Related Experiment Video
Updated: Mar 15, 2026
Measuring the Complete-arch Distortion of an Optical Dental Impression
Published on: May 30, 2019
Correlation and agreement among outcome measures in implant impression-making research: A validation study
Alvaro Limones1, Gülce Çakmak2, Pedro Molinero-Mourelle3
1Department of Conservative Dentistry and Prosthodontics, Faculty of Odontology, University Complutense of Madrid, Madrid, Spain; Department of Prosthodontics, Indiana University School of Dentistry, Indianapolis, Indiana, USA.
Objective:
To validate Root Mean Square (RMS) error calculation as an outcome measure of trueness by assessing its correlation and agreement with linear and angular outcome measures in implant impression-making research.
Methods:
A maxillary implant cast with four parallel implants was scanned with a high-resolution desktop 3D scanner (Artec Micro II; Artec 3D; Senningerberg, Luxembourg) and scanned 100 times using an intraoral scanner (TRIOS 5; v 22.1.10; 3 Shape; Copenhagen, Denmark) under controlled conditions. Primary outcome measure was trueness, assessed by RMS error, linear and angular inter-implant deviations, and linear and angular intra-implant (platform-level) deviations. Secondary outcome measures included scanning time and number of photograms. Correlations among outcome measures were evaluated using Pearson's correlation coefficients with Bonferroni correction (α = 0.05), and agreement was assessed using Bland-Altman analysis for outcome measures that were directly comparable.
Results:
Very large positive correlations were found between RMS error and both linear inter-implant deviation (r = 0.831, p = 0.001) and linear intra-implant deviation (r = 0.864, p = 0.001). Angular deviations were not correlated with other outcome measures except between themselves (r = 0.669, p = 0.001). Scanning time was moderately correlated with most deviation outcomes (RMS error: r = 0.478, p = 0.001; linear inter-implant: r = 0.318, p = 0.027; and linear intra-implant deviations: r = 0.319, p = 0.026). The number of photograms was correlated only with scanning time (r = 0.678, p = 0.001) and showed no correlation with deviation measures. RMS error and linear deviations showed partial agreement, whereas angular deviations showed acceptable agreement.
Conclusions:
RMS error calculation is a valid outcome measure of trueness in implant impression-making research, demonstrating a very large correlation and partial agreement with linear deviation measures, yielding comparable, albeit not identical, results due to the distinct physical aspects captured by each outcome.
