Accuracy in the digital workflow for complete-arch implant rehabilitation

Abstract

Fabrication of complete-arch implant-supported fixed prosthesis using the digital workflow has become increasingly popular for rehabilitation of edentulous patients. However, whether the accuracy of the workflow is clinically acceptable or not and where and by how much the distortions occurred in the workflow remain unknown. The aim of this study is to identify and quantify the errors generated from every stage of the digital workflow for fabrication of complete-arch implant-supported prosthesis.

Firstly, the accuracies of 4 laboratory scanners were evaluated using a novel calibration block particularly designed for complete-arch implant situations. Secondly, a standard edentulous maxillary model with 6 dental implants inserted was examined by a coordinate measuring machine (CMM) as the reference. The performance of an intraoral scanner was then evaluated with or without a novel auxiliary device with well-defined landmarks. Thirdly, 2 types of implant scan bodies were investigated to reveal the effects of geometry and repeated repositioning of the scan bodies on the scanning and the following computer-aided design (CAD) process. Fourthly, A systematic review was performed on methodologies for assessing the implant-framework misfit. Finally, milled complete-arch implant-supported frameworks using the complete-digital and the analogue-digital workflow were compared. The overall accuracy and the error distribution in each stage of the digital workflow were then computed.

The 4 laboratory scanners showed expanded uncertainties less than 20 µm/0.183°. Tilted implants significantly affected the laboratory scanning accuracy but the inter-implant distance did not. The trueness of intraoral scanning was significantly improved from 76 µm/0.283° to 38.8 µm/0.276° by the auxiliary device, particularly in the second quadrant. Besides, repeated and random repositioning of the scan bodies reduced the reproducibility of scanning. The virtual alignment process in the CAD stage induced 15 µm/0.088° of distortion. The dome-shaped scan body was more accurate than the cuboidal one. The systematic review summarized and discussed advantages and disadvantages of dimensional and technical techniques for assessing the implant-framework misfit. The last study showed the milling procedure contributed ~40 µm/0.074° of distortion. The expanded uncertainty of the overall digital workflow was ~ 150 µm and ~0.8°. The former was within the clinical threshold but not the later. Most distortions were derived from data acquisition and manufacturing, while the distortion induced at the CAD stage was minimal.

Laboratory scanning and CAD processing generated insignificant distortions in the digital workflow for complete-arch implant prosthesis. Intraoral scanning, on the other hand, induced greater errors although that can be reduced by adding well-defined landmarks. The implant scan body was a crucial link to transfer analogue to digital, of which accuracy depended on its geometry and positioning fit. Under the controlled experimental environment, data acquisition and manufacturing created most errors in the digital workflow and the overall linear distortion was clinical acceptable but not the angular one. It is expected that greater errors would be created under clinical conditions and the errors induced at every stage would propagate and accumulate along the digital workflow.