Comparison of Three-Dimensional Accuracy of Digital and

366 Volume 34, Number 2, 2019

©2019 by Quintessence Publishing Co Inc.

Comparison of Three-Dimensional Accuracy of Digital and Conventional Implant Impressions:

Effect of Interimplant Distance in an Edentulous ArchMing Yi Tan, BDS, MDS1/Sophia Hui Xin Yee, BDS, MDS2/Keng Mun Wong, BDS, MSD1/

Ying Han Tan, BDS, MSc1/Keson Beng Choon Tan, BDS, MSD1

Purpose: This study compared the three-dimensional (3D) accuracy of conventional impressions with digital

impression systems (intraoral scanners and dental laboratory scanners) for two different interimplant

distances in maxillary edentulous arches. Materials and Methods: Six impression systems comprising

one conventional impression material (Impregum), two intraoral scanners (TRIOS and True Definition), and

three dental laboratory scanners (Ceramill Map400, inEos X5, and D900) were evaluated on two completely

edentulous maxillary arch master models (A and B) with six and eight implants, respectively. Centroid positions

at the implant platform level were derived using either physical or virtual probe hits with a coordinate

measuring machine. Comparison of centroid positions between master and test models (n = 5) defined

linear distortions (dx, dy, dz), global linear distortions (dR), and 3D reference distance distortions between

implants (ΔR). The two-dimensional (2D) angles between the central axis of each implant to the x- or y-axes

were compared to derive absolute angular distortions (Absdθx, Absdθy). Results: Model A mean dR ranged

from 8.7 ± 8.3 µm to 731.7 ± 62.3 µm. Model B mean dR ranged from 16.3 ± 9 µm to 620.2 ± 63.2 µm.

Model A mean Absdθx ranged from 0.021 ± 0.205 degrees to –2.349 ± 0.166 degrees, and mean Absdθy

ranged from –0.002 ± 0.160 degrees to –0.932 ± 0.290 degrees. Model B mean Absdθx ranged from

–0.007 ± 0.076 degrees to –0.688 ± 0.574 degrees, and mean Absdθy ranged from –0.018 ± 0.048

degrees to –1.052 ± 0.297 degrees. One-way analysis of variance (ANOVA) by impression system revealed

significant differences among test groups for dR and ΔR in both models, with True Definition exhibiting the

poorest accuracy. Independent samples t tests for dR, between homologous implant location pairs in model

A vs B, revealed the presence of two to four significant pairings (out of seven possible) for the intraoral

scanner systems, in which instances dR was larger in model A by 110 to 150 µm. Conclusion: Reducing

interimplant distance may decrease global linear distortions for intraoral scanner systems, but had no

effect on Impregum and the dental laboratory scanner systems. Impregum consistently exhibited the best

or second-best accuracy at all implant locations, while True Definition exhibited the poorest accuracy for all

linear distortions in both models A and B. Impression systems could not be consistently ranked for absolute

angular distortions. Int J Oral MaxIllOfac IMplants 2019;34:366–380. doi: 10.11607/jomi.6855

Keywords: 3D accuracy, CMM, dental laboratory scanner, digital impression, implant, intraoral scanner

Digital dentistry appears to have reached a level ade-quate for clinical utility. For the impression-making

step, digitized mapping of intraoral architecture promises to eliminate the problems of conventional techniques. Intraoral scanner systems are generally ad-vocated for dentate arches or short edentulous spans, but there is mounting interest in extending their ap-plication to longer spans. Image acquisition is accom-plished by maneuvering a handheld intraoral scanner around relevant structures in the oral cavity, and these captured images are integrated to produce a digital three-dimensional (3D) replica of scanned structures.

Dental laboratory scanner systems similarly allow operators to adopt a digital workflow. The stone mod-el obtained from a conventional implant impression is scanned with a dental laboratory scanner to create a

1Faculty of Dentistry, National University of Singapore, Singapore, Republic of Singapore.

2Registrar, Khoo Teck Puat Hospital, Singapore, Republic of Singapore.

Correspondence to: Keson B. Tan, Faculty of Dentistry, National University of Singapore, 9 Lower Kent Ridge Road, Singapore 119085, Republic of Singapore. Fax: +65 67785742. Email: [email protected]

Submitted December 18, 2017; accepted July 3, 2018.

© 2019 BY QUINTESSENCE PUBLISHING CO, INC. PRINTING OF THIS DOCUMENT IS RESTRICTED TO PERSONAL USE ONLY. NO PART MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM WITHOUT WRITTEN PERMISSION FROM THE PUBLISHER.

The International Journal of Oral & Maxillofacial Implants 367

Tan et al

virtual model. However, problems inherent to conven-tional impressions remain.

The subsequent design and milling processes are similar for intraoral scanner and dental laboratory scanner systems. Resultant prostheses may be further characterized manually if needed.

Digital technology can provide much-anticipated technical, clinical, and procedural benefits, but a re-cent consensus conference highlighted the scarcity of appropriate clinical documentation to support wide-spread use of computer-aided design/computer-aided manufacturing (CAD/CAM) technology in implant reconstructions.1 There is an urgent need for good-quality long-term studies to validate the efficacy and effectiveness of digital options. However, new digital workflow parameters have been reported on but may not be directly relevant to clinical questions, and at-tempts to standardize terminology and communica-tion are still in their infancy. The American College of Prosthodontists (ACP) Digital Dentistry Glossary De-velopment Task Force acknowledged the dynamism of this field, and predicted the need for annual review of their newly minted glossary.2

Conventional impressions remain essential in implant prosthodontics, with several studies validating the supe-riority of polyether or addition polymerized silicones.3–5

