Diagnostic casts can incorporate different base designs and be manufactured using different vat-polymerization technologies. However, the influence of the interrelation between the base design and the 3D printing technology on the casts' final accuracy remains unclear.

A digital maxillary cast was obtained and used to generate 3 different base designs: solid (S group), honeycombed (HC group), and hollow (H group). The HC and H groups were subdivided based on the wall thickness of the cast design, resulting in 2 subgroups with thicknesses of 1 mm (HC1 and H1) and 2 mm (HC2 and H2) (N=100, n=10). Eleven reference cubes were added to each specimen for subsequent measurements. Specimens were manufactured by using 2 vat-polymerization 3D printers: Nextdent 5100 (ND group) and Sonic Mini 4K (SM4K group) and a resin material suitable for both 3D printers (Nextdent Model 2.0). A coordinate measuring machine quantified the linear and 3-dimensional discrepancies between the digital cast and each reference specimen. Trueness was defined as the average absolute dimensional discrepancy between the virtual cast and the specimens produced through additive manufacturing (AM), while precision was delineated as the standard deviation in dimensional discrepancies between the digital cast and the AM specimens. The data were analyzed using the Kruskal-Wallis and Mann-Whitney U pairwise comparison tests (α=.05).

For the NextDent group the trueness ranged from 21.83 µm to 28.35 µm, and the precision ranged from 17.82 µm to 37.70 µm. For the Phrozen group, the trueness ranged from 45.15 µm to 64.51 µm, and the precision ranged from 33.51 µm to 48.92 µm. The Kruskal-Wallis test showed significant differences on the x-, y-, and z-axes and in the 3D discrepancy (all P<.001). On the x-axis, the Mann-Whitney U test showed significant differences for the Phrozen group between the H-2 and H-1 groups (P=.001), H-2 and S groups (P<.001), and HC-2 and S groups (P=.012). On the y-axis, significant differences were found in the Phrozen group between the H-2 and H-1 groups (P=.001), the H-2 and S, H-1 and HC-1, and HC-1 and S groups (P<.001), the H-1 and HC-2 groups (P=.007), and the HC-2 and S groups (P=.009). The NextDent group exhibited significant differences, particularly among the HC-1 and H-2 groups (P=.004), H-1 (P=.020), and HC-2 (P=.001) groups; and on the z-axis significant differences were found in the Phrozen group between the H-2 and H-1 and S groups and the HC-2 group and H-1 and S groups (both P<.001). In the NextDent group, significant differences were found between the H-2 and HC-2 (P=.047) and HC-1 (P=.028) groups. For the 3D discrepancy analysis, significant differences were found in the Phrozen group between the H-2 and H-1 and S groups (P<.001), the H-1 and HC-2 groups (P=.001), the S and HC-1 and HC-2 groups (P<.001), and the H-1 and HC-1 groups (P=.002). In the NextDent group, significant differences were observed between the H-2 and HC-1 groups (P=.012).

The accuracy of digital casts depends on the manufacturing trinomial and base design of the casts. The honeycomb and hollow based designs provided the highest accuracy in the NextDent and Phrozen groups respectively for the material polymer tested. All specimens fell in the clinically acceptable range.