Abstract
PURPOSE
This study aimed to evaluate the influence of implant placement depth on the accuracy of five full-arch digital impression techniques: 2 photogrammetry systems (Imetric [PGI] and Oxo Core [PGO]) and 3 intraoral scanning (IOS) routes (standard IOS, IOS with crown-shaped [IOS+C], and IOS with laterally-extending [IOS+LE] scanning aids).
MATERIALS AND METHODS
Three master models including 5 implants with 2-, 4-, and 6-mm subgingival implant depths were fabricated. Implant positions were recorded using each of the 5 techniques (n = 15). The resultant scans were compared to reference data generated by a laboratory scanner. Deviations were 3-dimensionally assessed using root mean square values to quantify both trueness and precision. Angular deviations (AD) were calculated to provide a detailed assessment of spatial inconsistencies. Data were statistically analyzed (α = 0.05).
RESULTS
Implant depth and impression technique, as well as their interaction, significantly affected trueness, precision, and angular deviation values (P < .001). PGI demonstrated the highest accuracy, with the lowest mean AD values across all implant positions (0.24 – 0.32°) and the highest trueness (26.57 µm) and precision (7.72 µm). IOS exhibited the largest distortions, with AD values up to 1.77° and trueness values of 74.84 µm. Precision of both photogrammetry systems were not influenced by implant depth.
CONCLUSION
Implant depth significantly impacts the accuracy metrics of full-arch digital impressions. The Imetric system outperformed the other groups. Since laterally-extending scanning aids effectively improved accuracy, it may be recommended as a clinically viable alternative to standard IOS.
Keywords: Precision, Prefabricated auxiliary device, Photogrammetry, Scan-body visibility, Trueness
INTRODUCTION
Implant-supported restorations represent a reliable treatment method for the rehabilitation of completely edentulous arches, and 4 to 6 implants are usually used for full arch support.1 Achieving passive fit is important for the long-term viability of screw-retained restorations. Passive fit refers to the absence of static or dynamic stresses at the bone-implant interface and the restoration-implant interface when the prosthesis is completely secured, thus minimizing the risk of mechanical and biological complications associated with malalignment.2
The foundation for achieving passive fit is to execute a highly accurate transfer.3 Transfer accuracy is determined by trueness and precision, and it is thoroughly detailed in ISO 5725-1:1994.4 Trueness reflects how close a measurement is to a known reference value, while precision refers to the consistency or repeatability of measurements under the same conditions. Both components are critical for the reliability of transfer techniques, especially in complex clinical situations involving multiple implants placed at different angles or depths.5
Multiple techniques exist for acquiring the spatial positions of dental implants and transferring the captured mapping for subsequent prosthetic procedures.2 Of these, intraoral scanning technique (IOS) has gained increasing popularity due to its improved patient comfort, faster workflow, and ease of data storage and transfer.6 Intraoral scanners are non-contact devices that function by capturing multiple 2D images due to their limited sensor range. Integrated software then utilizes an alignment algorithm to computationally combine these 2D images into a single 3D representation. After scanning the scan body, dental CAD software uses alignment algorithms like the best-fit method to align the scan with its CAD reference or its manufactured 3D rendering, and position the implant analog in the virtual model.7,8,9 However, this method includes all data points—relevant or not—which can affect alignment accuracy.9 Unlike a direct impression coping that can be customized to accommodate deeper implant placement,10 a scan body lacks such customization capabilities to reduce alignment errors.11 These digital inaccuracies can accumulate and affect the fit of definitive prosthesis.12 Despite all this, IOS offers greater patient comfort, better time efficiency, convenient storage, and improved communication between dentists and patients.13
The accuracy of the image-stitching process depends on the software’s ability to spatially align 2D image captures, which is facilitated by the presence of stable anatomical landmarks.14 As the distance between scan bodies increases across the edentulous area, and as inherent inaccuracies stemming from the absence of anatomical landmarks accumulate during the stitching process, the resulting 3D image can suffer from cumulative distortion.15 Therefore, studies often suggest using IOSs as a substitute for traditional methods primarily for single-tooth and short-span fixed partial dentures.16,17 When natural anatomical guides are missing, the software may struggle to stitch the images correctly, or parts of the scan might be incorrectly identified as overlapping data.14 To address this, artificial landmarks should be introduced to create a continuous reference between scan bodies, thus ensuring accurate tracking of the scanning path.18 Several techniques are advised for incorporating artificial landmarks: splinting scan bodies,14,19 using scan bodies with extensions,20 and employing prefabricated auxiliary devices like scan aids.21,22,23,24,25 It is obvious that all these designs are difficult to apply in every single case, and this reveals the need for more adaptable designs suitable for clinical practice.
Photogrammetry (PG) is another route of transfer besides IOS.26 It allows photomapping which is a process of creating a map by enhancing a photo-mosaic, where individual images are corrected for tilt and standardized to a common scale. For this purpose, PG employs multiple cameras and captures 2 or more photographic images taken from different positions.26,27 Previous studies have indicated that PG can offer superior accuracy in transferring implant positions when compared to CT and standard IOS routes.26,28
Implants can be inserted at varying depths due to patient-related factors such as soft tissue thickness or bone availability,29 which may alter the scanner’s line of sight or influence marker visibility in photogrammetry.30 Therefore, this study aimed to evaluate the influence of implant placement depth (2-, 4-, and 6-mm subgingival) on the trueness of digital impressions obtained with 5 routes (Imetric PG system, Oxo Core PG system, standard IOS, IOS assisted with crown-shaped scanning aid [IOS+C], and IOS assisted with laterally-extending scanning aid [IOS+LE]). The null hypotheses were that there would be no significant differences in terms of trueness and precision among full-arch transfer routes and that there would be no significant influence of varying implant depth on the trueness and precision of full-arch transfer routes.
