Precision and practical usefulness of intraoral scanners in implant dentistry: A systematic literature review

Background This systematic review aimed to evaluate the efficiency and accuracy of digital impression techniques for implant-supported restorations, and to assess their economic feasibility. Material and Methods Two independent electronic database searches were conducted in the Pubmed/MedLine, Cochrane Library, and Lilacs databases complimented by a manual search, selecting relevant clinical and in vitro studies published between 1st January 2009 and 28st February 2019. All type of studies (in vivo and in vitro) were included in this systematic review. Results Twenty-seven studies (8 in vivo and 19 in vitro studies) fulfilled the inclusion criteria. No meta-analysis was performed due to a large heterogeneity of the study protocols. The passive fit of superstructures on dental implants presented similar results between digital and conventional impression techniques. The studies considered that several factors influence the accuracy of implant impression taking: distance and angulation between implants, depth of placement, type of scanner, scanning strategy, characteristics of scanbody, and operator experience. Regarding the economic viability of intraoral scanning systems, only one study reported any benefit in comparison with conventional techniques. Conclusions Digital impressions of dental implants can be considered a viable alternative in cases of one or two contiguous dental implants. However, more studies are needed to evaluate the accuracy of digital techniques in full-arch implant-supported restorations. Key words:Intraoral scanner, dental implant, prosthesis, misfit, systematic review.


Introduction
It is many years since the long-term success of dental implants was confirmed by Branemark et al. and Albrektsson et al. (1,2) Since then, numerous studies have described new surgical and prosthodontic techniques that aim to improve the clinical outcomes of implant-based treatments (3,4). In cases of implant-supported restorations, treatment success depends on the superstructure's passive fit, as failure to achieve adequate passive fit can produce biological and mechanical complications (5). Fit depends on the accuracy of implant impression taking, which may be realised using long-established conventional techniques or more recently introduced digital techniques. The fabrication of an implant-supported prosthesis in a conventional workflow must start with the aid of an implant transfer post. Conventional impression taking can be classified as direct (pick-up) or indirect (transfer). With the introduction of digital technologies in dentistry, intraoral scanners can now be used for digital impression taking. According to the manufacturers, the use of intraoral scanners are a key element in the digital workflow, providing greater comfort for the patient, decreased turnaround time, and even a better cost-benefit ratio when compared to conventional techniques (6). But to date, no systematic literature review has been conducted to confirm the advantages of digital impression taking. In this context, this systematic literature review aimed to: (a) to determine if it is possible to achieve an adequate level of accuracy and efficiency using intraoral digital impression systems and to compare them with various conventional techniques for implant-supported restorations and (b) to assess the economic feasibility of digital techniques.

Material and Methods
This systematic review was conducted following PRIS-MA guidelines (7) and was registered in the Prospero database (trial no. CRD42015029504). The systematic review focused question was based on the PICO format (Population, Intervention, Comparison, Outcome) as follows: Population: healthy adult human patients. Intervention: conventional impression techniques. Comparison: digital impression taking with intra-oral scanners. Outcome: accuracy of impression and efficiency for fixed implant-supported restorations.
-Study Selection Criteria In order to identify relevant articles, the following inclusion criteria were applied: Clinical studies without language restriction that evaluated the accuracy of digital impressions taken with intraoral scanners or compared digital impression taking with conventional impression taking in treatment protocols leading to fixed implant-su-pported restorations. As the initial search generated only a few articles, and so insufficient scientific evidence, the search was extended to include in vitro studies. Finally, due to the heterogeneity of different articles it was not possible implement a meta-analysis. -Search Strategy An electronic search was conducted in the following databases: PubMed, Cochrane Library, Lilacs. Key search terms were applied, combined using MesH terms, to locate relevant articles published between 1st January 2009 and 28st February 2019. A additional manual search was conducted in the following journals: The search was carried out by two independent reviewers. Any disagreement between the reviewers (IGG and JC-BB) regarding data collection or quality assessment was resolved by consensus. Inter-reviewer reliability was assessed obtained a Kappa coefficient of 0.88 (CI 95%), values above 0.8 being considered a good level of agreement (8). To assess the quality of in vivo articles, the Critical Appraisal Skills Program (CASP) proposed by the Public Health Resource Unit (2006) was used, and only studies with an overall score of at least 50% were included in the review. Due to the small number of in vivo studies available, a duplicate search was performed to obtain in vitro studies. Although in vitro research cannot reproduce the dynamic environment of the stomatognathic system or human variability, pre-clinical experiments can provide important information about the properties and characteristics of a new material or technique. It is therefore necessary to conduct in vitro research of the highest possible standard. Efforts have been made in recent years to improve the quality of reporting in scientific literature (9,10). Although the CASP consort checklist was not originally designed for analyzing in vitro trials, in 2012 a modified consort checklist was published of items selected to assess reporting in vitro studies of dental materials. 18 The authors of the present review adapted this checklist for the purpose of comparing the accuracy of different dental implant impression-taking techniques. Only studies with an overall score of at least 50% were included in the review.

