Evaluation of model accuracy and implant analog position in stone and additively manufactured models with different resin viscosities Authors: Gülşah Çınar, Necla Demir, Mehmet Esad Güven Published: 2025 ABSTRACT AbstractBackgroundThis in vitro study aimed to assess the trueness and precision of additively manufactured (AM) dental models using resins of different viscosities and conventional stone models, along with the positional accuracy of implant analogs.MethodsThree model groups were created using a reference model with four parallel implants on both posterior sides: stone models (ST), AM models fabricated with low-viscosity resin (AM-A), and high-viscosity resin (AM-K). Each group included eight samples (n = 8). ST models were generated from polyether impressions with Type IV dental stone, while AM models were manufactured 3D printer according to the manufacturers’ guidelines. Scan bodies (SBs) were placed at the implant analog (IA-1 and IA-2) and multi-unit abutment (MA-1 and MA-2) levels. All models were scanned using a high-resolution desktop scanner. Root mean square (RMS) values of the models and SBs, as well as 3D and angular deviations of the implant analogs and inter-implant deviations, were calculated using metrology software (Geomagic Control X). Statistical analysis was performed using the Kolmogorov–Smirnov test for normality, Levene’s test for homogeneity of variances, one-way ANOVA, and the Games–Howell post hoc test (α = 0.05).ResultsSignificant differences were found in RMS trueness among the groups (p < 0.043), with the ranking AM-A > AM-K > ST, whereas the AM-K group showed the lowest precision. Analysis of the RMS values of the SB regions, as well as the 3D and angular deviation measurements of the implant analogs, revealed comparable results between the AM-A and ST groups, while the AM-K group frequently demonstrated significant deviations. Interimplant deviation values were generally similar across the groups. In addition, significant differences between implant and abutment-level SBs were observed in all groups (p < 0.018).ConclusionsIn terms of model trueness, AM models exhibited higher trueness compared to stone models. Deviation analysis of analogs showed the AM-A group yielded outcomes comparable with ST, while AM-K showed the highest deviations, indicating resin viscosity affects the positional accuracy of implant analogs. SBs at the multi-unit abutment level showed greater accuracy than those at the implant level.Keywords: Trueness, Additive Manufacturing , Accuracy , Analog Position, Viscosity Background A definitive implant model is a 3-dimensional (3D) replica that represents the implant position and its relationship with the surrounding tissues with high accuracy, encompassing both trueness and precision [1–3]. Dental models are used for recording diagnostic data, evaluation of implant positions and 3D dental structure, pre-fitting of restorations, such as adjusting occlusal and proximal contacts of restorations, and pre-cementation of implant-supported fixed prostheses to abutments [1, 4, 5]. Conventional stone models have long been utilized as a dependable method for this purpose [1, 5]. Moreover, the 3D position of implants can be digitized through the implementation of intraoral scanners into clinical workflows and enabling the fabrication of digital implant casts using additive and subtractive manufacturing techniques [4, 6]. Additive manufacturing (AM) technology was introduced to resolve the constraints related to the subtractive technique [4, 5, 7]. AM technology provides several benefits, such as enabling the fabrication of definitive models that are easily reproducible. If an implant analog is compromised, it can be replaced. Additionally, the digital storage of STL files eliminates the need to maintain physical model records in private clinical settings [2]. Additively manufactured dental casts have been shown to exhibit higher dimensional accuracy compared to those fabricated by subtractive and conventional methods [8, 9]. In addition, previous studies have investigated several factors affecting the accuracy of printed dental models and the positional accuracy of implant analogs within these models [1, 2, 4, 5, 10, 11]. Among these, the resin used for fabrication is an essential factor affecting the accuracy of dental models [4, 12]. The viscosity of resin is considered a principal determinant in vat polymerization methods including SLA and DLP [13, 14]. Currently available resins exhibit variations in both chemical composition and viscosity [15, 16]. The viscosity of a resin is influential in the manufacturing process, especially during layering because the low fluidity of high viscosity resins, they may not flow sufficiently to the surface to form the next layer [15]. However, high viscosity has been reported to provide superior mechanical properties, making the material more resistant to post-curing deformation [12]. Previous authors compared implant dental models fabricated by additive manufacturing (AM), along with the accuracy of implant replica positions in these models, to those produced using conventional methods [2, 5, 7, 11, 17]. The resin used in AM dental applications is among the many factors that can affect fabrication accuracy and dimensional stability [12], which may influence the position of the implant analog and, consequently, the fit observed during clinical try-in. However, to the best of the authors’ knowledge, only limited investigations have explored the effect of different additively manufactured resins on the accuracy of AM definitive casts compared to the conventional method [4, 6, 16]. Therefore, the aim of this study was to assess AM models manufactured with two dental resins of different viscosities in comparison with conventional stone models, and to evaluate the accuracy of implant analog positions in these models. The null hypotheses were that (i) no difference exists in the accuracy between AM models with different viscosities and conventional dental stone models, (ii) no differences exist in the 3D and angular deviations of the implant analog positions among the different model groups, (iii) no differences exist in inter-implant linear and angular distance deviations, and (iv) no differences exist in the implant-analog deviations at the multi-unit abutment and implant levels within each model type. Methods Reference model A maxillary partially dentate typodont, featuring central, lateral, and canine denture teeth, was used as the master model. Four dental implants (T6 Implant System Ø4.1 mm × 10 mm; NucleOSS, Turkey) were placed at the sites corresponding to the right and left first premolars and first molars using a surgical guide. Multiunit abutments (NucleOSS, İzmir, Turkey) were attached to the implants in the right premolar (MA-1) and right molar (MA-2) and torqued to 25 Ncm, while the left first premolar (IA-1) and left molar (IA-2) were left at the implant level. To simulate clinical conditions and facilitate posterior measurements, the most coronal 2 mm of the left premolar implant and 3 mm part of the left molar and right premolar and molar implants were covered with tissue moulage (Gingifast Elastic, Zhermack, Badia Polesine, Italy) (Fig. 1). Subsequently, the maxillary typodont was duplicated using three different approaches: two distinct additive manufacturing resins (AM-A and AM-K), and a conventional method (ST). The power analysis (G*Power 3.1) indicated that a minimum total of 8 samples was deemed adequate to achieve 80% statistical power (actual power: 82%) at a 95% confidence level (α = 0.05). Therefore, 8 specimens were included per group.. The schematic workflow of the study is shown in Fig. 2. Master model. Used as the reference for fabricating both stone and additively manufactured dental models Experimental workflow Manufacturing of models For manufacturing the conventional stone model group (ST), eight open-tray splinted impressions were made using a polyether impression material (Impregum Penta Soft, 3 M ESPE, Neuss, Germany). The implant-level and abutment-level open tray impression abutments (Impression copings open tray; NucleOSS, Turkey) were positioned and tightened to 10 Ncm using a torque wrench. To standardize the procedure, implant impression copings were splinted with autopolymerizing modeling acrylic (Pattern Resin LS, GC, Tokyo, Japan). An impression with a custom tray and polyether impression material was made under ambient conditions (23 °C). The impression copings were tightened to 10 Ncm onto the implant replicas. Type IV stone (Triastone, GC, Tokyo, Japan) was used for casting the impressions after mixing 100 g of dental stone with 20 mL of water using a vacuum device (Twister Evolution Venturi, RENFERT, Germany) for 60 s. The procedure was repeated until eight specimens were obtained. Two AM subgroups were created based on dental resins with different viscosities: the first group (n = 8) dental models printed with low viscosity (DentaMODEL; ASIGA, Australia) (AM-A) and the second group (n = 8) dental models printed with high viscosity (KeyModel; Keyprint, Germany) (AM-K). The viscosities of the AM-A and AM-K groups were 400 mPa⋅s and 600 to 1200 mPa·s according to the manufacturer’s information. For the AM groups, polyetheretherketone (PEEK) SBs (Mono Scan Body; Nucleoss, Turkey) were tightened to the implant analogs (IA'S) and multi-unit abutment analogs (MA's) on the maxillary typodont with a torque of 10 Ncm. All SBs were unused to eliminate any potential influence on the results. The typodont was digitized using an intraoral scanner (TRIOS 5, 3Shape, Copenhagen, Denmark) by a single operator under stable ambient conditions (23 °C) to simulate clinical conditions. The STL files acquired from the scanner were transferred to CAD software (Exocad,Version 3.3, Exocad GmbH, Darmstadt, Germany) and processed by an experienced dental laboratory technician. Based on the SB positions, appropriate digital analogs and their corresponding analog sockets were designed. Since the study focused