Accuracy of impression scanning compared with stone casts of implant impressions

Statement of problem. Accurate virtual implant models are a necessity for the fabrication of precisely fitting superstructures. Purpose. The purpose of this in vitro study was to evaluate different methods with which to build an accurate virtual model of a 3-dimensional implant in the oral cavity; this model would then be used for iterative computer-aided design and computer-aided manufacturing (CAD-CAM) procedures. Material and methods. A titanium master model with 3 rigidly connected implants was manufactured and digitized with a noncontact industrial scanner to obtain a virtual master model. Impressions of the master model with the implant position locators (IPL) were made using vinyl siloxanether material. The impressions were scanned (Impression scanning technique group). For the transfer technique and pick-up technique groups (each group n=20), implant analogs were inserted into the impression copings, impressions were made using polyether, and casts were poured in Type 4 gypsum. The IPLs were screwed into the analogs and scanned. To compare the virtual master model with each virtual test model, a CAD interactive software, ATOS professional, was applied. The Kruskal-Wallis test was subsequently used to determine the overall difference between groups, with the Mann-Whitney U test used for pairwise comparisons. Through Bonferroni correction, the alpha-level was set to.017. Results. The outcome revealed a significant difference among the 3 groups (P<.01) in terms of accuracy. With regard to total deviation, for all axes, the transfer technique generated the greatest divergence, 0.078 mm (+/- 0.022), compared with the master model. Deviation with the pick-up technique was 0.041 mm (0.009), with impression scanning generating the most accurate models with a deviation of 0.022 mm (+/- 0.007). Conclusions. The impression scanning method improved the precision of CAD-CAM-fabricated superstructures.