With edentulous arches, Gherlone et al6 found that All-on-4 (Nobel Biocare) prostheses fabricated using

either conventional models or intraoral scans were comparable in terms of implant survival and crestal bone loss at 12 months. Papaspyridakos et al7 found that for a five-implant edentulous mandibular model, both the conventional impression technique with splinted impression copings and digital impression with the 3Shape TRIOS intraoral scanner were more accurate than conventional impressions with non-splinted impression copings. Vandeweghe et al8 evalu-ated four intraoral scanner systems on a six-implant edentulous mandibular model and found that 3M True Definition and 3Shape TRIOS fared better than Cerec Omnicam and Lava COS. In the latter two studies, the investigation of implant impressions in edentulous arches was limited to mandibular arches. Implants were placed within the interforaminal area, simulat-ing the common clinical scenario of full-arch implant-supported fixed prostheses in patients with early loss of posterior teeth. The interimplant distances in both studies were not clearly stated, but average mandibu-lar arch lengths interforaminally would result in im-plants placed relatively close together. This proximity may have skewed the results toward better accuracy if adjacent implants could be captured by the limited image field of the respective intraoral scanners.

The verdict on digital implant impression accuracy remains inconclusive. Many studies did not perform direct comparisons between conventional implant

Table 1 Features of Test Groups

Category

Conventional Digital (intraoral scanner) Digital (dental laboratory scanner)

Impregum, PentaSoft(3M ESPE)

TRIOS9

(3Shape)

True Definition10 (3M ESPE)

Ceramill Map40011,12

(Amann Girrbach)inEos X513

(Sirona)D90014

(3Shape)

Material/technology

Polyether Optical sectioning

Active wavefront sampling

Strip light projection Strip light projection with blue light

Triangulation using 4 × 5 megapixel cameras and blue LED

Impression coping/scan body

RC impression post (025.4202)

Core Scanbody (2077)


Range three Kit b Scan Body (792322) with screw (792341)

inPost Scan Body (6551639) on 2-CONnect abutment (L 810 M)


Straumann Core3D Centres

Core3D Centres

Amann Girrbach Sirona Core3D Centres

Need powder?

No No Yes No No No

Registers color?

No Yes No No No Yes

Manufacturer-specified accuracy

NA Not specified Not specified 20 µm 12 µm 15 µm

NA = not applicable.



Tan et al

impressions and the digital alternatives (intraoral scanner or dental laboratory scanner).8 Comparisons between studies could not be made due to the heter-ogenous study conditions and methodology. A perti-nent problem with the studies on edentulous arches was the lack of information on interimplant distance. Furthermore, many studies that utilized the superim-position method for distortion analysis7,8 did not com-pare test samples directly with the physical master models, but with digitized master models, the deriva-tion of which could introduce errors.

The purpose of this study, therefore, was to com-pare the 3D accuracy of implant positions obtained via a conventional impression system with that obtained from digital impression systems (intraoral scanner or dental laboratory scanner) for two different interim-plant distances in maxillary edentulous arches.

MATERIALS AND METHODS

Six impression systems, comprising one convention-al impression material, two intraoral scanners, and three dental laboratory scanners, were evaluated. The

features of each system are summarized in Table 1, and the experimental outline is illustrated in Fig 1.

Fabrication of Master ModelsTwo completely edentulous maxillary master models (Fig 2) that replicated the clinical situation for full-arch implant-supported fixed prostheses were fabrica-ted with heat-polymerized polymethyl methacrylate (PMMA) (Lucitone 199 Denture Base Resin, Dentsply). Master model A comprised six parallel implants (I, J, K, L, M, and N) spaced along the arch with approxi-mately 20-mm interimplant distance. Master model B comprised eight parallel implants (P, Q, R, S, T, U, V, and W) spaced approximately 13 mm apart.

The implants used were Straumann Bone Level implants RC (021.6412, 4.8 mmD and 12.0 mmL, Straumann) with the apical 9.0 mm secured with PMMA. Polyvinyl siloxane (PVS) impression material (Aquasil Ultra LV, Dentsply Caulk) simulated the peri-implant soft tissue around the coronal 3.0 mm of each implant.

Three Grade 5 silicon nitride ball bearings (8057222, Tsubaki Nakashima) of 8.0-mm diameter and manufacturer-specified sphericity of 0.2 µm were

Master models

Master model A(6 implants)

Master model B(8 implants)

Direct acquisition digital scanning with intraoral scanners

Conventional impression

(Impregum n = 5)

Direct acquisition digital scanning with intraoral scanners

Conventional impression

(Impregum n = 5)

Lab scanning on stone models using dental laboratory scanners

Lab scanning on stone models using dental laboratory scanners

TRIOSn = 5

True Definition

n = 5

Ceramill Map400

n = 5

inEos X5 n = 5

D900n = 5

TRIOSn = 5

True Definition

n = 5

Ceramill Map400

n = 5

inEos X5 n = 5

D900n = 5

Fig 1 Experimental outline, with six tested systems.

JK L

M

N

yz

I(0,0,0)

xI

III

II

QR

T

V

W

yz

P(0,0,0)

xI

III

II

SU

a b

Fig 2 (a) Schematic of master model A with six implants (I, J, K, L, M, N). (b) Schematic of master model B with eight implants (P, Q, R, S, T, U, V, W). I, II, III = silicon nitride ball bearings. Local coordi-nate system axes orientation with origin at centroid of implant I in model A and implant P in model B, respectively. X-axis is defined as I to N centroids and P to W centroids, respectively.



Tan et al

secured on the hard palate. Three-dimensional dis-tances between these ball bearings were monitored to ensure that the master models remained dimen-sionally stable throughout the study. Each master model was secured with PMMA to a custom aluminum block. The peripheral flat planes of this block served as vertical stops for the custom impression trays (Tray Resin II, Shofu). The aluminum block also facilitated securing of the master models to the worktable of the Coordinate Measuring Machine (CMM) (Global Silver Performance 7.10.7, Brown and Sharpe). Master mod-els were stored at 20°C for 30 days after completion to ensure that the models were dimensionally stable before proceeding.

