On a master model of a partially edentulous maxilla with four implants in posterior region, different impression techniques were investigated: conventional custom impression tray (CIT), customized foil tray (CFT), chairside 3-D printed impression tray using the SHERA system (3DS), and the Primeprint system (3DP). Each impression technique was investigated separately, resulting in four study groups.

All experiments were conducted under laboratory conditions (room temperature, 23° C ± 1° C; humidity, 50% ± 10%).

For a better overview, Fig. 1 displays a flow scheme of the investigation.

Fig. 1

Flow scheme of the study protocol

Implant master model and acquisition of reference dataset

An implant master model (IMM) of the maxilla, known from a previous study, was used as the patient equivalent [41]. On a stainless steel base plate (alloy 1.4301 [Ni–Cr], 100 × 100 × 15 mm), four stainless steel tubes were placed in the implant position of the first premolar (according to Federation Dentaire Internationale [FDI] schemes #14 and #24) and the first molar (#16 and #26, Fig. 2a). In each tube, an implant (4.1 mm diameter, 11.5 mm length, T3 non-platform switched tapered implants; Biomet 3i, Palm Beach Gardens, FL, USA) was adhesively luted (AGC-Cem Automix System, C. HAFNER, Wimsheim, Germany). The implants at positions #14 and #24 were inclined 15° in the buccal direction, while the implants were kept straight with 0° inclination at #16 and #26. In addition, a rectangular cube (10 × 10 × 20 mm) was placed at the center of the IMM perpendicular to the base plate to serve as a reference point. The IMM was finalized by modeling a partially edentulous maxilla with teeth in regions #17, #13, and #23 using pink and tooth-colored denture plastics (PalaXpress, Kulzer, Hanau, Germany; Fig. 2b).

Fig. 2

a IMM with implants #14/#24 and #16/#26. b Finalized IMM with model teeth #17,#13 to #23

The IMM was digitized with an X-ray computed tomography (TomoScope S, Werth Messtechnik, Giessen, Germany; measurement parameters: 225 kV, 100 ms, 1 mm tin filter, 60 µm voxel size, 2200 sections, surface resolution < 6 µm, linear accuracy < 4 µm). On the resulting standard tessellation language (STL) dataset, the implant–abutment interface points (IAIPs) were determined for each of the four implants using WinWerth software (Werth Messtechnik). Finally, a coordinate system was created on the reference cube. The z-axis was defined as the intersection of the two symmetry planes of the outer surfaces of the cube. The line of intersection between the left outer plane and upper plane of the cuboid formed the x-axis. The intersection line between the rear, outer, and upper planes of the cuboid formed the y-axis. The origin was placed at the intersection of the x- and y-axes. All the coordinate systems for the measurements were created in the similar manner (Fig. 3).

Fig. 3

Reference cube of IMM with coordinate system: x-axis (red), y-axis (green), and z-axis (blue)

Healing caps (ISHA42, Biomet 3i) were screwed into the four implants to manufacture impression trays.

Conventional custom impression tray (CIT)

To fabricate two CITs for open impression taking using the pick-up technique, an initial impression of the IMM with a stock metal tray and alginate (Cavex cream normal set, Cavex, Norden, Germany) was used. After manufacturing a plaster model (super-hard plaster type IV; Fujirock EP, GC, Leuven, Belgium), two CITs were fabricated using cold-cured polymethylmethacrylate (PMMA; C-plast, Candulor Dental, Rielasingen-Worblingen, Germany) with a layer thickness of 2.5 mm. In the implant position regions #14/#24 and #16/#26, a chimney-like opening with a diameter of 2.5 cm was designed, and a tray handle was placed in the anterior region (Fig. 4a).

Fig. 4

a Example of CIT and b CFT with impression posts in regions #14/#24 and #16/#26

Customized foil tray (CFT)

In contrast to CTI, no elaborate manufacturing process is required for CFT (Miratray, Hager und Werken, Duisburg, Germany). The correct size of the CFT was determined (size 3 for the maxilla), and before impression taking, the foil was perforated with a dental probe in the region of the four impression posts (#14/#24 and #16/#26, Fig. 4b). According to the manufacturer's recommendations, the foil tray is designed for single use. Therefore, a new CFT was used for each impression.