MATERIALS AND METHODS
A schematic representation of the study design is described in Fig. 1. Master models were fabricated by pouring self-curing acrylic resin (Meliodent Rapid Repair; Kulzer GnmH, Hanau, Germany) into a maxillary edentulous silicone mold (B-3 NMG Silicone Index; Frasaco GmbH, Baden-Württemberg, Germany). To simulate the gingival tissues, a 2-mm thick layer of silicone material (Prestige VDG Mask Rigid; Vannini Dental Industry, Florence, Italy) was applied to the master models. Multi-unit analogs (Nobel Active Multi-unit Analog; Ø4.8 mm, Nobel Biocare, Zurich, Switzerland) were positioned in 5 predetermined locations: the canine regions (#13 and #23), the first molar on the right quadrant (#16), and the first premolar and second molar on the left quadrant (#25 and #27). The sockets for these analogs were prepared using a rotary instrument with a parallelometer (Marathon 103; Saeyang Microtech, Daegu, Korea) to ensure controlled variations in parallelism and depth among the analogs.
Fig. 1. Schematic representation of study design.
Three unique master models were generated that differed by the vertical positioning (depth) of implants #25 and #27. In Model 1, all implant analogs were placed 2 mm subgingivally. In Model 2, analogs #25 and #27 were positioned at a depth of 4 mm, while the remaining analogs (#13, #23, and #16) remained at a depth of 2 mm. In Model 3, the depth of analogs #25 and #27 was further increased to 6 mm, with the others again maintained at 2 mm (Fig. 1). Each master model underwent laboratory scanning (inEOS X5, V22.3.0; Dentsply Sirona, Bensheim, Germany) facilitated by polyetheretherketone (PEEK) scan bodies (Elos Accurate; Scan Body IO 2C-A Elos Medtech, Timmersdala, Sweden), and the resultant reference data were archived in the standard tessellation language (STL) format.
The minimum requisite sample size was statistically determined utilizing a software program (G*Power, v.3.1.9.7; Heinrich Heine University Düsseldorf, Germany). An effect size of 0.25 was selected based on Cohen’s31 convention for a medium effect in ANOVA analyses, which was deemed appropriate for evaluating differences across multiple impression techniques.32,33 A statistical power of 85% was chosen to slightly enhance the sensitivity of the analysis while maintaining feasibility within the constraints of the in vitro setting. With a significance threshold of 0.05, the calculated minimum sample size was 15.
The digital impression methodologies evaluated within this study (n = 15) are comprehensively presented in Fig. 1. All digital scanning procedures were executed by a single, 5-year-experienced operator (T.M.). For the IOS group (CEREC Primescan, Version 5.2.7; Dentsply Sirona, Bensheim, Germany), PEEK scan bodies were securely affixed to the multi-unit analogs utilizing screws tightened by hand to a calibrated torque (Fig. 2). For the IOS+C and IOS+LE groups, the bespoke scanning aids were virtually designed using an advanced 3D modeling software program (SolidWorks; Dassault Systèmes,Vélizy-Villacoublay, France). The scanning aid used in the IOS+C group had a ring-shaped base with 8 symmetrically arranged protrusions designed to mimic the geometry of a crown (Fig. 2). The scanning aid used in the IOS+LE group employed a modular scan aid system consisting of a central ring integrated with adaptable, perforated extensions and interchangeable triangular rods in 6.5-, 9-, and 11.5-mm length (Fig. 2). The perforations allowed for anatomical customization to accommodate the arch’s curvature. This design facilitated enhanced scan accessibility and precision across a range of clinical configurations. All scan aids were fabricated using a fused deposition modeling (FDM) 3D printer (Prusa i3 MK3S; Prusa Research AS, Prague, Czech Republic) with polylactic acid plus (PLA+) filament. Across all IOS groups, we followed the manufacturer’s stipulated protocols for the scanning procedure. The scanning sequence started with the occlusal aspect of the #27 implant site, progressed along the dental arch toward the #16 implant, and then covered the buccal and palatal surfaces of the scan bodies. The resulting scan data were exported as STL files.
Fig. 2. Intraoral scanning groups: (A) IOS assisted with no scanning aid, (B) IOS assisted with crown-shaped scanning aid, and (C) IOS assisted with laterally-extending scanning aid.
For the PGI group, optical markers (iCamBodies; iCam CAD-CAM GmbH, Düsseldorf, Germany), characterized by a rectangular form and a matte black surface decorated with a white polka dot pattern, were manually fastened onto multi-unit analogs (Fig. 3A). Before each scanning session, the photogrammetry device was calibrated to ensure optimal performance. The device was positioned approximately 20 cm from the scan bodies, and the scanning head was methodically moved from the left to the right side to accurately register the three-dimensional spatial coordinates of the implants. Once implant positions were recorded, the scan bodies were replaced with reference markers (iCamRefs; Imetric4D Imaging Sàrl, Courgenay, Switzerland) to facilitate soft tissue capture using an intraoral scanner (TRIOS 3; 3Shape, Kopenhag, Denmark). The datasets—implant positions and soft-tissue morphology—were subsequently merged into a unified STL file using the best-fit alignment algorithm embedded within the proprietary PG software program (iScan 3D Dental v9; Imetric4D Imaging Sàrl, Courgenay, Switzerland).
Fig. 3. Photogrammetry systems: (A) Imetric group, (B) Oxo Core group.
For the PGO group, optical markers (Oxo Markers; Oxo Technology, Madrid, Spain), characterized by a flag-like configuration with white surfaces bearing gray geometric patterns, were affixed to the multi-unit analogs (Fig. 3B). Following device calibration, these markers were scanned to acquire implant position data. Upon completion, the markers were removed and replaced with healing caps (OXO Fit Healing Caps; Oxo Technology, Madrid, Spain), after which an intraoral scanner (AS200E; AlliedStar Medica, Shanghai, China) was used to capture soft-tissue contours. As with the PG system, the implant and soft-tissue datasets were integrated into a single STL file using the system’s proprietary software program (Oxo Fit, V2.0.0; Oxo Technology, Madrid, Spain).