on the accuracy of the working models, no relief space was created for the gingival silicone during the design phase. The finalized models were then transferred to the 3D-printer slicer software (Asiga Composer; Asiga, Sydney, Australia), where the respective print configuration files (.ini files) provided by each resin manufacturer were selected. AM models were fabricated using a digital light processing (DLP)-based 3D printer (ASIGA MAX UV; Sydney, Australia). For standardization, the layer thickness was set to 50 μm and the printing orientation was adjusted horizontally, consistent with previous studies [8, 18]. After fabrication, the samples underwent post-processing: they were cleaned in isopropyl alcohol (IPA) under vibration for 10 min. Subsequently, the printed models were light-cured with 2000 flashes applied to both the intaglio and occlusal surfaces (a total of 4000 flashes) using a light-curing unit (Otoflash G171, NK-Optik GmbH, Baierbrunn, Germany). All manufacturing and post-processing steps, including implant analog placement, were conducted following the manufacturers’ recommendations. Scanning procedure Unused digital implant analogs (Digital analogues; Nucleoss, Turkey) were positioned the housing of each analog into printed cast. The AM models were kept in dry and lightproof boxes to be scanned. All models were digitized using a reference scanner (AutoScan-DS-MIX, Shining 3D, Belgium) with high scanning accuracy (≤ 7 μm, ISO 12836). The typodont was scanned and used as a reference STL file (R-STL) to compare the 24 different casts and the scans of the manufactured models were converted into test STL files (T-STLs). The calibration of the scanner was performed before the procedure and after every two scans. All scanning procedures were performed by a single experienced clinician under standardized laboratory conditions. Digital measurements and analyses Industrial metrology software (Geomagic Control X 2024; Geomagic Inc., 3D Systems, NC, USA) was used for all analyses. For each manufactured model, the test STL file (T-STL) was designated as the “measured data,” while the STL file of the master model (R-STL) was designated as the “reference model.” Prior to alignment and measurement procedures, the “region” feature of the software was used on the reference model to identify and separate the virtual SB regions and other parts of the model (Fig. 3). Thus, the SB regions were defined as separate areas before the geometric features were identified. After alignment of both models using the “initial” and “best-fit alignment” tools of the software, the “3D Compare” tool was employed to compute the root mean square (RMS) values for both the individual SB regions and the overall model surface. Identification of the scan body regions and other parts of the model on the master model used for accuracy assessment For SB deviations, all predefined SB regions on the reference data were defined as cylindrical geometries using the “Geometry” tool of the software, and each SB was identified and named within the software. The software then automatically measured the 3D and angular deviations of the corresponding SB regions on the measured data. This allowed the program to automatically calculate the 3D and angular deviations between the corresponding cylindrical geometries in the reference and measured datasets. Three-dimensional deviation was described as the distance between the center point of the cylindrical geometry in the reference model and the corresponding center point in the test model. Three-dimensional deviation was described as the distance between the center point of the cylindrical geometry in the reference model and the corresponding center point in the test model. Angular deviation was defined as the angle between the vertical central axes of the corresponding cylinders in the reference and test models (Fig. 4). Angular deviation was defined as the angle between the vertical central axes of the corresponding cylinders in the reference and test models (Fig. 4). Defining the scanned bodies as cylindrical geometries and determining their vertical central axes for inter-implant and inter-implant angular deviation measurements For inter-implant linear deviations, the linear measurements between the center points of the SB cylinder geometries on the same side were taken on the reference data using the “Dimension > Linear Dimension” tool of the software. For angular deviations, the “Dimension > Angular Dimension” tool was used to measure the top angle between the axes of the SB cylinder geometries on the same side. The software then automatically recorded the deviations on the measured data. (Fig. 5). All measurements were performed by a more than 5 years experienced single investigator (M.E.G). (Fig. 5). All measurements were performed by a more than 5 years experienced single investigator (M.E.G). Illustration of inter-implant linear and angular deviations measurement (IA, implant analog; MA, multi-unit abutment analog) Before the statistical