Chairside 3-D printed impression trays

Two different chairside workflows for manufacturing of 3-D printed impression trays were analyzed: the SHERA system (3DS, SHERA Werkstoff-Technologie, Lemförde, Germany) and the Primeprint system (3DP, Dentsply Sirona, Bensheim, Germany). To simulate a close clinical study setup, the IMM was digitized using a laboratory scanner (D2000, 3Shape, Copenhagen, Denmark).

SHERA system (3DS)

First, the STL dataset of the IMM was imported into the computer-aided design (CAD) software SHERAeasy-base (version 2.0; SHERA Werkstoff-Technologie, Lemförde, Germany). The following features were selected: impression type, implant impression; impression material, polyether; implant system, Biomet 3i; and height of the healing cap, 2 mm. Further data on the respective implant systems were stored in the digital library of the software. Thus, the implant axis and the corresponding position of the chimney-like tray openings were calculated automatically, as well as the direction of insertion for the tray and the block out of the undercuts (Fig. 5a; chimney diameter 10 mm, thickness of the tray 3 mm).

Fig. 5

a Example of the 3DS with automatically calculated chimney-like tray openings and b the designed tray

Next, the contours of the trays were determined. Finally, buccal bars were designed in addition to a tray handle to facilitate the removal of trays from the IMM, and the CAD dataset was exported (Fig. 5b).

The computer-aided manufacturing (CAM) software Netfabb (version 2022, Autodesk, Munich, Germany) was used for nesting the CAD dataset. As two trays had to be fabricated, the CAD dataset was duplicated. Therefore, two trays were placed on the virtual printing platform in the software and support structures were added. The final dataset was exported to rapid-shape format, and transferred to the digital light-processing SHERAPrint 30 3D printer (SHERA Werkstoff-Technologie). For additive manufacturing, the light-curing pink-colored resin SHERAprint-tray clear (SHERA Werkstoff-Technologie) was used. After completing the printing process, the trays were manually detached from the platform. Followed by a post-processing cleaning in the SHERAprint-wash cleaning and drying unit (SHERA Werkstoff-Technologie), the support structures were cut off and trays were post-polymerized in the SHERA print-cure light-curing unit (SHERA Werkstoff-Technologie).

Primeprint system (3DP)

For the manufacturing of the chairside 3-D printed impression trays with the Primeprint system, the STL dataset of the IMM was imported into the inLab CAD software (version SW 22.1.1, Dentsply Sirona). After positioning the dataset in the coordinate system of the inLab CAD software, a virtual model of the IMM was created. The inLab splint software (version 22.0.3, Dentsply Sirona) was opened using inLab CAD software. In contrast to the SHERA system, all steps of the tray design had to be selected manually, except for the automatically determined direction of insertion of the tray and the block out of the undercuts. Therefore, a tray contour, gingiva former height of 2 mm, implant axis, and chimney-like openings were designed for the impression posts (Fig. 6a). The diameters of the openings were selected based on the implant and impression post used (10 mm in the present study design). The thickness of the impression tray was 3 mm. In the final design process, buccal bars were added next to the tray handle to facilitate the removal of the tray from the IMM (Fig. 6b).

Fig. 6

a Example of the 3DP with manually designed chimney-like tray openings and b the designed tray

For nesting the CAD dataset, the inLab CAM software (version 22.2.0, Dentsply Sirona) was used. As two trays had to be fabricated, the CAD dataset was duplicated. Two trays were placed on the virtual printing platform in the software and support structures were added. According to the manufacturer, the printing process was performed using the material Primeprint Tray (Dentsply Sirona). Finally, the tray was automatically cleaned and post-polymerized in the Primeprint Post Processing Unit (PPU, Dentsply Sirona). After the tray was detached from the platform, the support structures were removed manually.

Figure 7 shows an example of the 3-D printed impression trays.

Fig. 7

a Example of 3DS and b 3DP with inserted impression posts in regions #14/#24 and #16/#26

Implant impression taking

Ten implant impressions were obtained from each of the four study groups. After five impressions, the trays in the CIT, 3DS, and 3DP groups were replaced, whereas in the CFT group, a new customized tray was used for each impression.