In this study, three-dimensional surface deviations were assessed using root mean square (RMS) values, following previously established protocols for evaluating the accuracy of digital implant impressions.34 Additionally, angular deviations (AD) were calculated to provide a detailed assessment of spatial inconsistencies. For comprehensive 3D analysis, reference and test datasets were imported into a metrology software platform (Geomagic Control X, V2014.3.0; 3D Systems, Rock Hill, SC, USA), all performed by a single calibrated operator (T.M.). For AD analysis, the virtual scan bodies in the reference model were converted into hollow cylinders using the software’s “Create Feature” tool. This process began by manually selecting the superior planar surface of each scan body, from which a horizontal reference plane was generated. A second plane was then created 8 mm inferior to the first, corresponding to the height of the scan bodies, to define the boundaries of the cylindrical feature. Identical cylinders were automatically generated in the test scans to maintain methodological consistency (Fig. 4). Cartesian coordinates (x, y, z) of the central axes of these cylinders were recorded for all scan bodies (Fig. 5). The scan body at implant site #13 served as the reference axis. AD values were calculated by measuring the Euclidean distances between the central axis of implant #13 and those of the other implants (#23, #25, #27, and #16) (Fig. 1). ΔAD values were derived by comparing these angular measurements between the reference and test datasets, thereby quantifying the deviation in implant angulation across different scanning modalities.
Fig. 4. Conversion of scan bodies into hollow virtual cylinders.
Fig. 5. Record of Cartesian coordinates of reference and test virtual cylinders.
RMS values were calculated to quantify deviations between the reference and test data sets. These calculations were based on previously derived Euclidean linear distances (LD) obtained from the Cartesian coordinates of hollow cylinders representing the central axes of the scan bodies as defined during the AD analysis. RMS deviations were calculated using the standard RMS formula. To assess precision, the scan in each test group showing the lowest RMS deviation from the reference was designated as the internal reference scan. RMS values were then recalculated for all other scans in the same group relative to this internal reference. The resulting RMS values were reported as the precision metric for each group, with lower values indicating greater consistency and reproducibility of the scanning technique.
All data were subjected to statistical analysis using a software program (IBM SPSS Statistics v25.0; IBM Corp., Armonk, NY, USA). Normality of data distribution was assessed using the Shapiro-Wilk test, and homogeneity of variances was evaluated with Levene’s test. A two-way analysis of variance (ANOVA) was conducted to examine the main and interaction effects of the variables—digital impression technique and implant depth—on ΔAD, trueness, and precision values. Where significant differences were observed (P < .05), Tukey’s Honestly Significant Difference (HSD) test was applied for post-hoc comparisons. A significance level of P < .05 was adopted for all statistical tests.
RESULTS
The Shapiro-Wilk test indicated normal distribution (P > .05), and Levene’s test confirmed homogeneity of variances (P > .05). For angular distortion (ΔAD), technique and model complexity significantly influenced the measurements across all assessed implant positions. In accordance with the results of 2-way ANOVAs, ΔAD values were significantly affected by all variables and their interaction term (P < .001), except impression technique × implant depth on 13–25 (P = .596), 13–27 (P = .621), and 13–16 data (P = .800) (Table 1). The mean ΔAD values and standard deviations with Tukey post hoc comparisons are presented in Table 2. PGI showed the lowest overall distortions with means of 0.30 ± 0.25 (13–23), 0.24 ± 0.15 (13–25), 0.26 ± 0.24 (13–27), and 0.32 ± 0.25 (13–16). IOS had the highest overall distortions with means of 1.11 ± 0.63 (13–23), 0.95 ± 0.59 (13–25), 0.76 ± 0.28 (13–27), and 1.35 ± 0.51 (13–16). An increase in AD values was observed from Model 1 to Model 3.
Table 1. Two-way ANOVA results.
| Factor | Type III Sum of Squares | df | Mean Square | F | P | |
|---|---|---|---|---|---|---|
| Assessment of AD Data of 13-23 | ||||||
| Technique a | 13.712 | 4 | 3.428 | 91.582 | P < .001 | |
| Model b | 6.690 | 2 | 3.345 | 89.366 | P < .001 | |
| Technique × Model a | 4.872 | 8 | 0.609 | 16.269 | P < .001 | |
| Assessment of AD Data of 13-25 | ||||||
| Technique a | 9.229 | 4 | 2.307 | 11.393 | P < .001 | |
| Model b | 3.670 | 2 | 1.835 | 9.061 | P < .001 | |
| Technique × Model a | 1.311 | 8 | 0.164 | 0.809 | P = .596 | |
| Assessment of AD Data of 13-27 | ||||||
| Technique a | 4.172 | 4 | 1.043 | 14.613 | P < .001 | |
| Model b | 2.695 | 2 | 1.347 | 18.879 | P < .001 | |
| Technique × Model a | 0.445 | 8 | 0.056 | 0.780 | P = .621 | |
| Assessment of AD Data of 13-16 | ||||||
| Technique a | 20.876 | 4 | 5.219 | 40.263 | P < .001 | |
| Model b | 1.861 | 2 | 0.930 | 7.177 | P = .001 | |
| Technique × Model a | 0.592 | 8 | 0.074 | 0.571 | P = .800 | |
| Assessment of Trueness Data | ||||||
| Technique a | 62161.370 | 4 | 15540.343 | 1170.359 | P < .001 | |
| Model b | 7559.548 | 2 | 3779.774 | 284.659 | P < .001 | |
| Technique × Model a | 3731.994 | 8 | 466.499 | 35.133 | P < .001 | |
| Assessment of Precision Data | ||||||
| Technique a | 1207.689 | 4 | 301.922 | 82.792 | P < .001 | |
| Model b | 200.586 | 2 | 100.293 | 27.502 | P < .001 | |
| Technique × Model a | 64.826 | 8 | 8.103 | 2.222 | P = .027 | |
a Test within-subject contrasts, b Test between-subjects effects.