analysis, the absolute values of the 3D deviation and inter-implant linear deviation were calculated for trueness evaluation. For precision evaluation, the standard deviations (SDs) of the trueness values were used. Data analysis The study data were analyzed using statistical software (Jamovi, Version 2.6, 2024; Sydney, Australia). The Kolmogorov-Smirnov test confirmed normal distribution of the data, while Levene’s test showed that variance homogeneity was not consistent across all trueness and precision variables. Therefore, one-way ANOVA was performed, followed by the Games-Howell test for pairwise comparisons. Descriptive statistics were calculated, with the significance level set at α = 0.05. Results Table 1 presents the trueness and precision results of RMS measurements (mean and standard deviation values). The trueness and precision results (mean and standard deviation) of the 3D and angular deviation measurements of the implant analogs are presented in the Table 2. The trueness and precision results of the inter-implant linear and angular deviation measurements (mean and standard deviation) are presented in Table 3. Tables 4 and 5, and 6 show the differences between the groups and the p-values from the Games-Howell test. The results of the 3D and angular deviations of the SBs at the implant level and the abutment level within the same model are presented in Table 7. Table 1Absolute trueness and precision values (mean ± standard deviation) of RMS measurementGroupsTruenessPrecisionMODEL RMSAM-A118.15 ± 37.75a33.45 ± 12.10aAM-K160.30 ± 19.19b15.25 ± 10.13aST205.56 ± 31.39c25.88 ± 14.84aMA-1 RMSAM-A38.04 ± 10.48a8.89 ± 4.42aAM-K94.70 ± 43.45b37.35 ± 17.13bST44.21 ± 8.90a7.29 ± 4.31aMA-2 RMSAM-A45.21 ± 30.01a18.45 ± 22.61abAM-K78.06 ± 45.96a34.47 ± 27.47bST42.89 ± 3.83a2.82 ± 2.37aIA-1 RMSAM-A58.51 ± 13.58a11.48 ± 5.80aAM-K71.39 ± 20.43a14.56 ± 13.24aST70.76 ± 20.99a15.64 ± 12.69aIA-2 RMSAM-A107.85 ± 16.93b14.78 ± 6.11aAM-K166.97 ± 42.90c33.01 ± 24.40aST65.54 ± 25.05a18.07 ± 15.95 aUnits: micrometers (µm). Significance set at p < 0.05Lowercase letters represent significant differences between groupsAbbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Absolute trueness and precision values (mean ± standard deviation) of RMS measurement Units: micrometers (µm). Significance set at p < 0.05 Lowercase letters represent significant differences between groups Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Table 3Absolute trueness and precision values (mean ± standard deviation) of inter-implant linear and angular deviation measurementsInterimplant linear Deviation (µm)Interimplant angular Deviation (°)GroupsTruenessPrecisionTruenessPrecisionMA-1 and 2AM-A45.21 ± 30.01a18.45 ± 22.61ab0.14 ± 0.09a0.10 ± 0.05aAM-K78.06 ± 45.96a34.47 ± 27.47b0.16 ± 0.22a0.18 ± 0.19aST42.89 ± 3.83a2.82 ± 2.37a0.30 ± 0.20a0.26 ± 0.32aIA-1 and 2AM-A58.51 ± 13.58a11.48 ± 5.80a1.02 ± 0.13b0.10 ± 0.07aAM-K71.39 ± 20.43a14.56 ± 13.24a1.19 ± 0.46b0.31 ± 0.32abST70.76 ± 20.99a15.64 ± 12.69a0.32 ± 0.26a0.51 ± 0.32bUnits: micrometers (µm) for inter-implant linear deviations, degrees (°) for inter-implant angular deviations. Statistical significance set at p < 0.05 Lowercase letters represent significant differences between groupsAbbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Absolute trueness and precision values (mean ± standard deviation) of inter-implant linear and angular deviation measurements Units: micrometers (µm) for inter-implant linear deviations, degrees (°) for inter-implant angular deviations. Statistical significance set at p < 0.05 Lowercase letters represent significant differences between groups Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Table 5Games-Howell test p-values compared trueness and precision values of 3D and angular deviations3D DeviationAngular deviationCompared GroupsGroup 1Group 2Trueness (p)Precision (p)Trueness (p)Precision (p)MA-1AM-AAM-K0.1050.6480.2230.447AM-AST0.064 0.021* 0.6410.474AM-KST 0.01* 0.1310.5590.852MA-2AM-AAM-K 0.025* 0.9960.2610.486AM-AST0.2770.9340.460.931AM-KST < 0.001* 0.8630.0560.652IA-1AM-AAM-K0.7260.5220.9520.821AM-AST0.8910.373 0.007* 0.22AM-KST0.9410.996 0.004* 0.040* IA-2AM-AAM-K 0.014* 0.183 0.007* 0.216AM-AST < 0.001* 0.2760.1940.22AM-KST < 0.001* 0.0820.080.822Significance set at p < 0.05 Boldface values and asterisks (*) indicate statistically significant differences (p < 0.05)Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Games-Howell test p-values compared trueness and precision values of 3D and angular deviations Significance set at p < 0.05 Boldface values and asterisks (*) indicate statistically significant differences (p < 0.05) Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Table 2Absolute trueness and precision values (mean ± standard deviation) of the 3D and angular deviation3D Deviation (µm)Angular deviation (°)GroupsTruenessPrecisionTruenessPrecisionMA-1AM-A46.81 ± 20.38ab17.19 ± 8.82b0.30 ± 0.10a0.06 ± 0.07aAM-K80.64 ± 37.11b25.71 ± 24.94ab0.44 ± 0.20a0.13 ± 0.14aST26.51 ± 7.71a5.82 ± 