Effect of a modified scanning technique on the accuracy of complete-arch implant scanning: an in vitro study

Abstract

Background

This in vitro study aimed to investigate the accuracy of a modified scanning technique that incorporated a simple scan aid (SSA) and a novel scanning pattern (segmented scanning pattern, SSP) for complete-arch implant scan.

Methods

An edentulous maxillary stone cast with 4 parallel analogs was used as the master model. 10 scans were performed with an intraoral scanner (Trios3) using different scanning methods: a conventional scanning pattern (CSP) without SSA (Group I), CSP with SSA (Group II), SSP with SSA (Group III) and conventional splinted open-tray impression (Group IV). The reference scan was obtained by digitizing the master model using a dental laboratory scanner. The related files were imported into inspection software for trueness and precision assessment. One-way ANOVA, Kruskal-Wallis, SNK-q and Bonferroni test were used to evaluate the values. The level of significance was set at α=0.05.

Results

For the overall accuracy assessment (trueness and precision), group I exhibited the largest linear and angular deviation (p<0.001) followed by group II (p<0.001). For local accuracy assessment, all intraoral scan bodies (ISB A, B, C, and D) exhibited smaller linear and angular deviations in group III (p<0.05) and group IV (p<0.05) than that in group I. No differences were found between group III and IV (p>0.05), either in the overall accuracy or local accuracy assessment.

Conclusions

SSA combined with SSP could significantly improve the accuracy of complete-arch digital implant impressions.

Background

Intraoral scanners (IOS) have been widely used in various dental clinical scenarios owing to its simplicity, comfort, efficiency, space-saving benefits, and ease of communication between doctors, patients, and technicians [1–5]. In implant restorations, passivity of fit of the implant superstructure is crucial for long-term successful rehabilitation [6]. Several studies have indicated that the accuracy of an IOS is comparable to, or better than, that of conventional implant impressions with up to three units [7–9]. However, challenges remain when obtaining impressions for more units or edentulous cases [10–12]. An IOS cannot capture the entire object at once like a desktop scanner; instead, it relies on an algorithm involving continuously superimposed small sections of images over a large scanning area [13]. As the scan span is increased, errors in image stitching can accumulate [10–13], especially at the end of the scanning path [14]. Additionally, the gingival mucosa around the implant scan body (ISB) lacks identifiable landmarks, which are crucial for scanner recognition and image stitching during complete-arch implant scan [15, 16]. This can easily lead to the IOS confuse different ISBs and treating them as the same, resulting in erroneous 3D images [17, 18].

To address these issues, various auxiliary devices were used in complete-arch implant scanning to facilitate easier recognition by IOSs which led to some improvement in accuracy [19–22]. Additionally, the scanning pattern is considered an important factor affecting accuracy [23, 24]. Wu et al. [25] demonstrated that the manufacturer recommended scanning pattern leads to the highest scanning accuracy. Although intraoral scanning accuracy in edentulous conditions is improved using auxiliary devices and specific scanning patterns [17, 19–21, 25], these methods are still not widely adopted in clinical practice owing to the complexity of fabricating the devices and failure to address the issue of accumulating errors with increased scanning spans. Additionally, the ideal scanning pattern for complete-arch implant impressions remains unclear [26]. Therefore, the aim of this in vitro study was to evaluate the accuracy of a modified scanning technique using a simple scan aid (SSA) fabricated chairside using polyvinyl siloxane impression material (putty), along with a segmented scanning pattern (SSP) for obtaining complete-arch implant impressions. The null hypothesis was that the modified scanning technique will not affect the accuracy of complete-arch implant scan.

Methods

Master model and reference scan

Ethics approval was not required for this in-vitro study. An edentulous maxillary plaster model with a soft tissue replica was perfused by Type IV gypsum, as the master model. The master model included four parallel implant analogs (RC Implant Analog, Institute Straumann AG, Basel, Switzerland) in the bilateral cusp and first molar positions. The middle two analogs were 24.17 mm apart, and the two adjacent analogs in the anterior and posterior regions were 22.48 mm apart. Four ISBs (CARES RC Mono Scanbody, Institute Straumann AG, Basel, Switzerland), denoted as A–D from the upper right posterior area to the upper left posterior area, were sequentially attached to implant analogs with a manual torque wrench to 15 N·cm (Fig. 1). After calibration, the master model was digitized using a laboratory scanner (CERAMILL MAP 600, Amann Girrbach, Koblach, Austria) with 5 m precision and exported in STL format as reference data.

Fig. 1.