Table 2. Descriptive statistics for angular distortion values (°) with pairwise comparisons.
| AD Data | Technique | Model 1 | Model 2 | Model 3 | Total |
|---|---|---|---|---|---|
| 13-23 | PGI | 0.13 ± 0.01A,a | 0.26 ± 0.15A,a | 0.50 ± 0.33A,b | 0.30 ± 0.25A |
| PGO | 0.14 ± 0.02A,a | 0.43 ± 0.11A,b | 0.47 ± 0.23A,b | 0.35 ± 0.21A | |
| IOS | 0.40 ± 0.16B,a | 1.18 ± 0.41B,b | 1.77 ± 0.17B,c | 1.11 ± 0.63C | |
| IOS+C | 0.33 ± 0.03AB,a | 0.51 ± 0.12A,ab | 0.65 ± 0.21A,b | 0.49 ± 0.19B | |
| IOS+LE | 0.25 ± 0.09AB,a | 0.41 ± 0.13A,a | 0.44 ± 0.25A,a | 0.37 ± 0.18AB | |
| Total | 0.25 ± 0.12a | 0.55 ± 0.39b | 0.77 ± 0.56c | 0.52 ± 0.45 | |
| 13-25 | PGI | 0.16 ± 0.02 | 0.26 ± 0.24 | 0.32 ± 0.05 | 0.24 ± 0.15A |
| PGO | 0.17 ± 0.07 | 0.45 ± 0.25 | 0.41 ± 0.28 | 0.34 ± 0.25A | |
| IOS | 0.61 ± 0.06 | 0.99 ± 0.89 | 1.27 ± 0.28 | 0.95 ± 0.59B | |
| IOS+C | 0.36 ± 0.05 | 0.41 ± 0.23 | 0.66 ± 0.25 | 0.48 ± 0.23A | |
| IOS+LE | 0.28 ± 0.07 | 0.77 ± 0.34 | 0.79 ± 0.17 | 0.61 ± 0.49A | |
| Total | 0.31 ± 0.17a | 0.57 ± 0.46b | 0.69 ± 0.40b | 0.53 ± 0.43A | |
| 13-27 | PGI | 0.16 ± 0.02 | 0.18 ± 0.02 | 0.45 ± 0.35 | 0.26 ± 0.24A |
| PGO | 0.20 ± 0.06 | 0.46 ± 0.45 | 0.61 ± 0.38 | 0.42 ± 0.37A | |
| IOS | 0.65 ± 0.11 | 0.67 ± 0.33 | 0.97 ± 0.25 | 0.76 ± 0.28C | |
| IOS+C | 0.42 ± 0.05 | 0.61 ± 0.29 | 0.63 ± 0.14 | 0.55 ± 0.21B | |
| IOS+LE | 0.35 ± 0.07 | 0.58 ± 0.37 | 0.77 ± 0.41 | 0.57 ± 0.35B | |
| Total | 0.36 ± 0.19a | 0.50 ± 0.36b | 0.69 ± 0.36c | 0.51 ± 0.34 | |
| 13-16 | PGI | 0.20 ± 0.05 | 0.28 ± 0.07 | 0.49 ± 0.39 | 0.32 ± 0.25A |
| PGO | 0.29 ± 0.07 | 0.49 ± 0.22 | 0.38 ± 0.21 | 0.38 ± 0.19A | |
| IOS | 1.08 ± 0.20 | 1.41 ± 0.54 | 1.56 ± 0.62 | 1.35 ± 0.51C | |
| IOS+C | 0.56 ± 0.20 | 0.79 ± 0.56 | 0.81 ± 0.14 | 0.72 ± 0.36B | |
| IOS+LE | 0.39 ± 0.11 | 0.50 ± 0.46 | 0.61 ± 0.11 | 0.50 ± 0.44AB | |
| Total | 0.50 ± 0.34a | 0.69 ± 0.62b | 0.77 ± 0.54b | 0.66 ± 0.52 |
Superscript uppercase letters indicate statistically significant differences within the same column, while superscript lowercase letters denote significant differences within the same row.
An angular deviation threshold of 0.4° was established based on prior studies.20,35 Deviations surpassing these limits were deemed clinically unacceptable. IOS alone exceeded the 0.4° threshold in every region (1.11 ± 0.63 in 13–23 and 1.35 ± 0.51 in 13–16), rendering it clinically unacceptable without the use of scan aids.
In terms of trueness, a RMS threshold of 100 µm was determined based on prior studies.20,36,37 In accordance with the results of 2-way ANOVAs, trueness values were significantly affected by all variables and their interaction term (P < .05). The trueness values obtained from different impression techniques across the 3 models are presented in Table 3. For Model 1, the lowest mean deviation was recorded with PGI (19.07 ± 2.15 µm), followed by PGO (30.26 ± 5.89 µm), IOS+LE (33.24 ± 3.54 µm), IOS+C (36.01 ± 5.44 µm), and IOS (67.09 ± 8.21 µm). A similar pattern was noted in Model 2, where PGI (24.73 ± 2.02 µm) yielded significantly lower deviations compared to other techniques (P < .05). In Model 3, PGI demonstrated the lowest deviation (35.91 ± 1.59 µm), though the differences among techniques were narrower compared to the other models. When considering the overall means, PGI achieved the highest trueness (26.57 ± 7.29 µm), followed by IOS+C (38.26 ± 4.63 µm), IOS+LE (35.04 ± 3.05 µm), PGO (47.82 ± 13.26 µm), and IOS (74.84 ± 7.75 µm). Across all techniques, trueness decreased as the model complexity increased from Model 1 to Model 3, as indicated by statistically significant differences among models (P < .05). All RMS values remained well below the 100 µm clinical threshold, indicating that the deviations were within acceptable limits across all tested techniques and depths.
Table 3. Descriptive statistics for trueness values with pairwise comparisons (µm).
| Technique | Model 1 | Model 2 | Model 3 | Total |
|---|---|---|---|---|
| PGI | 19.07 ± 2.15A,a | 24.73 ± 2.02A,b | 35.91 ± 1.59A,c | 26.57 ± 7.29A |
| PGO | 30.26 ± 5.89B,a | 55.90 ± 2.10C,b | 57.31 ± 2.26C,b | 47.82 ± 13.26D |
| IOS | 67.09 ± 8.21C,a | 77.73 ± 2.05D,b | 79.70 ± 4.41D,b | 74.84 ± 7.75E |
| IOS+C | 36.01 ± 5.44B,a | 36.56 ± 1.34B,a | 43.15 ± 2.99B,b | 38.26 ± 4.63C |
| IOS+LE | 33.24 ± 3.54B,a | 35.63 ± 1.54B,b | 37.27 ± 1.33A,b | 35.04 ± 3.05B |
| Total | 36.74 ± 14.96a | 46.11 ± 18.94b | 50.67 ± 16.70c | 44.51 ± 18.45 |
Superscript uppercase letters indicate statistically significant differences within the same column, while superscript lowercase letters denote significant differences within the same row.