-Included Studies
An electronic search of the PubMed/MedLine, Cochrane Library and Lilacs databases located 1358 articles, which were reduced to 40 following title, abstract and full text analysis (PubMed/MedLine n=29; Lilacs n=7; Cochrane Library n=4). The articles from the different databases were compared to identify any duplicates, and a further 11 articles were eliminated on the basis of duplication (n=11) (Fig. 1). The ten remaining in vivo articles were categorized as follows: systematic reviews (n=5), randomized clinical trials (RCT) (n=1), prospective cohort studies (n=1), case-control studies (n=2), and case reports (n=1). The corresponding authors of the selected studies were contacted via email of whom four returned additional data. However, no additional data was included for analysis as all proved to be either replicate information or failed to meet the inclusion criteria. Due to the small number of in vivo studies available, the search was extended to include in vitro studies, using the same method, selecting 20 additional in vitro studies. These authors were also contacted via email, generating further data in three cases (n=3), but these were not included in the review for the same reasons as before. A modified CONSORT checklist of items for reporting in vitro studies was used to evaluate the risk of bias in the in vitro studies included (Fig. 2). When applying this modified CONSORT checklist to in vitro articles, points 5-9 could not be applied as they were designed to evaluate sample standardization. In the in vitro studies, the master model was the same in each study group, and so always standard. Of the articles evaluated, only one19 did not exceed the minimum score for inclusion in the review e787 (5/10), obtaining a score of 0/10 and so was eliminated (Table 1). Finally, the review included eight in vivo and 19 in vitro studies. The reasons for exclusion of various articles are specified in (Table 2).
-Outcomes Implant impressions can be obtained using open or closed tray, with or without splinting, using different impression materials (CI) or scanbody + an intraoral scanner system (DI). In order to carry out a complete analysis of the included articles, the outcomes were divided according to the technique(s) investigated: DI (17 studies), or CI vs. DI (12 studies) (Tables 3,4). 1. DI Seventeen studies used DI to take impression of dental implants: five systematic reviews, one case report, and eleven in vitro studies.

In Vivo
This case report describes DI in a patient with a fully edentulous jaw rehabilitated with six dental implants; three clinical tests were carried out to evaluate the accu-racy of the superstructure: saliva intrusion, the Sheffield test, and the screw resistance test, although the authors did not specify the fit values obtained (11).

In Vitro
Eleven in vitro studies were located that investigated the accuracy of IOS, divided into three subgroups: partially edentulous (PE), completely edentulous (CE), and partially and completely edentulous models (CE-PE). In Vitro -PE Three in vitro studies used DI-PE models (12)(13)(14). In 2012, Van der Meer et al. (12) carried out a study using a PE model with the aim of evaluating the accuracy of three different IOS. The authors concluded that the Lava COS was more accurate than the other IOS. Flugge et al. (13) employed two models bearing dental implants to compare the precision of three IOS with a laboratory scanner, obtaining a decrease in precision of the IOS when the distance between scan bodies increased, whereas with the dental lab scanner this was not dependent. Koch    Use of extraoral scanner, not intraoral scanner deviations between single tessellation language (STL) datasets of a master model, and milled model, and IOS from a previous single implant model. The authors concluded that direct digitization using the IOS presented less systematic error than physical model fabrication by milling from IOS.
In the studies carried out by Giménez et al., (15)(16)(17)(18)(19) pre      concluded that the accuracy of impressions with iTero® IOS (Cadent) decreased with the increased length of the scanned section but the angulation of dental implants did not affect scanning accuracy. In 2015, Giménez et al. (18) performed a study to assess the accuracy of two different IOS: ZFX Intrascan® (Zimmer Biomet, Dachau Germany) and 3D Progress® (MHT, Verona, Italy), concluding that neither IOS was suitable for taking impressions of dental implants in the full arch. In the same way, Giménez et al. (17) concluded that angulated and deep implant placement did not seem to decrease the system's accuracy with Lava COS® intraoral scanning system (3M ESPE), although accuracy was higher among experienced operators. Also in 2015, the same authors published another in vitro study of the CEREC AC Bluecam (Sirona) intraoral scanner. They concluded that neither angulation nor implant depth significantly affected scanner accuracy but operator experience did, with a tendency for less experienced operators to commit lower levels of error (16). In 2017, Giménez-González et al.