The mster model with scan bodies

Preparation of the SSA

Putty impression material (Silagum Putty Soft; DMG Medical Devices, Hamburg, Germany) was prepared, following the manufacturer’s operating recommendations, and placed on the soft tissue area around ISBs, following the profile of the dental arch. Circular patterns (size: inner diameter 3.5 mm, outer diameter 5.5 mm) were impressed into the surface of the putty. After 5 min’ polymerization, the putty was taken off the dental arch and properly trimmed to fully expose the characteristic structure of the ISB. The surface oil film was then wiped off with a dry cotton ball to complete the personalized manufacture of the SSA that with a thickness of approximately 4 mm (Fig. 2).

Fig. 2.

SSA fabricated chairside using polyvinyl siloxane impression material (putty). A Occlusal surface of the SSA. B Tissue surface of the SSA

Scanning procedures

Calibration of the IOS (Trios3, 3Shape A/S, Copenhagen, Denmark) was completed before the start of the scanning process, following the manufacturer’s recommendations. In addition, the torque (15 N·cm) of ISBs was re-verified through another wrench calibrated by the manufacturer before each scan. The specific scanning procedure for each test group was as follows.

Group I: Using the CSP recommended by the manufacturer (from ISB A along the arch to ISB D following the occlusal-palatal-buccal (OPB) sequence) to scan the master model without the SSA (Fig. 3a).

Fig. 3.

Group division. A Group I: using the CSP without SSA. B Group II: using the CSP with SSA. C Group Ⅲ: using the SSP with SSA. D Group Ⅳ: conventional open-tray splinted impressions

Group II: Using the same scanning pattern as in group I to scan the master model with the SSA (Fig. 3b).

Group III: Using an SSP (starting from the area between ISBs B and C, to complete the scanning of sections B–A and C–D, following the OPB sequence) to complete the segmented scanning of the master model with the SSA. The scans (B–A and C–D) were conducted separately in two independent scanning scenarios (Fig. 3c). Then, registering the scans of B–A and C–D via the n-point alignment and best fit algorithm, based on geometric features of the SSA in reverse engineering software (Geomagic studio 2014, 3D systems, South Carolina, USA) (Fig. 4). Next, merging the registered scans to obtain complete scan data of the master model.

Fig. 4.

The registration of segmented scans. A B–A segmented scan. B C–D segmented san. C The alignment of segmented scans

Group IV: Polyether impression material was used to make the splinted open-tray impression (implant-level) at room temperature (Fig. 3d). The plaster model, containing the implant analogs and artificial gingiva, was then infused with IV gypsum. A laboratory scanner was used to digitize the plaster model after mounting the ISBs.

Calculation of the sample size based on pre-experimentation showed that 9 scans were required in each group for a two-sided difference in scanning accuracy, with α=0.05 and power of 80% (β=0.2), and considering previous related studies [24, 25], the sample size for each group was 10 in the current study. The procedures for each group were repeated 10 times by an experienced operator under the same environment conditions in the same room. As a result, 10 STL files for the master model were obtained in each group. A 5-minute interval was required for scanner cooling between each scan, and the SSA was ensured with no displacement during the scanning of groups II and III.

Data processing and evaluation

The standard data of the ISB, provided by the manufacturer, were aligned to all scans (the reference scan and all test scans) using the n-point alignment and best fit algorithms to replace the original ISB information in Geomagic Studio 2014 software (Fig. 5). The center axis and vertices were created for all the ISBs in an interactive medical image processing software (Mimics 11.0; Materialise, Leuven, Belgium). The test scans were then aligned to the reference data using ISBs A–D as the alignment feature.

Fig. 5.

Registering the standard data of ISB to scan data. (The orange area is the scanning image of ISBs and the blue area is the image of standard ISBs)

According to ISO 5725-1, accuracy consists of trueness and precision. Trueness refers to the closeness of agreement between the measured value and the true (or reference) value of the object, while precision refers to the repeatability or stability of repeated measurements under specified conditions. Common metrics for assessing accuracy include linear deviation, which quantifies the Euclidean distance between two measured points, and angular deviation, which describes the discrepancy between two measured axes.

The difference between the test scan and reference data was calculated, serving as the trueness including the linear and angular trueness of the ISBs (Fig. 6). Ten scans in each group were superimposed on each other using ISBs A–D as the alignment feature to generate a comparison, which served as precision evaluation for each group. Additionally, the accuracy of each position ISB was calculated separately as the local accuracy, and the accuracy of ISBs A–D was calculated together as the overall accuracy in each group.

Fig. 6.

Deviations between reference sacn and test scan: linear deviation (a) and angular deviation (α)

Statistical analysis

Statistical analyses were performed using SPSS 22.0 (IBM Corp., Armonk, NY, USA). The Shapiro-Wilk test and Levene’s test were used to assess the normality and variance homogeneity of the data. Differences in scanning accuracy between the groups were compared using one-way ANOVA and SNK-q tests for data with a normal distribution and variance assumptions. Otherwise, the Kruskal-Wallis and Bonferroni tests were applied. The level of significance was set at α = 0.05.