Precision values were, in accordance with the results of 2-way ANOVAs, significantly affected by all variables and their interaction term (P < .05). The mean precision values and standard deviations for each technique across the 3 models are presented in Table 4. For Model 1, PGI demonstrated the highest precision (7.11 ± 1.58 µm), showing significantly lower variability than PGO (10.80 ± 2.91 µm), IOS+C (10.44 ± 1.78 µm), IOS+LE (11.71 ± 3.22 µm), and IOS (13.71 ± 2.32 µm). This trend persisted in Model 2, with PGI (7.70 ± 0.91 µm) maintaining superior precision relative to the other techniques. Similarly, in Model 3, PGI (8.03 ± 1.12 µm) exhibited the lowest deviation, while IOS (16.40 ± 1.32 µm) presented the highest deviation, indicating significantly lower precision. When considering the overall means, PGI achieved the highest precision across all models (7.72 ± 1.27 µm), followed by PGO (11.30 ± 2.11 µm), IOS+C (12.11 ± 2.07 µm), IOS+LE (12.14 ± 2.91 µm), and IOS (14.94 ± 2.15 µm). Across all techniques, a significant decline in precision was observed as model complexity increased from Model 1 to Model 3 (P < .05).
Table 4. Descriptive statistics for precision values with pairwise comparisons (µm).
| Technique | Model 1 | Model 2 | Model 3 | Total |
|---|---|---|---|---|
| PGI | 7.11 ± 1.58A,a | 7.70 ± 0.91A,a | 8.03 ± 1.12A,a | 7.72 ± 1.27A |
| PGO | 10.80 ± 2.91B,a | 11.30 ± 1.13B,a | 11.80 ± 2.66B,a | 11.30 ± 2.11B |
| IOS | 13.71 ± 2.32C,a | 14.72 ± 1.79C,a | 16.40 ± 1.32D,b | 14.94 ± 2.15C |
| IOS+C | 10.44 ± 1.78B,a | 11.90 ± 1.32B,a | 13.90 ± 1.34C,b | 12.11 ± 2.07B |
| IOS+LE | 11.71 ± 3.22B,a | 11.60 ± 1.32B,a | 14.30 ± 1.32C,b | 12.14 ± 2.91B |
| Total | 10.70 ± 3.35a | 11.44 ± 2.60b | 12.89 ± 3.27c | 11.64 ± 3.16 |
Superscript uppercase letters indicate statistically significant differences within the same column, while superscript lowercase letters denote significant differences within the same row.
DISCUSSION
This in vitro study evaluated the effect of implant placement depth (2 mm, 4 mm, and 6 mm subgingival corresponding to 8 mm, 6 mm, and 4 mm scan body visibility) on the accuracy of full-arch digital implant transfer methods. Based on statistical analyses, both null hypotheses were rejected as significant differences in trueness and precision were detected among transfer methods and implant depth was shown to have a significant effect on these accuracy parameters.
Although digital intraoral scanning has been widely adopted in implant dentistry, the effect of subgingival implant placement and hence scan body visibility on the accuracy of IOSs has been investigated in a limited number of studies, particularly in edentulous full-arch models. Giménez et al.38 evaluated the effect of implant angulation and depth by using an active wavefront sampling system in a completely edentulous model and suggested that implant depth had a minimal effect on accuracy, although this may be due to the specific scanner technology used in the study and the relatively short scan intervals.38 Giménez-González et al.39 extended this research by showing that implants placed subgingivally in a full-arch maxillary model exhibited greater deviation than those placed at the equigingival level. This increased deviation was attributed to decreased viewing distance and restricted lines of sight, increasing the risk of stitching errors at long scanning distances.39 Arcuri et al.40 reinforced this perspective with a study in which implant depth (0 – 6 mm) significantly affected IOS accuracy in completely edentulous arches. More deviations were detected in deeper placements, and the authors emphasized that reduced supragingival visibility compromised optical detection and alignment, and consequently reduced the accuracy of the virtual model.40 Similarly, Choi et al.41 also demonstrated that reduced scan body exposure negatively influenced the accuracy of image matching.41 Recently, Gómez-Polo et al.42 tested the impact of both clinical scan body height and implant angulation in a full-arch model systematically and compared the perceived heights of 10 mm, 6 mm, and 3 mm. They proved that the lowest visibility group (3 mm) exhibited the highest linear and angular deviations significantly, especially when the implants were angulated. The authors attributed improved scanning accuracy in the groups with greater scan body visibility to enhanced geometric referencing during intraoral scanning. Specifically, greater supragingival exposure facilitates the stitching process by providing more consistent and accessible reference points across the arch, thereby improving alignment and reducing cumulative error propagation.42
Implementation of scan aids has been proven to improve the accuracy of IOSs,21,23,25 but their accuracy in the case of subgingival implant placement with reduced visibility of the scan body has not been researched extensively. Additionally, the impact of implant depth on PG accuracy has not been examined in previous studies. Therefore, in the present study, the efficiency of scan aid designs in reducing the negative effects of implant depth on accuracy was examined, and PG were also assessed to determine whether their performance was less influenced by implant depth. To minimize the impact of variables such as inter-implant distance or implant position within the arch, comparisons were made using implants placed on the same side. In this way, the effect of implant depth was evaluated under controlled conditions.
