Accuracy of current intraoral scanning systems for full-arch impressions: An in vitro study

This in vitro study aimed to evaluate the accuracy of seven intraoral scanners (IOSs) for full-arch impressions. Correlations among three evaluation parameters summarising three-dimensional (3D) difference values were also analysed.

Custom-made complete-arch reference models were initially scanned using a reference scanner (ATOS III, Triple Scan MV60). Full-arch scans were then performed using seven IOSs (n = 20 each): Aoralscan 3, Emerald Green (EME), iTero 5D, Medit i900, Primescan (PRI), Primescan 2, and Trios 5. Conventional impressions (CONVs) (n = 20) were obtained and digitised using a laboratory scanner (inEOS X5). All data were exported as standard tessellation language files. Trueness and precision were determined using (90_10)/2-percentile values. Statistical analyses were performed using analysis of variance with Games-Howell post-hoc test and linear mixed-effects model. Correlations among absolute 80-percentile, absolute mean, and (90_10)/2-percentile were analysed using linear regression with Pearson's correlation coefficient.

Trueness ranged from 24.5 ± 3.2 μm (CONV) to 95.0 ± 18.8 μm (EME), while precision ranged from 13.3 ± 2.6 μm (CONV) to 59.6 ± 22.4 μm (EME). CONV outperformed all IOSs. PRI showed the highest accuracy (trueness: 30.5 ± 5.3 μm, precision: 18.7 ± 5.6 μm). The evaluated parameters showed highly significant linear correlations (p < 0.001).

All tested IOSs show full-arch deviations of <100 µm. The evaluation parameters are highly correlated, allowing for comparability.

IOSs reach clinically acceptable accuracy values for full-arch impressions. However, high-precision CONVs still achieve the highest trueness and precision for full-arch impressions.