Accuracy of intraoral scanning using modified scan bodies for complete arch implant-supported fixed prostheses

The accuracy of intraoral scanning techniques for complete arch implant-supported prostheses remains unclear.

The purpose of this in vitro study was to evaluate the accuracy of complete arch intraoral scanning using newly modified scan bodies.

A definitive cast with 6 parallel dental implants (6-246 subgroup, right first molar, right first premolar, right lateral incisor, left lateral incisor, left first premolar, and left first molar) was fabricated. By masking the implants with artificial gingiva, 2 other distinct definitive casts were obtained for 2 subgroups: the 4-24 subgroup, which included 4 implants (right first premolar, right lateral incisor, left lateral incisor, and left first premolar) and the 4-26 subgroup, which also included 4 implants (right first molar, right lateral incisor, left lateral incisor, and left first molar). Three methods were used to record implant location in these 3 subgroups: conventional impression making using the open-tray splinted technique (group CNV), intraoral scanning with the use of conventional scan bodies (group IOS-C), and intraoral scanning using newly modified scan bodies (group IOS-M). To assess accuracy, the best-fit algorithm was used, and root mean square (RMS) values were calculated. Descriptive statistics, including the median, interquartile range, and minimum and maximum values, were used to summarize the variables. Accuracy among different groups was compared, and the influence of the number of implants and the scan distance on the accuracy of group IOS-M was investigated. Appropriate methods were chosen based on the examination of normal distribution and homogeneity of variance, with 1-way analysis of variance (ANOVA) and the Tukey multiple comparison test for data normally (or log-normally) distributed and having equal variances and the Brown-Forsythe ANOVA test and Dunnett T3 multiple comparisons test for data normally (or log-normally) distributed but having unequal variances (α=.05). For data that did not follow a normal or log-normal distribution, the nonparametric Kruskal-Wallis test and Dunn multiple comparisons test was used.

The trueness of group IOS-M ranged from 15.5 to 37.5 µm, with a median (Q1, Q3) of 22.8 (20.3, 25.5) μm, better than that of group IOS-C (P<.001), ranging from 10.1 to 110.0 µm, with a median (Q1, Q3) of 32.1 (26.3, 47.6) μm. Although the trueness of group IOS-M was worse than group CNV (P<.001), ranging from 6.7 to 22.5 µm, with a median (Q1, Q3) of 14.9 (10.5, 17.8) μm, it was within the threshold deemed acceptable to produce clinically suitable complete arch restorations (<59 to 72 µm). The precision of group IOS-M, ranging from 7.2 to 40.8 µm, with a median (Q1, Q3) of 19.5 (16.4, 23.0) μm, was better than that of group IOS-C (P<.001), ranging from 9.8 to 86.8 µm, with a median (Q1, Q3) of 33.7 (25.2, 44.5) μm, but not as good as group CNV (P<.001), ranging from 7.0 to 34.3 µm, with a median (Q1, Q3) of 18.8 (14.3, 21.4) μm. No significant difference in accuracy was found in group IOS-M among subgroups 6-246, 4-26, and 4-24 (P>.05).

For complete arch implant scans, the modified scan body significantly improved the accuracy of intraoral scanning, with trueness <59 to 72 µm (threshold deemed acceptable to produce clinically suitable complete arch restorations). The accuracy of intraoral scanning using the modified scan bodies was not affected by the number of implants or the scan distance.