4.57a0.35 ± 0.13a0.10 ± 0.06aMA-2AM-A40.73 ± 17.94a12.75 ± 11.67a0.25 ± 0.10a0.08 ± 0.06aAM-K67.06 ± 17.33b13.23 ± 10.01a0.36 ± 0.16a0.12 ± 0.08aST28.14 ± 12.95a11.16 ± 5.03a0.26 ± 0.11a0.09 ± 0.06aIA-1AM-A65.33 ± 21.43a13.90 ± 15.44a0.45 ± 0.18a0.13 ± 0.11abAM-K76.72 ± 35.74a25.06 ± 23.67a0.48 ± 0.14a0.10 ± 0.09aST71.26 ± 29.48a24.19 ± 14.17a0.87 ± 0.26b0.23 ± 0.09bIA-2AM-A159.43 ± 25.44b21.18 ± 11.59a0.78 ± 0.13a0.12 ± 0.05aAM-K248.54 ± 65.24c49.15 ± 38.67a1.46 ± 0.44b0.31 ± 0.29aST44.46 ± 15.69a13.78 ± 5.39a1.01 ± 0.31ab0.24 ± 0.18aUnits: micrometers (µm) for 3D deviations, degrees (°) for angular deviations. Significance set at p < 0.05 Lowercase letters represent significant differences between groupsAbbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Absolute trueness and precision values (mean ± standard deviation) of the 3D and angular deviation Units: micrometers (µm) for 3D deviations, degrees (°) for angular deviations. Significance set at p < 0.05 Lowercase letters represent significant differences between groups Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Table 4P-values from Games-Howell tests comparing absolute trueness and precision of RMS values between groupsCompared GroupsGroup 1Group 2Trueness (p)Precision (p)MODEL RMSAM-AAM-K 0.043* 0.015* AM-AST < 0.001* 0.52AM-KST 0.012* 0.254MA-1 RMSAM-AAM-K 0.018* 0.005* AM-AST0.4350.748AM-KST 0.031* 0.003* MA-2 RMSAM-AAM-K0.2480.433AM-AST0.9740.195AM-KST0.147 0.033* IA-1 RMSAM-AAM-K0.3320.822AM-AST0.3780.687AM-KST0.9980.985IA-2 RMSAM-AAM-K 0.013* 0.163AM-AST 0.005* 0.851AM-KST < 0.001* 0.348Significance set at p < 0.05 Boldface values and asterisks (*) indicate statistically significant differences (p < 0.05)Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model P-values from Games-Howell tests comparing absolute trueness and precision of RMS values between groups Significance set at p < 0.05 Boldface values and asterisks (*) indicate statistically significant differences (p < 0.05) Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Table 6Games-Howell test p-values compared trueness and precision of inter-implant linear and angular deviationsInter-implant Linear DeviationInter-implant Angular Deviation Compared Groups Group 1 Group 2 Trueness (p)Precision (p)Trueness (p)Precision (p)MA-1 and 2AM-AAM-K0.2480.4330.9660.524AM-AST0.9740.1950.1750.402AM-KST0.147 0.033* 0.4380.815IA-1 and 2AM-AAM-K0.3320.8220.590.219AM-AST0.3780.687 < 0.001* 0.02* AM-KST0.9980.985 0.002* 0.455Statistical significance set at p < 0.05 Boldface values and asterisks (*) indicate statistically significant differences (p < 0.05)Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Games-Howell test p-values compared trueness and precision of inter-implant linear and angular deviations Statistical significance set at p < 0.05 Boldface values and asterisks (*) indicate statistically significant differences (p < 0.05) Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Table 7Comparative analysis of 3D and angular deviations at implant and abutment levels within modelsRepeated Measures3D Deviation (p)Angular Deviation (p)AM-A (implant vs. multi-unit) < 0.001 0.001 AM-K (implant vs. multi-unit) < 0.001 < 0.001 ST (implant vs. multi-unit) 0.018 < 0.001 Significance set at p < 0.05Boldface values indicate statistically significant differences (p < 0.05)Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model Comparative analysis of 3D and angular deviations at implant and abutment levels within models Significance set at p < 0.05 Boldface values indicate statistically significant differences (p < 0.05) Abbreviations: IA implant analog, MA multi-unit abutment analog, ST stone model, AM-A low-viscosity resin model, AM-K high-viscosity resin model When the RMS measurements of model trueness were considered, statistical analysis revealed significant differences between the groups (p ≤ 0.043). The AM-A group (118.15 ± 37.75 μm) exhibited the highest trueness, followed by the AM-K group (160.3 ± 19.19 μm) and the ST group (205.56 ± 31.39 μm). The precision of the AM-K group (15.25 ± 10.13 μm) was significantly lower compared to the results of the AM-A group (33.45 ± 12.10 μm) (p = 0.015) (Table 1 and Table 4; Fig. 6). When the RMS measurements of SB regions, the AM-K group showed significantly lower trueness than the other groups in the MA-1 (p ≤ 0.031). In the IA-2, all groups exhibited statistically significant differences (p ≤ 0.013). The ST (65.54 ± 25.05 μm) group exhibited the highest trueness, followed by the AM-A (107.85 ± 16.93 μm) group and the AM-K group (166.97 ± 42.90 μm). Regarding the precision results, the AM-K group showed significantly lower precision than the remaining groups in the MA-1 (p ≤ 0.005). In the MA-2, the ST group had higher precision than the AM-K group (p = 0.033) (Table 1 and Table 4; Fig. 6). RMS values across groups. Box plots showing the RMS values for (a) trueness and (b) precision across all model groups and implant analog positions When assessing the 3D deviations, the ST group