Fabrication of Conventional Impression Test ModelsFive polyether (Impregum) impressions, dispensed from a Pentamix Automatic Mixing Unit (3M ESPE), were made of each master model via the open-tray impression technique and with torque application of 15 Ncm15–17 to the impression copings (RC impression post, 025.4202, Straumann), which were splinted with pattern resin (GC Pattern Resin LS, GC Corp) incremen-tally placed on a floss scaffold. To minimize the effect of polymerization shrinkage, the pattern resin segments on individual impression copings were initially kept separate, and final connection of the residual gap was done with a small increment. On removal of each im-pression, implant analogs (Bone Level Implant Analog, 025.4101, Straumann) were attached to the impression copings. Care was taken to prevent rotation of impres-sion copings within the impression material by stabi-lizing the implant analog. Type IV dental stone models (Silky Rock, Whip Mix) were poured using the manu-facturer’s recommended powder:water ratio, removed after 1 hour, and stored at 20°C for at least 7 days to mitigate the effects of dental stone expansion.

Scanning ProcedureThree types of scan bodies were used. Implant posi-tions in test groups TRIOS, True Definition, and D900 were registered using new precision-milled Core3D scan bodies (Core Scanbody, 2077, Core3D Centres). Test group Ceramill Map400 utilized Range three Kit b Scan bodies (792322, Amann Girrbach) and the corre-sponding Range three Kit a/b screws (792341, Amann Girrbach). Test group inEos X5 utilized a two-piece as-sembly comprising a 2-CONnect abutment for fixed prostheses and bars (2-CONnect L 810 M, Sirona) that inserts into the implant analog, and an inPost scan body (inPost for 2-CONnect KS61, 6551639, Sirona), which was hand-tightened on the abutment.

Dental Laboratory ScannerThe respective scan bodies for the dental laboratory scanner systems were secured with 10-Ncm torque to the implant analogs in the test stone models derived from Impregum conventional impressions, and one laboratory scan of each stone model was done using all three dental laboratory scanner systems. Additional scans were prescribed at specific locations on the stone model if the initial scan was deemed to be deficient.

Intraoral ScannerFive intraoral scans of each master model were per-formed using the intraoral scanner systems (TRIOS and True Definition), following insertion of the assigned scan bodies with 10-Ncm torque. Because True Defini-tion required the use of a powder spray (3M High Res-olution Scanning Spray, 3M ESPE), TRIOS scans were conducted first to avoid powder contamination of the master models. Both intraoral scanner systems have recommended scan strategies for dentate arches with natural teeth surfaces providing unique features for the stitching of images. However, these scan strategies do not work well for the edentulous arches with iden-tical scan bodies, and the manufacturers do not have published recommendations for edentulous arches. Help was sought from the respective manufactur-ers, and modified scan strategies were adopted with approval. The distal-most scan body in quadrant one was scanned first. Each scan body was captured fully in the buccal and palatal directions, before the scan pro-gressed further anteriorly. Increasing amounts of the palate were captured during the palatal scan of each scan body.

CMM Measurements of Master Models and Test ModelsImplant positions were determined by a CMM with manufacturer-specified measurement accuracy of 2 μm. For physical models comprising the two master models and the 10 stone models derived from Imp-regum conventional impressions, physical probe hits by the CMM spring-loaded ruby ball stylus of 1.0-mm diameter and 20.0-mm shank length (Renishaw Plc) allowed derivation of the Cartesian coordinates of da-tum points. The CMM metrology software (PC-DMIS CAD++ Version 2013 MR1, Wilcox Associates) then constructed geometric forms from these measured datum points. Virtual models from the intraoral scan-ner and dental laboratory scanner test groups were ex-ported as STL files from their respective software suites and imported into the same CMM metrology software, where virtual probe hits were made in the same man-ner as on physical models.



Tan et al

For master model implants and stone model im-plant analogs, the coronal implant platform was mea-sured via six probe hits to define a plane, while the internal conical portion was measured via 10 probe hits at two levels to define a cone. The centroid was found by constructing a pierce point between the cen-tral axis of the internal cone and the coronal flat plane (Fig 3).

For virtual model implants, the top platform of the virtual scan body was measured via four virtual probe hits to define a plane. The axial portion of each scan body was measured via eight virtual probe hits at two levels to define a cylinder (TRIOS, True Definition, inEos X5 and D900) or cone (Ceramill Map400). CMM measurements confirmed that for the three scan body types from Core3D, Amann Girrbach, and Sirona, the nominal scan-body heights were 10.00, 10.40, and 12.18 mm, respectively. The virtual implant centroid was defined by constructing a pierce point between the central axis of the virtual scan-body cylinder or cone and a constructed plane offset by the magnitude of the respective scan-body heights, apical to the top platform plane (Fig 3).

Each geometric feature was measured three times to minimize the magnitude of errors. A verification check using the CMM software tolerance criterion functions was conducted after each measurement. The tolerance value for all physical geometric features was set at 6 μm, while that for virtual geometric fea-tures was set at 20 μm. The higher tolerance limit in virtual models served to compensate for minor surface

irregularities inherent in STL files. Three-dimensional distances between the three centroids of the same constructed feature were calculated, and a limit of ac-ceptable deviation was set at 10 μm for both physical and virtual models.

The alignment setup for model A consisted of an XY-plane formed by implants I, K, and N; X-axis formed by implants I and N; and origin of local coordinate sys-tem determined by the centroid position of implant I. The alignment setup for model B consisted of an XY-plane formed by implants P, S, and W; X-axis formed by implants P and W; and origin of local coordinate sys-tem determined by the centroid position of implant P (Fig 2).

