Optical 3D scans for orthodontic diagnostics performed on full-arch impressions. Completeness of surface structure...

The purpose of this work was to evaluate the completeness of surface structure representation offered by full-arch impression scans in different situations of tooth (mal)alignment and whether this completeness could be improved by performing rescans on the same impressions reduced sequentially to different levels of gingival height and by adding extra single scans to the number of single scans recommended by the manufacturer.

Three pairs of full-arch resin models were used as reference, characterized either by normal occlusion, by anterior diastematic protrusion (and edentulous spaces in the lower posterior segments), or by anterior crowding. An alginate impression of each arch was taken and digitized with a structured-light scanner, followed by three rescans with the impression cut back to 10, 5, and 1 mm of gingival height. Both the initial scan and the rescans were performed both with 19 basic single scans and with 10 extra single scans. Each impression scan was analyzed for quantitative completeness relative to its homologous direct scan of the original resin model. In addition, the topography of voids in the resultant digital model was assessed by visual inspection.

Compared to the homologous reference scans of the original resin models, completeness of the original impression scans--in the absence of both gingival cutback and extra single scans--was 97.23 ± 0.066% in the maxilla or 95.72 ± 0.070% in the mandible with normal occlusion, 91.11 ± 0.132% or 96.07 ± 0.109% in the arches with anterior diastematic protrusion, and 98.24 ± 0.085% or 93.39 ± 0.146% in those with anterior crowding. Gingival cutback and extra single scans were found to improve these values up to 100.35 ± 0.066% or 99.53 ± 0.070% in the arches with normal occlusion, 91.77 ± 0.132% or 97.95 ± 0.109% in those with anterior diastematic protrusion, and 98.59 ± 0.085% or 98.96 ± 0.146% in those with anterior crowding.

In strictly quantitative terms, the impression scans did capture relatively large percentages of the total surface. However, the topographic examinations revealed that regions essential for orthodontic model analysis were missing. The malocclusion models were particularly affected. Thus, impression scans performed with structured-light scanners cannot replace scans of positive casts for diagnostic use in orthodontics.