All the trays were coated with a thin layer of Polyether Adhesive (3 M, Neuss, Germany). For implants in positions #16 and #26, implant system-specific impression posts with anti-rotation protection ((IIIC42—non-hexed, Biomet 3i) were used, whereas for implants in regions #14 and #24, impression posts without anti-rotation protection (IIIC41, Biomet 3i) were applied. All impression posts were tightened with a torque of 10 Ncm according to the manufacturers’ instructions.

Polyether (Impregum Penta, 3 M) was used as the impression material and automatically mixed with the corresponding Pentamix 3 mixing device (3 M). After a setting time of 6 min, the screws of the impression posts were unscrewed, and the impression tray with the impression posts embedded in the impression material were removed from the IMM.

Fabrication of the plaster models

To ensure recovery of the polyether impression material, all impressions were stored for at least 45 min, and laboratory analogs (H51, H-series, nt-Trading, Karlsruhe, Germany) with a diameter of 4.1 mm were screwed with a torque of 10 Ncm into the impression posts in the implant impression to reproduce the implant position during model fabrication.

Type IV plaster (Fujirock EP, GC, Leuven, Belgium) was used to fabricate the plaster models according to the manufacturer’s instructions. The plaster models were demolded from the impressions after 60 min. Model trimming was omitted because of the potential dimensional changes in the models due to water absorption. Prior to subsequent measurements, the models were stored for 7 days.

Measurement and evaluation of the plaster models

The 40 plaster models were measured using the coordinate measuring machine (CMM) CNC Thome RAPID (Thome Präzision, Messel, Germany) with the corresponding measuring software Metrolog X4 (version 10, Metrologic, Meylan, France). To determine the implant position on the plaster casts, scanbodies (H-series, nt-Trading) with a polyetherketone (PEEK) base and titanium wing surfaces were screwed into laboratory analogs with a torque of 10 Ncm. The scanbodies used were measured individually in the CMM prior to the study to obtain the exact length and determine the implant position as accurately as possible during evaluation.

The first measurement was performed manually and recorded using the measurement software as a measurement and inspection template, based on which the measurement process was repeated five times. First, all five planes of the reference cube (Fig. 1) were probed at four points using a 3-mm-diameter ruby head probe (Renishaw, Pliezhausen, Germany). Subsequently, these planes were circularly measured automatically at 7000 points. Next, the scanbodies were probed with a 1.5-mm-diameter ruby head. The upper surface of each scanbody was probed as a plane with three points and measured in an automatic circle to avoid faulty touches. Subsequently, a cylinder was constructed by probing the scanbodies at 12 points.

After all the elements were measured, a coordinate system with an origin point was created on the reference cube. The points between the implant and abutments (implant-abutment interface points/IASPs) were constructed by shifting the intersection points of the cylinder and planes of the scanbodies using the previously determined length of the scanbodies. After each pass, the collected data were saved and arithmetically averaged for each model. The determined x-, y-, and z-coordinates of the IASP were imported into the inspection software GOM Inspect 2022 (GOM, Braunschweig, Germany) and aligned to the original coordinate system of the IMM. Subsequently, the distances in the x-, y-, and z-directions between the respective determined and reference points were constructed and measured to obtain the deviation from the master model.

Statistical analysis

Statistical analyses were performed using SPSS 26 (IBM, Armonk, NY, USA) with an alpha error of 5%. To investigate whether the impression technique and implant position differed significantly in terms of absolute deviation, a two-factor 4 × 4 ANOVA (analysis of variance (ANOVA) was performed. In addition, a robustness analysis was performed to exclude the possibility of distortions in the statistical analysis owing to outliers. Only a marginal difference was observed between the results of all cases and those without outliers. Because clear variance heterogeneity was observed, the model was calculated using MIXED (estimation method REML, degrees of freedom according to Satterthwaite).

The multiple pairwise comparisons were corrected for alpha error accumulation according to SIDAK.

Data are presented as boxplot diagrams. Trueness (mean) and precision (SD) are reported according to the International Organization for Standardization (ISO) 5725 [42].