Results

Descriptive statistics are shown as mean (SD) and median [IQR]. The overall accuracy for the groups is shown in Table 1; Fig. 7. The local trueness of the groups is shown in Fig. 8; The local precision of the groups is shown in Fig. 9.

Table 1.

Descriptive statistics of groups for the global accuracy

	Trueness		Precision
Group	Linear (μm)	Angular (°)	Linear (μm)	Angular (°)
Ⅰ	171.50[161.55–211.01]a	0.81 [0.77–0.83]a	180.55 [158.28–189.48]a	0.74 [0.68–0.86]a
Ⅱ	65.90[58.14–73.68]b	0.48 [0.45–0.51]b	61.74 [54.59–69.19]b	0.45 [0.42–0.48]b
Ⅲ	42.88 [39.98–45.86]c	0.32 [0.27–0.37]c	41.03 [35.68–43.83]c	0.30 [0.27–0.33]c
Ⅳ	40.21 [34.35–48.28]c	0.24 [0.20–0.32]c	38.58 [34.95–43.23]c	0.27 [0.23–0.29]c
p value	<0.001		<0.001
Test	Kruskal-Wallis test		Kruskal-Wallis test

Different lowercase letters indicate the presence of significant differences between groups in multiple comparison tests (p<0.05)

Fig. 7.

The results of the global accuracy. A Linear deviations.B Angular deviations. Different lowercase letters indicate the presence of significant differences between groups in multiple comparison tests (p<0.05)

Fig. 8.

The results of the local accuracy for trueness.A Linear deviations.B Angular deviations. Different lowercase letters indicate the presence of significant differences between groups in multiple comparison tests (p<0.05)

Fig. 9.

The results of the local accuracy for precision. A Linear deviations. B Angular deviations. Different lowercase letters indicate the presence of significant differences between groups in multiple comparison tests (p<0.05)

Overall accuracy

Group I exhibited the largest linear and angular deviations in overall accuracy (p < 0.05), followed by group II (p < 0.05). A non-significant difference was found in the overall accuracy between groups III and IV (p > 0.05).

Local trueness

The linear and angular deviations of ISBs A, B, C and D in group I were significantly larger than that in groups II (p < 0.05), III (p < 0.05) and IV (p < 0.05), except for the linear deviations of ISB D, which were comparable to those in group II (p > 0.05). The linear deviations of ISBs C and D as well as the angular deviations of ISBs A, B, C and D in group II were significantly greater than that in groups III (p < 0.05) and IV (p < 0.05). A non-significant difference was found in local trueness between groups III and IV (p > 0.05).

Local precision

The linear and angular deviations of ISBs A, B, C, and D in group I were significantly larger than that in groups II (p < 0.05), III (p < 0.05) and Ⅳ (p < 0.05), except for the angular deviations of ISB D, which were comparable to those in group II (p > 0.05). The linear and angular deviations of ISBs A, B, C, and D in group II were significantly larger than those in groups III (p < 0.05) and IV (p < 0.05), except for the linear deviations of ISB B, which were comparable to those in group IV (p > 0.05). A non-significant difference was found in local precision between groups III and IV (p > 0.05).

Discussion

The present study compared differences in scanning accuracy between a modified scanning technique incorporating a SSP and SSA and the conventional scanning technique using the traditional scanning pattern, with or without a scan aid. The results indicated that the accuracy of complete-arch implant scan was significantly enhanced in both overall accuracy and local accuracy using the modified scanning technique (group III), achieving accuracy comparable to that of conventional splinted open-tray impression (group IV). Therefore, the null hypothesis was rejected.

Auxiliary devices are beneficial for improving the accuracy of complete-arch implant scan [17, 20, 27]. In previous studies, different types of scan aids have been used to provide geometric features that are easy for scanner recognition and image stitching, which have obtained improved scanning accuracy [20, 21]. Group I (without the SSA) and group II (with the SSA) were set up in this study to verify the role of the SSA. The results indicated that under the same scanning pattern recommended by the manufacturer, the overall accuracy of group II was greater than that of group I [25]. In terms of local accuracy, group II significantly outperformed group I with respect to accuracy of the ISBs A, B, and C. However, for the end ISB (ISB D), group II only outperformed group I in angle trueness and linear precision, suggesting that the advantage of the auxiliary device was no longer evident in the end ISB. Additionally, the overall accuracy of group II was significantly lower than that of group IV, indicating that use of the SSA alone was insufficient to achieve the same scanning accuracy as the splinted open-tray impression. However, this result was inconsistent with the findings of Wu et al. [22], who achieved an accuracy similar to that of the conventional splinted open-tray impression. This discrepancy may be attributed to differences in the design of the two experiments, such as the diameter of the ISB, the distance between the implants, and the type of IOS used.