In the present study, all evaluated parameters—AD, trueness, and precision—were significantly influenced by both the transfer technique and implant depth. Among the evaluated systems, the PGI showed the lowest AD values across all models, remaining below the clinical threshold of 0.4°.20,35 In contrast, IOS without scan aids revealed the highest AD, exceeding 1.7° in the deepest placement condition (Model 3), which is well above the clinically acceptable limit. IOS without scan aid showed the highest trueness deviations, especially at deeper placements, reaching up to 79.70 µm RMS in Model 3. Although these values remained below the 100 µm clinical threshold, they were notably higher compared to IOS+C and IOS+LE groups, where scan aids helped to reduce the deviations. The IOS+LE subgroup showed slightly more favorable values than IOS+C, particularly in the deeper models, suggesting that lateral extensions may provide better geometric reference during stitching. Among the evaluated methods, PGI consistently provided the lowest trueness deviations across all implant depths, while PGO demonstrated increased deviation at greater depths—surpassing IOS+C in the deepest condition. These findings indicate that trueness performance may vary considerably between PG systems, depending on scanning protocol and implant configuration.
Until now, only 2 recent studies43,44 have compared the Imetric system with the Oxo Core system in full-arch implant models. In both studies, the Imetric system (iCam4D) provided superior trueness outcomes compared to the OxoFit system used by Oxo Core for different implant configurations and angulations. Although the specific reasons for this difference were not detailed in those studies, it may be attributable to the differences in scanning workflows between these systems. According to user experience (T.M.) and manufacturer-reported protocols, the Imetric system is believed to capture scan bodies sequentially, while Oxo Core employs simultaneous acquisition. This difference may lead to greater spatial error accumulation in the presence of subgingival placement, which may help explain the greater loss of trueness observed in the PGO group in this study. Precision results showed that PGI provided the most consistent performance across all depths, with no significant within-group differences and the lowest SD overall. Although implant depth did not affect the precision of PGO; its overall repeatability was significantly lower than PGI suggesting that PGO may be more susceptible to inherent system-related variability. These findings are in agreement with the recent studies43,44 where the Imetric system demonstrated the best linear precision among the evaluated photogrammetry methods. IOS without scan aid demonstrated reduced precision at greater implant depths. In all IOS-based groups, precision values in Model 3—where scan body visibility was reduced to 4 mm—were significantly higher than those in Models 1 and 2, indicating that decreased visibility detrimentally affected repeatability even when scan aids (IOS+C and IOS+LE) were used. This suggests that while scan aids help improve precision, they may not fully compensate for the loss of geometric reference when minimal supragingival height is available.
This study has several limitations. One important limitation is that it was conducted under in vitro conditions without using a phantom head model, which reduces clinical relevance. Although a phantom head could have provided more realistic conditions—such as limited intraoral access, presence of saliva, and soft tissue movement—it was not included in order to focus specifically on the effects of implant depth and impression technique under consistent and controlled conditions. In addition, using a phantom head might have made it more difficult to apply the same scanning conditions across all systems, particularly for PG methods that require clear access to the scan bodies. Only the maxillary arch was evaluated, and implants were placed in a parallel configuration; therefore, the findings may not be directly applicable to mandibular or angulated implant scenarios. While 3 different depth conditions were tested, other contributing factors like varying implant numbers, inter-implant distances, and tissue morphology were kept constant to isolate the effect of depth. A laboratory desktop scanner was utilized as a reference scanner for alignment with clinical digital workflows; however, high-resolution industrial scanners might yield more precise baseline readings. Additionally, the evaluation was conducted on metrology software rather than prosthesis fit, which may limit clinical interpretability. Therefore, further studies are needed.
CONCLUSION
This study highlights that both the digital transfer method and the visibility of the scan body—particularly when influenced by subgingival placement—are pivotal factors affecting the 3D positional accuracy of implant impressions. Among the evaluated techniques, photogrammetry systems demonstrated the highest levels of trueness and precision, underscoring their reliability for accurate implant positioning. Intraoral scanning augmented with scanning aids significantly outperformed standard intraoral scanning, offering improved accuracy metrics. Optimizing scan body visibility by using a longer marker or scan body and selecting advanced digital workflows are essential for enhancing the overall quality of digital implant impressions.
References
- 1.Kosago P, Ungurawasaporn C, Kukiattrakoon B. Comparison of the accuracy between conventional and various digital implant impressions for an implant-supported mandibular complete arch-fixed prosthesis: an in vitro study. J Prosthodont. 2023;32:616–624. doi: 10.1111/jopr.13604. [DOI] [PubMed] [Google Scholar]