exhibited significantly lower deviations than both AM groups in most cases. The ST group (26.51 ± 7.71 μm) showed significantly higher trueness compared to the AM-K group (80.64 ± 37.11 μm) (p = 0.01) in the MA-1, while in the MA-2, the AM-K group showed lower trueness than the other groups (p ≤ 0.025). Statistical analysis revealed a significant difference across all groups in IA-2, with the ST group exhibiting the highest trueness (p ≤ 0.014). When taking the precision into account, the AM groups were mostly similar to the ST group. Regarding precision in 3D deviations, the ST group (5.82 ± 4.57 μm) had higher precision than the AM-A group (17.19 ± 8.82 μm) in MA-1 (p = 0.021) (Table 2 and Table 5; Fig. 7). When the angular deviations were considered, the ST group had the lowest trueness among all groups in IA-1 (p < 0.007). In IA-2, the AM-A (0.78 ± 0.13°) group had higher trueness than AM-K group (1.46 ± 0.44°) (p = 0.007). In terms of angular precision, the ST group (0.23 ± 0.09°) had lower precision than the AM-K group (0.10 ± 0.09°) in IA-1 (p = 0.040) (Table 2 and Table 5; Fig. 8). Three-dimensional deviations across groups. Box-plot graphs of measured 3D deviations showing (a) trueness and (b) precision across implant analog positions within each model group Angular deviations across groups. Box-plot graphs of measured angular deviations showing (a) trueness and (b) precision across implant analog positions within each model group The trueness and precision of the inter-implant linear and angular deviations were assessed between MA-1 and MA-2, and between IA-1 and IA-1 (Table 3). When the inter-implant linear deviations were considered, no statistically significant differences were detected between groups (p > 0.05). In terms of precision, the ST group (2.82 ± 2.37 μm) showed higher values than the AM-K group (34.47 ± 27.47 μm) in the measurements between MA-1 and MA-2 (p = 0.033) (Table 3 and Table 6; Fig. 9). For the trueness of inter-implant angular deviations, the ST group (0.32 ± 0.26°) showed the highest trueness among all groups between IA-1 and IA-2 (p ≤ 0.002). When considering precision, the ST group (0.51 ± 0.32°) showed lower precision than the AM-A group (0.10 ± 0.07°) in the same region (p = 0.020) (Table 3 and Table 6; Fig. 10). Interimplant linear deviations. Box plots showing inter-implant linear deviations for (a) trueness and (b) precision at implant analog positions across all model groups Interimplant angular deviations. Box plots showing inter-implant angular deviations for (a) trueness and (b) precision at implant analog positions across all model groups Additionally, a statistically significant difference in the mean 3D and angular deviations between implant analogs at both the implant level and multi-unit abutment level was found across all model groups (Table 7, p ≤ 0.05). Discussion This research primarily aimed to evaluate the accuracy of dental stone models and AM models manufactured using low-viscosity and high-viscosity resins based on five key metrics; the RMS values, the 3D and angular deviations of the implant analogs position, and the linear and angular deviations between implants. The analysis of the accuracy of the tested dental models revealed significant differences among the groups. Furthermore, statistically significant differences were observed in the 3D and angular deviations, and in the inter-implant deviation measurements. Therefore, the null hypothesis was rejected. For RMS measurements, significant differences in trueness were observed among all groups. The difference between the AM and ST groups might be explained by the observation that detailed areas, such as the tooth region, exhibited higher deviation in the ST group within this study. Similar trends were reported by Donmez et al., who found that, regardless of the time interval, stone models exhibited the highest deviations in the anterior tooth region [6]. Earlier studies have indicated that the acceptable threshold for prosthetic procedures in dental models is between 150 and 200 μm [7]. In our study, only the ST group exceeded this value, showing a deviation of 205.56 μm. However, among all groups, only the AM-A group remained below the 150 μm threshold proposed by Papaspyridakos et al. [19]. This may be due to significant differences between models manufactured with low- and high-viscosity resins, which may result from differences in chemical composition and from the influence of resin viscosity on model accuracy. In vat polymerization processes such as DLP, higher-viscosity materials exhibit reduced flowability, which may lead to inconsistencies in layer deposition and decreased surface accuracy [15]. Among the AM groups, the low-viscosity resin group (AM-A) had higher accuracy than the high-viscosity resin group (AM-K). These findings correspond to those reported in previous studies [16, 20, 21]. According to the RMS trueness and precision measurements of the SB regions, no significant difference was found between AM-A and ST except for