners, the trueness values obtained in the PEM were significantly better than those obtained in the FEM. In the PEM, TrueDefinition((R)) had the best precision (19.5 +/- 3.1 mu m), followed by Trios3((R)) (24.5 +/- 3.7 mu m), CS3600((R)) (24.8 +/- 4.6 mu m) and Omnicam((R)) (26.3 +/- 1.5 mu m); no statistically significant differences were found among different IOS. In the FEM, Trios3((R)) had the best precision (31.5 +/- 9.8 mu m), followed by Omnicam((R)) (57.2 +/- 9.1 mu m), CS3600((R)) (65.5 +/- 16.7 mu m) and TrueDefinition((R)) (75.3 +/- 43.8 mu m); no statistically significant differences were found among different IOS. For CS3600((R)), For CS3600((R)), Omnicam((R)) and TrueDefinition((R)), the values obtained in the PEM were significantly better than those obtained in the FEM; no significant differences were found for Trios3((R)). Conclusions: Significant differences in trueness were found among different IOS; for each scanner, the trueness was higher in the PEM than in the FEM. Conversely, the IOS did not significantly differ in precision; for CS3600((R)), Omnicam((R)) and TrueDefinition((R)), the precision was higher in the PEM than in the FEM. These findings may have important clinical implications.