- 2.Albanchez-González MI, Brinkmann JC, Peláez-Rico J, López-Suárez C, Rodríguez-Alonso V, Suárez-García MJ. Accuracy of digital dental implants impression taking with intraoral scanners compared with conventional impression techniques: a systematic review of in vitro studies. Int J Environ Res Public Health. 2022;19:2026. doi: 10.3390/ijerph19042026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Abduo J, Bennani V, Waddell N, Lyons K, Swain M. Assessing the fit of implant fixed prostheses: a critical review. Int J Oral Maxillofac Implants. 2010;25:506–515. [PubMed] [Google Scholar]
- 4.ISO 5725-1. Accuracy (trueness and precision) of measurement methods and results. Part 1: general principles and definitions. Geneva; Switzerland: International Standards Organization (ISO); 2023. [Google Scholar]
- 5.Imburgia M, Logozzo S, Hauschild U, Veronesi G, Mangano C, Mangano FG. Accuracy of four intraoral scanners in oral implantology: a comparative in vitro study. BMC Oral Health. 2017;17:92. doi: 10.1186/s12903-017-0383-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Eggmann F, Blatz MB. Recent advances in intraoral scanners. J Dent Res. 2024;103:1349–1357. doi: 10.1177/00220345241271937. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Pozzi A, Arcuri L, Laureti A, Carosi P, Gallucci G, Londono J. Image-guided photogrammetry accuracy: In vitro evaluation of an implant-supported complete arch digital scanning technology. J Prosthet Dent. :S0022-3913(25)00365-8. doi: 10.1016/j.prosdent.2025.03.047. [DOI] [PubMed] [Google Scholar]
- 8.Gómez-Polo M, Ballesteros J, Perales-Padilla P, Perales-Pulido P, Gómez-Polo C, Ortega R. Guided implant scanning: a procedure for improving the accuracy of implant-supported complete-arch fixed dental prostheses. J Prosthet Dent. 2020;124:135–139. doi: 10.1016/j.prosdent.2019.09.022. [DOI] [PubMed] [Google Scholar]
- 9.Sequeira V, Harper MT, Lilly CL, Bryington MS. Accuracy of digital impressions at varying implant depths: an in vitro study. J Prosthodont. 2023;32:54–61. doi: 10.1111/jopr.13496. [DOI] [PubMed] [Google Scholar]
- 10.Linkevicius T, Svediene O, Vindasiute E, Puisys A, Linkeviciene L. The influence of implant placement depth and impression material on the stability of an open tray impression coping. J Prosthet Dent. 2012;108:238–243. doi: 10.1016/S0022-3913(12)60169-3. [DOI] [PubMed] [Google Scholar]
- 11.Gómez-Polo M, Donmez MB, Çakmak G, Yilmaz B, Revilla-León M. Influence of implant scan body design (height, diameter, geometry, material, and retention system) on intraoral scanning accuracy: a systematic review. J Prosthodont. 2023;32(S2):165–180. doi: 10.1111/jopr.13774. [DOI] [PubMed] [Google Scholar]
- 12.Gracis S, Appiani A, Noè G. Digital workflow in implant prosthodontics: The critical aspects for reliable accuracy. J Esthet Restor Dent. 2023;35:250–261. doi: 10.1111/jerd.13004. [DOI] [PubMed] [Google Scholar]
- 13.Lee SJ, Gallucci GO. Digital vs. conventional implant impressions: efficiency outcomes. Clin Oral Implants Res. 2013;24:111–115. doi: 10.1111/j.1600-0501.2012.02430.x. [DOI] [PubMed] [Google Scholar]
- 14.Knechtle N, Wiedemeier D, Mehl A, Ender A. Accuracy of digital complete-arch, multi-implant scans made in the edentulous jaw with gingival movement simulation: An in vitro study. J Prosthet Dent. 2022;128:468–478. doi: 10.1016/j.prosdent.2020.12.037. [DOI] [PubMed] [Google Scholar]
- 15.Wulfman C, Naveau A, Rignon-Bret C. Digital scanning for complete-arch implant-supported restorations: a systematic review. J Prosthet Dent. 2020;124:161–167. doi: 10.1016/j.prosdent.2019.06.014. [DOI] [PubMed] [Google Scholar]
- 16.Ferrini F, Sannino G, Chiola C, Capparé P, Gastaldi G, Gherlone EF. Influence of Intra-Oral Scanner (I.O.S.) on the marginal accuracy of CAD/CAM single crowns. Int J Environ Res Public Health. 2019;16:544. doi: 10.3390/ijerph16040544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nagata K, Fuchigami K, Okuhama Y, Wakamori K, Tsuruoka H, Nakashizu T, Hoshi N, Atsumi M, Kimoto K, Kawana H. Comparison of digital and silicone impressions for single-tooth implants and two- and three-unit implants for a free-end edentulous saddle. BMC Oral Health. 2021;21:464. doi: 10.1186/s12903-021-01836-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kim JE, Amelya A, Shin Y, Shim JS. Accuracy of intraoral digital impressions using an artificial landmark. J Prosthet Dent. 2017;117:755–761. doi: 10.1016/j.prosdent.2016.09.016. [DOI] [PubMed] [Google Scholar]
- 19.Nagy Z, Simon B, Mennito A, Evans Z, Renne W, Vág J. Comparing the trueness of seven intraoral scanners and a physical impression on dentate human maxilla by a novel method. BMC Oral Health. 2020;20:97. doi: 10.1186/s12903-020-01090-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Huang R, Liu Y, Huang B, Zhang C, Chen Z, Li Z. Improved scanning accuracy with newly designed scan bodies: an in vitro study comparing digital versus conventional impression techniques for complete-arch implant rehabilitation. Clin Oral Implants Res. 2020;31:625–633. doi: 10.1111/clr.13598. [DOI] [PubMed] [Google Scholar]
- 21.Pan Y, Tsoi JKH, Lam WY, Zhao K, Pow EH. Improving intraoral implant scanning with a novel auxiliary device: an in-vitro study. Clin Oral Implants Res. 2021;32:1466–1473. doi: 10.1111/clr.13847. [DOI] [PubMed] [Google Scholar]
- 22.Kernen F, Schlager S, Seidel Alvarez V, Mehrhof J, Vach K, Kohal R, Nelson K, Flügge T. Accuracy of intraoral scans: an in vivo study of different scanning devices. J Prosthet Dent. 2022;128:1303–1309. doi: 10.1016/j.prosdent.2021.03.007. [DOI] [PubMed] [Google Scholar]
- 23.Wu HK, Wang J, Chen G, Huang X, Deng F, Li Y. Effect of novel prefabricated auxiliary devices attaching to scan bodies on the accuracy of intraoral scanning of complete-arch with multiple implants: an in-vitro study. J Dent. 2023;138:104702. doi: 10.1016/j.jdent.2023.104702. [DOI] [PubMed] [Google Scholar]