IA-2. However, the AM-K group had the highest mean values. The cause might be the differing polymerization shrinkage and viscosity characteristics of the resins. Due to the small and intricate structure of the analog holes, these factors may have influenced the fit of the analogs within the holes and, consequently, affected the deviations in the SB regions [17]. In terms of 3D deviations, except for IA-1, the ST group had the lowest 3D deviation among all groups and its significantly higher trueness than the AM-K group. Previous studies have reported that the main differences between ST and AM implant models are that the implant analog is placed in AM models after manufacturing, and that digital scanning is performed for the production of AM models [4, 22]. However, the fact that this significant difference between the ST and AM-K groups was not observed between the ST and AM-A groups may be explained by the more limited effect of the resin, in addition to other sources of error. Furthermore, the fact that both RMS values and 3D deviation of IA-2 in the AM groups were higher compared to both the ST group and the other region within the AM groups suggests that the AM models were influenced by the digital scanning process. Earlier research has reported that factors such as the visibility of the scan body [23, 24], the scanned region [25, 26], and the presence of reference teeth affect scanning accuracy [27]. The results of this study are consistent with those of previous studies. In this study, the implant at the position of the left first molar was placed 3 mm below the crest level to simulate clinical conditions, making it the deepest implant. Moreover, it was the last region to be scanned, as the intraoral scanner (IOS) was used in accordance with the manufacturer’s instructions, and no adjacent reference teeth were present. The present authors compared the angular deviation of the implant analogs between the ST and AM groups. Consistent with previous studies [2, 5], our findings on trueness suggest that there was no significant difference between the groups, except for IA-1. When the AM groups were evaluated, the AM-A group showed lower trueness than the AM-K group, suggesting a possible influence of resin viscosity on angular deviation; however, the differences were not statistically significant, except for IA-2. No prior studies have been found that compare angular deviation in AM models manufactured using different resins. Inaccuracies in manufacturing can result in discrepancies between the implant and prosthesis, which may cause micro-gaps and mechanical stress [5]. Rutkūnas et al. reported that AM models showed higher trueness in inter-implant linear deviations than the ST group [5]. In contrast, another study reported that, for the same parameter, trueness was higher in the ST group compared to the AM group [17]. In the present study, the AM-A and ST groups exhibited comparable results, but the AM-K group showed lower trueness, yet the observed differences were not statistically significant. When similar studies were considered, the studies by Rutkūnas and Tan Shaun reported no significant differences in inter-implant angular deviation between the ST and AM groups. The findings correspond to the inter-implant angular deviations recorded in MA-1 and MA-2 in this study. Nevertheless, at positions IA-1 and 2, a statistically significant difference was identified between the ST and AM groups, which may be attributed to the higher deviation of the IA-2 in the AM groups. No consensus currently exists regarding the clinically acceptable threshold for implant-prosthesis misfit; however, the recommended range is 100–150 μm for linear discrepancy [28, 29] and 0.4° for angular discrepancy [5]. However, in many studies, the maximum mean angular deviation has been reported to exceed this threshold, rising to 1.25° [30–32]. In our study, the acceptable threshold for 3D deviation was exceeded only in IA-2 of the AM groups. However, the mean deviation reported in the AM-A group was 159.43 μm, which slightly exceeds the clinically accepted threshold. According to the clinically acceptable angular deviation threshold, only IA-2 in the AM-K group exceeded the maximum allowable limit of 1.25°, while all other values remained below this threshold limit. Previous studies have evaluated the biocompatibility, mechanical properties, and dimensional stability of various resins [4, 6, 12, 13, 16]. Additionally, previous studies reported that digital analogs are influenced by resin properties such as hardness, flexibility, and deformation [5, 33, 34]. However, as far as the authors are aware, no previous research has exclusively examined the effect of resin viscosity on the accuracy of models and the positional deviations of implant analogs. According to the analyses in our study, although not all results reached statistical significance, the mean values indicated that the AM-A group exhibited lower deviations compared to the AM-K group. Only for IA-2, the angular deviation precision