PURPOSE: To compare the accuracy (trueness, precision) of direct and indirect scanning
CAD/CAM methods. METHODS: A master cast with prepared abutments and edentulous parts
was created from polymethyl methacrylate (PMMA). A high-resolution industrial scanner
was used to create a reference model. Polyvinyl-siloxane (PVS) impressions and digital
impressions with three intraoral scanners (iTero, Cerec, Trios) were made (n=10 for
each) from the PMMA model. A laboratory scanner (Scan CS2) was used to digitize the
sectioned cast made from the PVS impressions. The stereolithographic (STL) files of
the impressions (n=40) were exported. Each file was compared to the reference using
Geomagic Verify software. Six points were assigned to enable virtual calliper measurement
of three distances of varying size within the arch. Methods were compared using interquartile
range regression and equality-of-variance tests for precision, and mixed-effects linear
regression for trueness. RESULTS: The mean (SD) deviation of short distance measurements
from the reference value was -40.3 (79.7) mum using the indirect, and 22.3 (40.0)
mum using the direct method. For the medium distance, indirect measurements deviated
by 5.2 (SD: 111.3) mum, and direct measurements by 115.8 (SD: 50.7) mum, on average;
for the long distance, the corresponding estimates were -325.8 (SD: 134.1) mum with
the indirect, and -163.5 (SD: 145.5) mum with the direct method. Significant differences
were found between the two methods (p<0.05). CONCLUSIONS: With both methods, the shorter
the distance, the more accurate results were achieved. Virtual models obtained by
digital impressions can be more accurate than their conventional counterparts.