Distortion ParametersDistortion values were derived by calculating the differ-ence in 3D coordinate values between master and test models. Three linear distortions (dx, dy, dz) were com-puted for each implant, indicating distortion along the x-, y-, and z-axes, respectively. Global linear distortion (dR) is defined by the formula dR = √(dx

2 + dy2 + dz

2).Two absolute angular distortions (Absdθx and

Absdθy), occurring about the x- and y-axes, respec-tively, were computed. Absdθx reflects anteropos-terior angular distortions, while Absdθy reflects mediolateral angular distortions.

Three-dimensional reference distances (R) between the centroids of implants in test models were com-pared with those in the master models to derive 3D reference distance distortions (ΔR).

Scan-body central axis

Scan-body top platform plane

Distance between planes = scan-body height

Implant platform plane

Implant internal cone

Implant central axis

Implant centroid

Fig 3 (a) Implant centroid was defined by constructing a pierce point between the central axis of the virtual scan body/implant inter-nal cone and the implant platform plane. (b) Scan-body cylinder (TRIOS, True Definition, inEos X5, D900) or cone (Ceramill Map400) was defined by eight virtual probe hits at two levels. (c) Scan-body top plane was defined by four virtual probe hits. (d) Virtual scan body with measured features.

a

b c

d



Tan et al

To allow comparison between models A and B, homologous implant location pairs were assigned (Table 2).

The dependent variables were dx, dy, dz, dR, Absdθx, Absdθy, and ΔR, as detailed earlier for each of the six- or eight-implant models. One-way analysis of variance (ANOVA) and post hoc Tukey honestly significant dif-ference (HSD) test were used for the comparison of the six impression systems in each of the interimplant dis-tance models (A or B), by implant location. Independent samples t tests between homologous implant location pairs were conducted to compare model A with model B. Significance was reported at a level of alpha = .05. The

standard deviations of the measured variables reflect the precision of each impression system. All statistical calculations were performed using statistical software (SPSS Statistics v21.0, IBM).

RESULTS

The 3D linear distortions (dx, dy, dz) as well as the glob-al linear distortion (dR) for models A and B are shown in Figs 4 and 5 and Tables 3 and 4. Model A mean dR ranged from 8.7 ± 8.3 µm to 731.7 ± 62.3 µm. Model B mean dR ranged from 16.3 ± 9 µm to 620.2 ± 63.2 µm.

Table 2 Homologous Implant Location Pairs

Homologous implant location pairs (Model A ↔ Model B) Rationale

J ↔ Q Implant in right premolar region

J ↔ R Implant in right premolar region

K ↔ S Most anterior, right implant location

L ↔ T Most anterior, left implant location

M ↔ U Implant in left premolar region

M ↔ V Implant in left premolar region

N ↔ W Direct cross-arch implant location

Table 3 Model A Global Linear, Absolute Angular, and 3D Reference Distance Distortions

Impregum TRIOS True Definition Ceramill Map400 inEos X5 D900

Mean dR, μm (SD)I 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0)J 10.9 (5.0) 172.5 (74.2) 291.6 (79.8) 52.8 (25.3) 25.4 (8.6) 62.3 (14.4)K 15.6 (6.6) 153.6 (40.7) 482.6 (93.1) 49.9 (13.1) 54.2 (45.5) 53.9 (9.1)L 24.6 (10.2) 143.4 (70.7) 716.6 (116.0) 49.8 (20.3) 53.3 (14.5) 59.5 (11.5)M 24.4 (9.6) 100.0 (57.9) 731.7 (62.3) 42.5 (12.7) 43.9 (13.9) 87.2 (31.1)N 20.1 (9.3) 36.3 (28.6) 677.1 (50.3) 8.7 (8.3) 35.0 (19.2) 59.8 (40.0)

Mean Absd𝛉x, deg (SD)I –0.273 (0.069) –0.676 (0.338) –0.875 (0.866) –0.827 (0.185) –0.618 (0.071) –0.537 (0.269)J –0.035 (0.163) 0.021 (0.205) –0.037 (0.520) –0.359 (0.078) –0.129 (0.154) –0.150 (0.175)K –0.096 (0.133) –2.189 (0.453) –2.249 (1.100) –2.349 (0.166) –1.620 (0.539) –1.944 (0.671)L 0.153 (0.133) –0.096 (0.241) –0.243 (0.684) –0.073 (0.074) 0.054 (0.039) –0.159 (0.130)M 0.276 (0.250) 0.306 (0.106) –0.063 (0.414) –0.036 (0.305) 0.224 (0.081) –0.437 (0.606)N 0.151 (0.399) –0.313 (0.374) –0.437 (1.302) –0.757 (0.240) –0.350 (0.319) –0.269 (0.358)

Mean Absd𝛉y, deg (SD)I –0.089 (0.069) 0.215 (0.292) 0.469 (0.320) 0.611 (0.351) 0.533 (0.133) 0.295 (0.208)J –0.075 (0.127) –0.141 (0.304) –0.219 (0.272) –0.229 (0.302) –0.233 (0.071) –0.222 (0.051)K –0.039 (0.082) –0.932 (0.290) –0.715 (0.360) –0.774 (0.180) –0.537 (0.158) –0.347 (0.187)L 0.138 (0.135) 0.136 (0.241) –0.143 (0.245) 0.125 (0.098) 0.064 (0.059) 0.220 (0.349)M 0.322 (0.106) 0.186 (0.072) –0.304 (0.346) 0.031 (0.266) 0.106 (0.127) 0.472 (0.362)N 0.061 (0.151) 0.110 (0.201) –0.155 (0.257) 0.307 (0.116) –0.002 (0.160) 0.446 (0.462)Mean ΔR, μm (SD)I-J 1.9 (10.3) 166.8 (78.0) –267.5 (85.4) 28.9 (36.1) 11.1 (9.3) –12.3 (37.0)I-K 11.1 (12.6) 119.0 (61.4) –480.9 (92.9) 29.5 (32.9) 25.4 (10.8) 2.7 (32.6)I-L 22.0 (12.2) 71.2 (107.8) –703.9 (121.1) 35.8 (31.9) 45.4 (20.2) –6.4 (39.1)I-M 23.4 (9.5) 26.4 (39.6) –709.2 (66.8) 30.5 (24.6) 40.1 (15.1) –49.8 (23.1)I-N 20.1 (9.3) 13.3 (47.4) –677.1 (50.3) –4.8 (11.6) 35.0 (19.2) –59.8 (40.0)



