Accuracy of implant abutment level digital impressions using stereophotogrammetry in edentulous jaws: an in vitro pilot study

Background In edentulous jaws, factors such as the number of implants, cross-arch distribution, and the angle among implants may affect the accuracy of the implant impression. This study explored factors influencing the accuracy of implant abutment-level digital impressions using stereophotogrammetry in edentulous jaws. Methods Two standard all-on-4 and all-on-6 models of edentulous jaws were constructed in vitro. In the stereophotogrammetry group (PG), the implant digital impression was made using stereophotogrammetry and saved as an STL file. In the conventional group (CNV), the impression was made using the open-tray splint impression technique. An electronic and optical 3D measuring instrument was used to scan the standard model and the conventional plaster model to obtain STL files. Using 3D data processing software (GOM Inspect Pro, Zeiss), the distance and angle between the abutments in the CNV impression and the PG impression were measured and compared with the data from the standard model. Results The distance deviation in the PG and the CNV was 145 +/- 196 mu m and 96 +/- 150 mu m, respectively, with a significant difference (P < 0.001). The angle deviation in the PG and the CNV was 0.82 +/- 0.88 degrees and 0.74 +/- 0.62 degrees, respectively, with no significant difference (P = 0.267). In the PG, the distance deviation was negatively correlated with the distance between implants (r = -0.145, P = 0.028) and positively correlated with the angle of implants (r = 0.205, P = 0.002). The angle deviation was negatively correlated with the distance between implants (r = -0.198, P = 0.003) and positively correlated with the angle of implants (r = 0.172, P = 0.009). In the CNV, the effect of inter-implant distance on impression accuracy was also shown by Spearman correlation analysis: r = 0.347 (P < 0.001) for distance deviation and r = -0.012 (P = 0.859) for angle deviation. The effect of inter-implant angulation on impression accuracy deviation was r = -0.026 (P = 0.698) for distance deviation and r = 0.056 (P = 0.399) for angle deviation. Conclusions The CNV method is closer to the real value of the original model. The distance between implants and the distribution angle had a weak correlation with the accuracy of digital impressions but no significant correlation with the accuracy of traditional impressions.