Purpose: The present study was conducted both in-vivo and in-vitro. The aim was to evaluate the trueness and precision of three-dimensional datasets acquired from digital impressions by chair-side intraoral scanning, or extraoral scanning of the impressions and gypsum casts of completely edentulous patient with multiple implants (All-on-4 protocol). Materials and Methods: Test groups of the in-vivo study:
Group (I): (IOS n=5) Direct digital scans using intraoral digital impression technique, Group (II) :(IMPR IL n=5) Digital scans of the conventional implant level impressions using extraoral scanner, Group (III):(CAST IL n=5) Digital scans of the stone casts - obtained from the conventional implant level impressions, Group (IV):(IMPR AL n=5) Digital scans of the conventional abutment level impressions using extraoral scanner and Group(V): (CAST AL n=5) Digital scans of the stone casts- obtained from the conventional abutment level impressions. Test groups of the in-vitro study: Control group: reference scan of the resin master models, Group (I): (IOS n=5) Direct digital scans of the master model using intraoral scanner, Group (II): (IMPR IL n=5) Digital scans of the conventional implant level impressions using extraoral scanner, Group(III):(CAST IL n=5) Digital scans of the stone casts - obtained from the conventional implant level impressions, Group (IV):(IMPR AL n=5) Digital scans of the conventional abutment level impressions, using extraoral scanner and Group (V): (CAST AL n=5) Digital scans of the stone casts- obtained from the conventional abutment level impressions. All STL datasets (IOS 1–5, IMPR 1–10 and CAST 1–10) for both in-vivo and invitro studies were imported into industrial reverse engineering software. The distance data were saved as an STL file and imported into a statistical program. The measurements were noted in tables and compared with the same measurements made with other scans. Results: The results of the in-vivo study revealed that the highest agreement (reliability-precision) between distance measurements was found with Group 1 (IOS), while the lowest agreement was found with Group 5 (CAST-AL)
[insignificant difference] However; all Cronbach's alpha coefficients showed very good agreement (range between 0.812 - 1.000). As regards angle measurements; Group 2 (IMPR-IL) showed the highest agreement, while Group 5 (CAST-AL) showed the lowest agreement [insignificant difference]. Cronbach's alpha coefficients showed very good agreement (range between 0.912 – 0.988). Regarding the invitro study, the highest agreement (reliability-precision) between distance measurements was found with Group 1(IOS) , while the lowest agreement was found with Group 5 (CAST-AL) [insignificant difference]. However; all Cronbach's alpha coefficients showed very good agreement (range between 0.871 - 1.000). As regards angle measurements; Group 5 (CAST-AL) showed the highest agreement while Group 3 (CAST-IL) showed the lowest agreement [insignificant difference]. Cronbach's alpha coefficients showed very good agreement (range between 0.749 – 0.992). Overall comparison between errors in distance and angle measurements (In-vitro study) revealed that there were no statistically significant differences between overall dimensional changes in distance or angle measurements in all groups. Conclusion: It was concluded that the five digital impression techniques tested were comparable under all conditions, and the misfits were all within the clinically acceptable range