Tan et al

I J K LMN I J K LMN I J K LMN I J K LMN

800

600

400

200

0

–200

–400

–600

–800

dy dz dR

Mod

el A

line

ar d

isto

rtion

s (µ

m)

dx

ImpregumCeramill Map400inEos X5

TRIOSTrue DefinitionD900

Fig 4 Model A linear distortions (µm) by implant location.

800

600

400

200

0

–200

–400

–600

–800

dy dz dR

Mod

el B

line

ar d

isto

rtion

s (µ

m)

dx



Fig 5 Model B linear distortions (µm) by implant location.

Table 4 Model B Global Linear, Absolute Angular, and 3D Reference Distance Distortions


Mean dR, μm (SD)P 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0)Q 16.3 (9.0) 40.2 (19.4) 160.9 (20.1) 48.9 (7.3) 23.4 (9.9) 68.7 (27.6)R 19.1 (9.8) 40.8 (36.3) 369.1 (64.5) 30.2 (7.4) 31.0 (5.7) 62.7 (35.3)S 27.3 (10.6) 73.5 (35.9) 464.8 (100.0) 36.2 (13.5) 48.0 (9.9) 46.1 (25.3)T 35.3 (8.4) 148.0 (35.4) 611.1 (85.2) 56.1 (15.8) 60.3 (15.0) 65.6 (23.9)U 39.3 (5.8) 75.3 (31.7) 620.2 (63.2) 46.3 (14.7) 65.4 (10.7) 47.1 (27.9)V 34.8 (2.5) 110.4 (86.6) 582.1 (61.5) 37.4 (21.5) 64.7 (4.7) 53.0 (42.8)W 35.5 (9.2) 59.9 (47.9) 567.8 (30.5) 32.3 (25.7) 66.4 (5.9) 73.4 (24.0)

Mean Absd𝛉x, deg (SD)P –0.151 (0.142) 0.532 (0.326) 0.511 (0.379) 0.596 (0.141) 0.136 (0.170) 0.140 (0.371)Q 0.022 (0.239) –0.112 (0.214) 0.021 (0.157) –0.153 (0.187) –0.034 (0.031) 0.226 (0.314)R 0.205 (0.147) 0.432 (0.305) 0.130 (0.303) 0.226 (0.261) 0.134 (0.276) 0.306 (0.261)S –0.243 (0.417) –0.476 (0.252) –0.688 (0.574) –0.459 (0.222) –0.349 (0.105) –0.312 (0.321)T –0.038 (0.269) 0.330 (0.340) 0.335 (0.242) 0.329 (0.112) 0.160 (0.054) 0.041 (0.123)U 0.233 (0.112) –0.361 (0.313) –0.135 (0.304) –0.109 (0.146) –0.032 (0.067) –0.146 (0.130)V –0.110 (0.130) –0.108 (0.293) –0.260 (0.183) 0.069 (0.137) –0.182 (0.046) –0.227 (0.075)W –0.007 (0.076) 0.232 (0.164) 0.226 (0.162) 0.448 (0.141) 0.224 (0.101) 0.161 (0.071)

Mean Absd𝛉y, deg (SD)P –0.091 (0.116) 0.150 (0.188) 0.185 (0.451) 0.284 (0.244) –0.018 (0.048) –0.153 (0.342)Q 0.162 (0.091) 0.255 (0.055) 0.144 (0.178) 0.484 (0.123) 0.328 (0.022) 0.196 (0.447)R –0.066 (0.102) 0.377 (0.081) 0.706 (0.547) 0.706 (0.141) 0.136 (0.252) 0.387 (0.104)S –0.084 (0.129) –0.812 (0.330) –1.052 (0.297) –0.795 (0.112) –0.526 (0.139) –0.376 (0.175)T 0.067 (0.152) 0.117 (0.141) 0.145 (0.068) 0.435 (0.104) 0.154 (0.180) 0.155 (0.200)U 0.267 (0.605) 0.161 (0.107) 0.083 (0.150) 0.493 (0.176) 0.178 (0.068) 0.043 (0.109)V –0.053 (0.100) 0.333 (0.314) 0.134 (0.288) 0.632 (0.187) 0.180 (0.084) 0.207 (0.149)W 0.080 (0.146) –0.714 (0.315) –0.784 (0.340) –0.839 (0.296) –0.877 (0.381) –0.707 (0.240)Mean ΔR, μm (SD)P-Q 6.0 (2.3) 12.6 (17.2) –151.1 (32.8) 22.6 (10.3) 14.6 (7.5) –34.7 (28.8)P-R 8.6 (8.5) –9.1 (28.9) –365.3 (68.7) 9.8 (15.7) 27.6 (9.1) –16.8 (40.6)P-S 16.9 (9.5) 28.7 (69.0) –463.4 (99.9) 29.1 (10.9) 42.2 (14.0) –4.2 (26.3)P-T 28.6 (11.6) 38.8 (126.6) –596.4 (90.3) 50.2 (20.9) 57.3 (16.1) 9.4 (28.9)P-U 31.2 (11.2) 33.3 (41.6) –602.5 (70.0) 37.8 (20.9) 62.6 (12.2) 9.5 (45.9)P-V 30.5 (5.9) 69.8 (109.2) –574.2 (63.0) 23.4 (31.0) 62.3 (5.1) 23.9 (61.6)P-W 35.5 (9.2) 50.1 (60.4) –567.8 (30.5) 28.8 (30.4) 66.4 (5.9) –22.2 (81.8)