In Vivo
Comparisons between CI and DI were analyzed in four in vivo studies: a randomized crossover trial (23), two pilot studies (24,25), and one randomized clinical trial (26). Andriessen et al. (24) assessed the accuracy of IOS (iTero) in edentulous mandibles rehabilitated with overdentures compared with an extraoral laboratory scanner. They concluded that inter-implant distance and implant angulation were critical factors influencing the accuracy of intraoral scanning. Gherlone et al. (25) carried out two cases series studies with a similar design: CE rehabilitated with the "All on Four" protocol. In 2015, CI and DI (LAVA C.O.S scanner, 3M ESPE) were performed, assessing the accuracy of metallic structures through the use of an X-Ray (intraoral digital radiographs). In 2016, the patients were allocated either to the control group (CI) or test group (DI, using the Trios (3Shape). The authors concluded that it is possible to manufacture cobalt-chromium full-arch rehabilitations using computer-aided design/ computer-assisted manufacturing (CAD/CAM) from DI with satisfactory accuracy (26). Joda et al. (23) concluded that in addition to the multiple benefits offered by digital technology, DI allows a more efficient workflow in terms of cost when compared with CI.

Discussion
This systematic review was designed to evaluate the accuracy and efficiency of IOS for dental implant impression taking, compared with CI, and to assess the economic feasibility of introducing digital techniques.
The in vivo evidence located in the first search was scarce, further reduced by risk of bias determined by the CASP quality assessment (8 studies). So in order to expand the amount of information on the topic, an additional search was carried out expanding the criteria to include in vitro studies. In order to critically appraise the works identified, the authors adapted a previously published checklist18 for assessing the potential bias of in vitro studies. This checklist was initially designed to evaluate the quality of in vitro studies investigating dental materials. However, applying the checklist to the stu-dies selected in the present review, none fulfilled points 5 to 9. Point 5 of this checklist analyzes sample size, while points 6-9 analyze randomization (sequence generation, allocation concealment mechanism, implementation, and blinding). An in vitro study which evaluates dental implant impression-taking employs a previously designed model, with replicas of dental implants from which impressions are taken. The choice of model does not alter the results, as the models are manufactured industrially in advance and so the rate of error from model to model is negligible. In turn, there is no need for randomization, and sample size does not affect the results obtained. In this way, the authors of the present review used a modified version of the checklist published in 2012 by Faggion et al. (35), removing questions 5-9. In this way, the risk of bias and the quality of the in vitro studies analyzed were assessed by an appropriate, simple, and practical method. Because of the variability between the in vivo studies included and the fact that it was unclear how passive fit had been evaluated, comparisons of the results were not possible (11,(23)(24)(25)(26). Likewise, the in vitro studies reviewed could not be compared because of the different methods and IOS employed in both partial (27)(28)(29)(30)(31) and completely edentulous model (32)(33)(34). Nevertheless, most of the studies analyzed obtained results indicating sufficient accuracy, precision or trueness to guarantee adequate passive fit; especially on partially edentulous models. Several authors concluded that dental implant angulation and depth did not influence outcomes in terms of passive fit (15)(16)(17). Regarding the economic feasibility of DI, in comparisons between DI and CI, only a single in vivo study found that DI allowed a more efficient workflow than CI (23). Nevertheless, four systematic reviews have been conducted evaluating if there are any significant differences in accuracy between CI and DI (one in vitro study (36), two in vivo (37,38) studies and one that analyzed both in vivo and in vitro (39)studies) and all authors have concluded that the quality and quantity of the articles analyzed were insufficient. The present systematic review studied the same issue, analyzing both in vivo and in vitro studies, and adding one further objective, to determine the economic feasibility of DI.

Conclusions
Based on the data extracted from the articles analyzed in this systematic review, objectives could not be clearly and objectively addressed. It was not possible to determine which implant impression technique leads to better passive fit of superstructures. Digital techniques with intraoral scan impressions offer promising results, although improvements are still needed, particularly in full-arch impression taking. The available in vivo evidence is scarce, mainly case reports, which only provided low quality evidence. Randomized clinical studies comparing conventional and digital implant impression techniques are needed to generate decisive evidence. Finally, insufficient evidence was found regarding the economic feasibility of DI for implant-supported restorations, so additional research is needed to clarify this.