was found to be lower in the AM-K group compared to the AM-A group. This outcome, based on previous studies, is thought to be related to differences in resin flow characteristics, which may directly affect the layer-by-layer manufacturing process. The mechanical and chemical properties of the resin determine the structural integrity of the analog holes, where even minor deformations in these small structures can lead to disproportionately large positional deviations [17, 33–35]. The secondary purpose of this research was to examine the differences in SB deviations between implants and multi-unit abutments. While deformations may occur at the SB–implant interface under tightening forces in implants with a deep conical connection [36], the mechanical stop provided in the vertical direction by the abutment shoulder in abutment-level impressions [37] may explain the differences in deviations between SBs at the multi-unit and implant levels within the same model. Nam et al. reported that attaching the (SB) to the abutment rather than directly to the implant would improve impression accuracy [38]. This finding is consistent with the results of a previous study evaluating the accuracy of digital impressions at the abutment and implant levels, in which all models showed higher accuracy at the abutment level [37]. Consistent with prior research, this study also demonstrated that SBs placed at the implant level exhibited significantly higher mean deviation compared to those placed at the abutment level. Present study offers significant insights; however, some limitations exist. Only two resins (high- and low-viscosity), one 3D printer, and one post-curing unit were used. Additionally, the resins were evaluated primarily based on their viscosity, without consideration of their chemical compositions. Minor deviations may have occurred during the placement or screwing of the analogs into their sockets, which could affect the accuracy independently of the analog position. The use of an intraoral scanner (IOS) in an in vitro setting represents another limitation of this study, as such devices may not fully replicate clinical conditions; using a high-precision coordinate measuring machine instead of a scanner could potentially affect the results [5]. As implant components from only one manufacturer were used, the behavior of analogs from different brands within the model resin remains unclear. In terms of clinical applicability, further advanced experimental studies are required in which a greater variety of resins with different viscosities are assessed, and comparisons are made using additively manufactured and stone dental models, including full-arch edentulous jaw models with implants placed in varying numbers, positions, angulations, and depths. Conclusions The outcomes of this in vitro investigation allow for the following conclusions: The positional accuracy of analogs in additively manufactured models were, in most cases, comparable to that observed in conventional stone casts. The positional accuracy of analogs in additively manufactured models were, in most cases, comparable to that observed in conventional stone casts. The resin viscosity employed in the manufacturing process was identified as a significant factor affecting the positional accuracy of the implant analogs. The resin viscosity employed in the manufacturing process was identified as a significant factor affecting the positional accuracy of the implant analogs. The tested AM casts appear to be a feasible substitute for stone models when assessing implant-supported prosthetic restorations.. The tested AM casts appear to be a feasible substitute for stone models when assessing implant-supported prosthetic restorations.. Acknowledgements Not applicable. Availability and requirements Not applicable. Clinical trial number Not applicable. Abbreviations Additively manufactured Three-Dimensional Implant analog Multi-unit abutment analog Root mean square Scan body Stone model Standard Tessellation Language (3D file format) Authors’ contributions G.Ç. carried out the methodology, data curation, investigation, and drafted the original manuscript. N.D. contributed to conceptualization, supervision, visualization, and project administration. M.E.G. contributed to software, validation, and project administration. G.Ç. and M.E.G. performed the formal analysis, while N.D. and M.E.G. were involved in manuscript review and editing. All authors reviewed the manuscript, approved the final version, and agree to be accountable for all aspects of the work. Funding There was no funding used in our study. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate Ethics approval was obtained from Selcuk University Medical Faculty Ethical Committee (decision no: 2023/51). This study did not involve human participants, human data, or animals. The data analyzed in this study were derived from a previously completed dental speciality of prosthodontics thesis project by Dr. Çınar. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.