P R TUVWQ S P R TUVWQ S P R TUVWQ S P R TUVWQ S



Tan et al

The absolute angular distortions (Absdθx and Absdθy) for models A and B are shown in Figs 6 and 7 and Tables 3 and 4. Model A mean Absdθx ranged from 0.021 ± 0.205 degrees to –2.349 ± 0.166 degrees, while mean Absdθy ranged from –0.002 ± 0.160 degrees to –0.932 ± 0.290 degrees. Model B mean Absdθx ranged from –0.007 ± 0.076 degrees to –0.688 ± 0.574 degrees, while mean Absdθy ranged from –0.018 ± 0.048 de-grees to –1.052 ± 0.297 degrees.

Mean 3D reference distance distortions (ΔR) for models A and B are shown in Figs 8 and 9.

The data were initially submitted to the Shapiro-Wilk test to check for the assumption of normality, and it was not rejected. One-way ANOVA by impression system revealed significant differences among the test groups for dR and ΔR in both models A and B for all implant locations except the origin. For Absdθx, signifi-cant differences among the test groups were found for implants K and M in model A, as well as implants P, T, U, and W in model B. For Absdθy, significant differences among the test groups were found for implants I, K, M, and N in model A, as well as R, S, T, V, and W in model B. Tables 5 and 6 summarize the results of the one-way ANOVA and Tukey HSD procedures performed.

Independent samples t tests for dR, between ho-mologous implant location pairs in model A vs model B, revealed the presence of two to four significant pair-ings out of seven possible pairings for test groups Im-pregum, TRIOS, True Definition, and inEos X5 (Table 7). dR was larger in model A than in model B for the intra-oral scanner systems by 110 to 150 µm. dR was larger in

model B than in model A for Impregum and inEos X5, but by a much smaller extent of less than 32 µm. In-terimplant distance did not affect Absdθx or Absdθy in any consistent manner (Table 7).

DISCUSSION

Implant impression systems may be broadly catego-rized into three groups: (1) conventional impression, (2) digital impression with intraoral scanner, and (3) digital impression with dental laboratory scanner. In general, the systems utilizing intraoral scanners were found to fare more poorly, although this reached sig-nificance only for True Definition in most instances. These trends were applicable to both the six-implant (model A) and eight-implant (model B) configurations.

Linear DistortionsThe results showed significantly poorer performance of True Definition for all linear distortions. Another interesting finding related to True Definition was the consistent negative linear distortions in the x- and y-directions for both models A and B. Negative dx and dy indicated “shrinkage” in virtual model arch width and height, respectively, which can lead to fabrication of a smaller framework.

Absolute Angular DistortionsThe impression systems could not be consistently ranked for absolute angular distortions. Negative

I IJ JK KL LM MN N

1.00.5

0–0.5–1.0–1.5–2.0–2.5

Mod

el A

abs

olut

e

angu

lar d

isto

rtion

s (d

eg)

Absdθx Absdθy



Fig 6 Model A absolute angular distortions (deg) by implant location.

P Q S UR T V W P Q S UR T V W

1.00.5

0–0.5–1.0–1.5–2.0–2.5

Mod

el B

abs

olut

e

angu

lar d

isto

rtion

s (d

eg)

Absdθx Absdθy



Fig 7 Model B absolute angular distortions (deg) by implant location.

I-J I-K I-L I-M I-N

300200100

0–100–200–300–400–500–600–700–800

Mod

el A

3D

refe

renc

e di

stan

ce

dist

ortio

ns (∆

R) (µ

m)

∆R

Fig 8 Model A 3D reference distance distortions (µm).

ImpregumCeramill Map400inEos X5TRIOSTrue DefinitionD900

P-Q P-R P-S P-T P-U P-V P-W

300200100

0–100–200–300–400–500–600–700–800

Mod

el B

3D

refe

renc

e di

stan

ce

dist

ortio

ns (∆

R) (µ

m)

∆R

Fig 9 Model B 3D reference distance distortions (µm).

ImpregumCeramill Map400inEos X5TRIOSTrue DefinitionD900



Tan et al

Table 5 Summary of One-Way ANOVA and Tukey HSD for Global Linear and Absolute Angular Distortions, by Implant Location

Variable/model/implant location P value

Statistical significant subsets


dR

Model A I .000 – – – – – – J .000 a b c a a a K .000 a b c a a a L .000 a b c a b a b a b M .000 a a b a a a N .000 a a b a a a Model B P .000 – – – – – – Q .000 a a b c a b a b R .000 a a b a a a S .000 a a b a a a T .000 a b c a a a U .000 a a b a a a V .000 a a b a a a W .000 a a b a a a

Absd𝛉x

Model A I .243 a a a a a a J .265 a a a a a a K .000 b a a a a a L .392 a a a a a a M .018 b b a b a b a b a N .372 a a a a a a Model B P .001 a b b b a b a b Q .109 a a a a a a R .471 a a a a a a S .426 a a a a a a T .034 a a a a a a U .005 b a a b a b a b a b V .054 a b a b a b a b a b W .000 a a b a b b a b a

Absd𝛉y

Model A I .002 a a b b b b a b J .812 a a a a a a K .000 c a a b a b a b b c L .168 a a a a a a M .001 b b a a b a b b N .013 a b a b a a b a b b Model B P .105 a a a a a a Q .131 a a a a a a R .000 a a b b b a a b S .000 d a b a a b c b c c d T .010 a a a b a b a b U .161 a a a a a a V .001 a a b a b a a W .000 b a a a a a

Letters indicate specific implant locations as defined in Fig 2.



