Assessment of the reliability of measurements taken on digital orthodontic models obtained from scans of plaster models in laboratory scanners. A systematic review and meta-analysis.

Abstract

In recent years, digital models have become increasingly popular among orthodontists, both for clinical and scientific purposes. It is, therefore, crucial to appropriately investigate their reliability. To this date, however, there has been no scientific, statistical investigation of their reliability as compared to the traditional gold standard-plaster models in the form of a meta-analysis. To evaluate the reliability and reproducibility of measurements taken on digital orthodontic models obtained from scanning plaster models in laboratory scanners compared to measurements taken directly on plaster models. Multiple electronic databases (PubMed, Scopus, Web of Science, Google Scholar and Cochrane Central Register of Controlled Trials) were searched for articles with no year or language limitations. The included original papers should have dealt with the accuracy and repeatability of the measurements conducted on plaster and digital models derived from laboratory scanners. In order to provide an adequate statistical analysis, the studies should have provided sufficient data, that is the difference of means (MDs) with standard deviations (SDs) for analysed measurements. In total, 25 types of non-standardised measurements were found in the evaluated studies. The quantitative analysis included papers that compared at least one of the parameters: upper/lower intermolar width, upper/lower intercanine width, overjet and overbite and provided standard deviation of the mean differences between measurements obtained on plaster and digital models from a laboratory scanner. GRADE and QUADAS tools were used to assess the quality of evidence, and they revealed substantial heterogeneity. Random-effects meta-analysis revealed no statistically significant differences for analysed measurements. Four of the analysed papers reported differences that may be considered clinically significant. No statistical significance between the direct measurements on plaster models and the digital ones taken from laboratory scanners could be identified by means of random-effects meta-analysis. The systematic review was registered in the PROSPERO database (ID CRD42020215411).