- 24.Canullo L, Pesce P, Caponio VCA, Iacono R, Luciani FS, Raffone C, Menini M. Effect of auxiliary geometric devices on the accuracy of intraoral scans in full-arch implant-supported rehabilitations: An in vitro study. J Dent. 2024;145:104979. doi: 10.1016/j.jdent.2024.104979. [DOI] [PubMed] [Google Scholar]
- 25.Bader Eddin MB, Önöral Ö. Influence of splinting scan bodies or incorporating three-dimensionally printed scan aids on the trueness of complete arch digital scans. J Prosthet Dent. 2024;132:828.e1–828.e12. doi: 10.1016/j.prosdent.2024.07.001. [DOI] [PubMed] [Google Scholar]
- 26.Tohme H, Lawand G, Chmielewska M, Makhzoume J. Comparison between stereophotogrammetric, digital, and conventional impression techniques in implant-supported fixed complete arch prostheses: an in vitro study. J Prosthet Dent. 2023;129:354–362. doi: 10.1016/j.prosdent.2021.05.006. [DOI] [PubMed] [Google Scholar]
- 27.Hussein MO. Photogrammetry technology in implant dentistry: a systematic review. J Prosthet Dent. 2023;130:318–326. doi: 10.1016/j.prosdent.2021.09.015. [DOI] [PubMed] [Google Scholar]
- 28.Joensahakij N, Serichetaphongse P, Chengprapakorn W. The accuracy of conventional versus digital (intraoral scanner or photogrammetry) impression techniques in full-arch implant-supported prostheses: a systematic review. Evid Based Dent. 2024;25:216–217. doi: 10.1038/s41432-024-01045-z. [DOI] [PubMed] [Google Scholar]
- 29.Awad EM, ElSheikh MM, El-Segai AAEM. Effect of soft tissue thickness on accuracy of conventional and digital implant impression techniques. BMC Oral Health. 2024;24:1318. doi: 10.1186/s12903-024-05037-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ma B, Yue X, Sun Y, Peng L, Geng W. Accuracy of photogrammetry, intraoral scanning, and conventional impression techniques for complete-arch implant rehabilitation: an in vitro comparative study. BMC Oral Health. 2021;21:636. doi: 10.1186/s12903-021-02005-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. New York: Routledge; 1988. 567 [Google Scholar]
- 32.Önöral Ö, Kurtulmus-Yilmaz S, Toksoy D, Ozan O. Effect of angulation on the 3D trueness of conventional and digital implant impressions for multi-unit restorations. J Adv Prosthodont. 2023;15:290–301. doi: 10.4047/jap.2023.15.6.290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ortensi L, La Rosa GRM, Ciletta S, Grande F, Pedullà E. Evaluation of the accuracy of digital impressions with different scanning strategies: an in vitro study. J Dent. 2024;151:105433. doi: 10.1016/j.jdent.2024.105433. [DOI] [PubMed] [Google Scholar]
- 34.Tallarico M, Qaddomi M, De Rosa E, Cacciò C, Jung YJ, Meloni SM, Ceruso FM, Lumbau AI, Pisano M. Accuracy and precision of digital impression with reverse scan body prototypes and all-on-4 protocol: an in vitro research. Prosthesis. 2025;7:36 [Google Scholar]
- 35.Cheng J, Zhang H, Liu H, Li J, Wang HL, Tao X. Accuracy of edentulous full-arch implant impression: an in vitro comparison between conventional impression, intraoral scan with and without splinting, and photogrammetry. Clin Oral Implants Res. 2024;35:560–572. doi: 10.1111/clr.14252. [DOI] [PubMed] [Google Scholar]
- 36.Andriessen FS, Rijkens DR, van der Meer WJ, Wismeijer DW. Applicability and accuracy of an intraoral scanner for scanning multiple implants in edentulous mandibles: a pilot study. J Prosthet Dent. 2014;111:186–194. doi: 10.1016/j.prosdent.2013.07.010. [DOI] [PubMed] [Google Scholar]
- 37.Kernen-Gintaute A, Akulauskas M, Kernen F, Zitzmann NU, Spies BC, Burkhardt F. Accuracy of digital and conventional implant impressions in edentulous jaws: a clinical comparative study. J Dent. 2025;153:105559. doi: 10.1016/j.jdent.2025.105559. [DOI] [PubMed] [Google Scholar]
- 38.Giménez B, Özcan M, Martínez-Rus F, Pradíes G. Accuracy of a digital impression system based on active triangulation technology with blue light for implants: effect of clinically relevant parameters. Implant Dent. 2015;24:498–504. doi: 10.1097/ID.0000000000000283. [DOI] [PubMed] [Google Scholar]
- 39.Gimenez-Gonzalez B, Hassan B, Özcan M, Pradíes G. An in vitro study of factors influencing the performance of digital intraoral impressions operating on active wavefront sampling technology with multiple implants in the edentulous maxilla. J Prosthodont. 2017;26:650–655. doi: 10.1111/jopr.12457. [DOI] [PubMed] [Google Scholar]
- 40.Arcuri L, Pozzi A, Lio F, Rompen E, Zechner W, Nardi A. Influence of implant scanbody material, position and operator on the accuracy of digital impression for complete-arch: A randomized in vitro trial. J Prosthodont Res. 2020;64:128–136. doi: 10.1016/j.jpor.2019.06.001. [DOI] [PubMed] [Google Scholar]
- 41.Choi YD, Lee KE, Mai HN, Lee DH. Effects of scan body exposure and operator on the accuracy of image matching of implant impressions with scan bodies. J Prosthet Dent. 2020;124:379.e1–379.e6. doi: 10.1016/j.prosdent.2020.04.004. [DOI] [PubMed] [Google Scholar]
- 42.Gómez-Polo M, Sallorenzo A, Ortega R, Gómez-Polo C, Barmak AB, Att W, Revilla-León M. Influence of implant angulation and clinical implant scan body height on the accuracy of complete arch intraoral digital scans. J Prosthet Dent. 2024;131:119–127. doi: 10.1016/j.prosdent.2021.11.018. [DOI] [PubMed] [Google Scholar]
- 43.Revilla-León M, Gómez-Polo M, Drone M, Barmak AB, Kois JC, Alonso Pérez-Barquero J. Accuracy of complete arch implant scans recorded by using intraoral and extraoral photogrammetry systems. Prosthet Dent. 2025:S0022-3913(25)00084-8. doi: 10.1016/j.prosdent.2025.01.041. [DOI] [PubMed] [Google Scholar]
- 44.Revilla-León M, Barmak AB, Drone M, Kois JC, Alonso Pérez-Barquero J. Capturing complete arch implant positions in two partial photogrammetry scans: Does it impact the accuracy of complete arch implant scans? J Prosthet Dent. 2025:S0022-3913(25)00118-0. doi: 10.1016/j.prosdent.2025.02.024. [DOI] [PubMed] [Google Scholar]