Tan et al

Absdθx values indicate posterior tilting of implants to-ward the palate, and negative Absdθy values indicate the tilting of implants toward the patient’s right side (Figs 2a and 2b). Except for Impregum, it was observed that the greatest absolute angular distortions typically occurred at implant K for model A and implants S or W for model B. Implants K and S were diagonally placed with respect to their respective origins (I or P), while implants W and N were located across the arch from origins (I or P). The consistency in sites affected by the greatest absolute angular distortions hints at possible challenges when using digital systems to register di-agonally placed implants or when crossing the dental arch. However, this did not reach statistical significance.

3D Reference Distance DistortionsThe mathematical derivation of the global linear dis-tortion (dR) in itself will always derive a positive value. Therefore, 3D reference distance distortions, ΔR, be-tween all implant locations to the respective origins (I in model A and P in model B) were used to reveal increases or decreases in R.

The results showed significantly poorer perfor-mance of True Definition for all 3D reference distance distortions. The consistently negative 3D reference distance distortion values correspond to the negative linear distortions (dx and dy) that were described ear-lier. This implied shrinkage of virtual models when True Definition was utilized, and this distortion ranged from

Table 6 Summary of One-way ANOVA and Tukey HSD for 3D Reference Distance (ΔR) Distortions

Variable/model/implant location P value

Statistical significant subsets


ΔR Model A

I-J .000 b c a b b b I-K .000 b c a b b b I-L .000 b b a b b b I-M .000 c c a c c b I-N .000 c c a b c c b Model B

P-Q .000 c c a c c b P-R .000 b b a b b b P-S .000 b b a b b b P-T .000 b b a b b b P-U .000 b b a b b b P-V .000 b b a b b b P-W .000 b b a b b b

Letters indicate specific implant location pairings to define the reference distances (see Fig 2).

Table 7 Summary of Independent Samples t Test for Homologous Implant Location Pairs Between Models A and B

Homologous implant location pairs (A ↔ B)

dR Absd𝛉x Absd𝛉y

A > B B > A A > B B > A A > B B > A

J ↔ Q TRIOS True Definition

TRIOS True Definition

Impregum

J ↔ R TRIOS TRIOS inEos X5

K ↔ S

L ↔ T True Definition

M ↔ U True Definition Impregum inEos X5

True Definition D900

M ↔ V True Definition Impregum inEos X5

True Definition True Definition D900

inEos X5

N ↔ W True Definition Impregum inEos X5

True Definition Ceramill Map400inEos X5

Impregum True Definition inEos X5

Significantly different test groups listed.



Tan et al

–268 to –709 µm in model A and from –151 to –602 µm in model B.

D900 showed a similar tendency for negative 3D distance distortions for some reference distances, but the magnitude of distortion was much lower with ΔR ranging from –60 to +3 µm in model A and from –4 to –35 µm in model B. The remaining four impression systems displayed an opposite trend, with distortions tending to be positive, implying an expansion of the resultant model.

Mean ΔR results for TRIOS in model A appeared to be counterintuitive, with mean ΔR decreasing as the magnitude of the reference distances increased. The results were verified to be correct, and taking into ac-count the large standard deviations, ΔR for TRIOS for the various 3D reference distances were actually not significantly different.

Comparison Between Models A and BFor Impregum, mean dR was found to be statistically greater in model B than in model A for three out of sev-en possible homologous implant location pairs. The magnitude of this difference was, however, less than 16 µm and unlikely to be clinically significant.

Differences in global linear distortion magnitude between models A and B were detected more readily for the intraoral scanner systems. Of the homologous implant location pairs that showed significance, the distortions noted for model A were consistently great-er than model B by 110 to 150 µm. The limited image fields of the respective intraoral scanner systems may have been a limitation in image acquisition and stitch-ing of the longer edentulous spans in model A.

For the dental laboratory scanner systems, significant homologous implant location pairs were noted only for inEos X5, for which mean dR was found to be signifi-cantly greater in model B than in model A for three out of seven possible pairings. The magnitude of this dif-ference was greater than for Impregum, which may be because a conventional impression was required in the workflow for dental laboratory scanner systems. None-theless, the difference between models was less than 32 µm and unlikely to be clinically significant.

Comparison with Previous StudiesThree-dimensional reference distance distortions for Impregum were found to be comparable to the re-sults shown by Wee,4 who compared four linear dis-tances between test and master models replicating the clinical scenario of an edentulous arch with five implants. For polyether, mean deviation was reported as 16.2 ± 8.8 µm. In the present study, ΔR for Impregum ranged from 1.9 ± 10.3 µm to 23.4 ± 9.5 µm in model A and ranged from 6.0 ± 2.3 µm to 35.5 ± 9.2 µm in model B.

Papaspyridakos et al7 reported that the median 3D deviation was 5 to 13 µm for conventional impressions with splinted impression copings, 13 to 132 µm for conventional impressions with nonsplinted impres-sion copings, and 8 to 29 µm for digital impression with a 3Shape TRIOS intraoral scanner. Vandeweghe et al8 reported mean deviations for Lava COS, 3M True Definition, Cerec Omnicam, and 3Shape TRIOS to be 112 ± 25 µm, 35 ± 12 µm, 61 ± 23 µm, and 28 ± 7 µm, respectively. However, these values should not be compared directly with the present study, due to fun-damental differences in methodologies for assessment of accuracy. The methodologies previously reported have utilized linear distance measurements,18 scan-ning of calibrated objects of known dimensions,19 measuring fit of definitive prostheses,20 superimpo-sition of digital images,21 and use of a CMM.15,16 The aforementioned two studies on multiple implants for the edentulous arch7,8 compared test and control groups by superimposing test scans with reference scans; hence, all surfaces of each scan body contrib-uted to the measured distortion. On the other hand, this study utilized the CMM for distortion analysis at the implant platform level. For model A, dR for TRIOS ranged from 36.3 ± 28.6 µm to 172.5 ± 74.2 µm, while that for True Definition ranged from 291.6 ± 79.8 µm to 731.7 ± 62.3 µm. For model B, dR for TRIOS ranged from 40.2 ± 19.4 µm to 148.0 ± 35.4 µm, while that for True Definition ranged from 160.9 ± 20.1 µm to 620.2 ± 63.2 µm.

