To assess the accuracy of scanning technologies for constructing facial prostheses
on human faces.Our systematic search was performed on five databases. Studies reporting
on human volunteers (P) whose faces were scanned with a scanning technology were eligible.
The anthropometrical interlandmark distances (ILDs) were used as indicators of accuracy;
the ILDs are measured on the virtual models (I) and directly on the faces (C). The
virtual models deviated from their true values (O). Studies reporting the measurements
on patients with or without facial deformities were included, but cadavers or inanimate
objects were reasons for exclusion. We performed a mean difference (MD) / standardized
MD analysis with a random effect model. The difficulties regarding the scanning procedure
mentioned in the articles were also assessed.We found 3723 records after duplicate
removal. Twenty five articles were eligible for the qualitative review, and ten articles
were included in the quantitative synthesis. Eight different ILDs were compared in
MD analyses. The differences were between -0.54-0.43 mm. We also performed a regional
three-dimensional analysis to compare scanning technologies in each major region.
No significant differences were found in any of the regions and axes. The most mentioned
difficulties were artifacts due to motion or blinking.The results suggest no systematic
skew in linear dimensions neither between direct caliper measurements nor between
measurements on the scanned models, scanning technologies, or facial regions.
