Can lithium disilicate ceramic crowns be fabricated on the basis of CBCT data?

Evaluating the fit of CAD/CAM lithium disilicate ceramic crowns fabricated on basis of direct and indirect digitalization of impressions by CBCT or of dental casts.

A metal model with a molar chamfer preparation was digitized (n = 12 per group) in four ways: IOS-direct digitalization using an Intra-Oral scanner (CS3600), cone-beam computed tomography scan (CBCT 1)-indirect digitalization of impression (CBCT-CS9300), CBCT 2-indirect digitalization of impression (CBCT-CS8100), and Extra-Oral scanner (EOS)-indirect digitalization of gypsum-cast (CeramillMap400). Accuracy of 3D datasets was evaluated in relation to a reference dataset by best-fit superimposition. Marginal fit of lithium disilicate crowns after grinding was evaluated by replica technique. Significant differences were detected for 3D accuracy by Mann-Whitney U and for fit of crowns by One-way ANOVA followed by Scheffe's post hoc (p = 0.05).

3D analysis revealed mean positive and negative deviations for the groups IOS (- 0.011 ± 0.007 mm/0.010 ± 0.003 mm), CBCT 1 (- 0.046 ± 0.008 mm/0.093 ± 0.004 mm), CBCT 2 (- 0.049 ± 0.030 mm/0.072 ± 0.015 mm), and EOS (- 0.023 ± 0.007 mm/0.028 ± 0.007 mm). Marginal fit presented the results IOS (0.056 ± 0.022 mm), CBCT 1 (0.096 ± 0.034 mm), CBCT 2 (0.068 ± 0,026 mm), and EOS (0.051 ± 0.017 mm).

The marginal fit of EOS and IOS, IOS and CBCT 2, and CBCT 2 and CBCT 1 showed statistical differences. The marginal fit of CBCT 1 and CBCT 2 is within the range of clinical acceptance; however, it is significant inferior to EOS and IOS.

The use of a CBCT enables clinicians to digitize conventional impressions. Despite presenting results within clinical acceptable levels, the CBCT base method seems to be inferior to Intra-Oral scans or to scanning gypsum models regarding the resulting accuracy and fit.