The scanning pattern has an important effect on the scanning accuracy of complete-arch implant scan [23, 24]. Wu et al. [25] compared the impact of eight patterns on scanning accuracy and found that using the pattern recommended by the manufacturer achieved the best results, which may be attributed to less twisting of the scanning tip [28]. However, long scan paths still exist in previous scan patterns, which cause accumulated errors with increased span [10, 11, 13]. In this study, two short scans were performed in group III to shorten the span of a single scan and the resulting large deviation in the end ISB. The two independent scans were then aligned and merged in vitro via the n-point alignment and best fit algorithm, based on geometric features of the SSA between ISB B and ISB C. To verify the role of the novel scanning pattern, this study conducted a comparison between group II using the regular pattern and group III using the segmented scanning pattern. With both groups assisted by the SSA, group III achieved higher overall accuracy in all ISBs and local accuracy of ISB D, as compared with group II. Considering that the starting point of the scan for group III was in the middle of the edentulous arch, ISB A was also the endpoint of the scan. The results showed that the accuracy of ISB A in group III was also greater than that in group II.

Polyvinyl siloxane impression material (putty) provides a new option for the manufacturing of a scan aid. Zhang et al. [17] suggested that arcuate aids were advantageous for complete-arch implant scan. In this study, the putty was precisely shaped according to the curvature of the edentulous arch before curing. Then, irregular textures without sharp edges were added to the surface of the putty, which was beneficial for IOS identification [29]. Moreover, the putty had favorable optical properties for scanner recognition, due to its opacity and lack of a shiny surface [30]. In addition to the synthetic elastomeric impression materials used in this experiment, specialized scannable elastomeric materials with improved surface properties for scanning can also be considered [31]. Except for the above advantages, the SSA’s tissue surface information obtained using a desktop scanner can be used as soft tissue data, thereby saving chairside time.

When using the modified scanning technique, it is crucial to bear several essential points in mind. Longer ISB may be required in clinical practice to ensure that the putty is of appropriate thickness. The oil film on the putty surface, after curing, should be wiped clean to prevent the influence of a shiny surface on the scanning process. The SSA should be ensured with no displacement during the scanning process to avoid misalignment between segmented scans.

The primary limitations of this study was its in vitro nature including the lack of simulation of saliva, soft-tissue mobility, and patient movement. These factors may affect scanner recognition and increase image stitching errors [32–34], precluding direct extrapolation of the results to clinical applications. Additionally, the scans were performed in grouping order without randomization, which may introduce sequence effects or scanner-cooling biases despite the five-minute intervals.

Another potential bias in this study arose from the alignment strategy used for accuracy assessment. The global best-fit alignment of all four ISBs could result in bias averaging, as positive and negative errors might have offset each other [35]. The absolute linear deviation method was used in some studies [36, 37], as it could prevent the averaging of discrepancies. However, since deviations may involve six degrees of freedom, and this method primarily focuses on selected points rather than overall scanning deformation, it was limited by its inability to capture three-dimensional deformations. In addition, the passive fit of the framework was tested to evaluate scanning accuracy [38], which provided clinically meaningful data. However, a potential limitation was that the 3D marginal gap of the frameworks could not be fully assessed when capturing the vertical misfit, as gaps in the horizontal plane were not visible.

In subsequent studies, the impact of factors such as arch location, IOS platform, higher implant numbers (six or eight), and evaluation methods on this modified technique should be assessed. Additionally, in vivo accuracy evaluation of this technique is also needed.

Conclusions

Based on the findings of this in vitro study, the following conclusions could be drawn: Significant deviations were still observed at the end of the scanning path in group II when using a scan aid alone for assistance. The SSA combined with the SSP could significantly improve the accuracy of complete-arch implant scan.

Abbreviations

SSA

Simple scan aid

SSP

Segmented scanning pattern

CSP

Conventional scanning pattern

IOSs

Intraoral scanners

STL

Standard tessellation language

ANOVA

One-way analysis of variance

Authors’ contributions

L.W. contributed to the conceptualization, methodology, validation, investigation, formal analysis, and writing - review & editing. J.J. contributed to the conceptualization, methodology, validation, investigation, formal analysis, and writing - Original Draft. X.Z. contributed to the methodology, validation, formal analysis, and writing - Original Draft preparation. C.Y. contributed to the conceptualization, methodology, investigation, and resources. P.W. contributed to the conceptualization, methodology, data curation, writing - review & editing, supervision, project administration, funding acquisition. All authors have read and agreed to the published version of the manuscript.

Funding

This in vitro study was supported by the Science and Technology Bureau of Xuzhou city [grant numbers KC23150].

Data availability

The complete data and materials described in the research article are freely available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.