Considerations for Selecting Implant Impression System for the Edentulous ArchThe dental laboratory scanner systems’ fabrication workflow incorporates a conventional impression. By calculating the mean differences in 3D reference dis-tance distortions between Impregum and the dental laboratory scanner systems, the magnitude of distor-tion resulting from the digital laboratory scanning step only may be quantified. Reference distances I-L and P-T in models A and B, respectively, were analyzed to de-termine accuracy between diagonally placed implants in an edentulous arch, while reference distances I-N and P-W were analyzed to determine cross-arch ac-curacy (Table 8). The magnitude of distortion result-ing from the dental laboratory scanner scanning step alone for Ceramill Map400, inEos X5, and D900 was found to range from 13.8 to –28.4 μm for I-L; 28.7 to –19.2 μm for P-T; 14.9 to –79.9 μm for I-N; and 6.7 to –57.7 μm for P-W.

This study demonstrated that impression system contributed to varying levels of distortion. However, the ensuing fabrication steps also contribute to overall prosthesis accuracy. Tan et al,22 using an edentulous mandibular master cast with five parallel implants,



Tan et al

quantified distortions in full-arch implant prosthesis castings. Overall linear displacements for the five cylin-ders (implants) were found to be 20.2 ± 14.5 μm. Over-all angular displacements for all five cylinders were –0.098 ± 0.109 degrees for dθx and 0.020 ± 0.097 de-grees for dθy. Unfortunately, studies evaluating CAD/CAM fabrication of full-arch implant prostheses using a similar distortion measurement methodology are lacking. Örtorp et al23 compared conventional castings with milled frameworks, and concluded that the milled frameworks exhibited better fit and precision. Howev-er, distortions were calculated by the superimposition technique, and the results cannot be directly extrap-olated to the present study. Further investigation is warranted to identify the direction and magnitude of distortion at each stage of CAD/CAM prosthesis fab-rication. The clinician or technician may then balance positive distortions with negative distortions to mini-mize overall inaccuracy.

Software compatibility with a variety of scan bodies and implant systems influences utility. Two of the in-vestigated dental laboratory scanner systems (Ceramill Map400 and inEos X5) required use of manufacturer-specified scan bodies, while the investigated intraoral scanner systems (TRIOS and True Definition) and one dental laboratory scanner (D900) allowed use of third-party scan bodies. Impression systems that allow use of third-party scan bodies may simplify a clinic’s invento-ry. However, it was noted that the standard deviations

for those impression systems that utilize third-party scan bodies were larger, though these trends did not reach statistical significance. One speculation for this lower precision is a lack of optimization by the vari-ous scanners, possibly leading to poorer scan body recognition.

Fit of scan body into the implant/implant analog is another potential source of error; the coaxiality of the scan body and implant will influence angular dis-tortions in particular. All three scan bodies used in this study were equally challenged in this aspect, as none were produced by the implant manufacturer. Stimmelmayr et al24 evaluated the reproducibility of implant scan body fit, but the protocol included digi-tization of models with their attached scan bodies. Er-rors inherent to this digitization process may obscure the true distortions that had resulted from poor fit of scan bodies to the implant or implant analog. More specific investigations are required to determine the effect of scan body fit on scan accuracy.

Due to the lack of precise instructions from manu-facturers on the recommended torque magnitude for scan body insertion, pilot studies were conducted to investigate the effect of torque magnitude on scan body distortion and thus scan body height. Five scan bodies from each type were tested on a single implant analog. Each scan body was secured with increasing torque magnitudes of 5, 10, and 15 Ncm. Six probe hits were used for the registration of the top flat surface

Table 8 3D Distance Distortion Attributable to Laboratory Scanning

Reference distanceMean 3D distance distortion

in Impregum (µm)

Mean 3D distance distortion in dental laboratory scanner

systems (µm)Mean 3D distance distortion

attributable to laboratory scanning (µm)

Diagonal

I-L (Model A) 22.0

Ceramill Map400 35.8 13.8

inEos X5 45.4 23.4

D900 –6.4 –28.4

P-T (Model B) 28.6

Ceramill Map400 50.2 21.6

inEos X5 57.3 28.7

D900 9.4 –19.2

Cross-arch

I-N (Model A) 20.1

Ceramill Map400 –4.8 –24.9

inEos X5 35.0 14.9

D900 –59.8 –79.9

P-W (Model B) 35.5

Ceramill Map400 28.8 –6.7

inEos X5 66.4 30.9

D900 –22.2 –57.7



Tan et al

of the scan body, while eight probe hits were used to register the external cylindrical (TRIOS, True Definition, inEos X5, and D900) or conical (Ceramill Map400) sur-faces. The pierce point between the central axis and the top plane of each scan body defined the “scan body centroid.” Scan bodies were subsequently re-moved to allow for CMM registration of the implant analog coronal flat plane and internal conical surface; these registrations then allowed for the pierce point to be calculated, deriving the “analog centroid.” The 3D distance between each scan body centroid and the analog centroid was calculated. At the same time, co-axiality values were monitored.

Mean coaxiality values for Core3D scan bodies were found to be much higher than those of the Amann Gir-rbach and Sirona scan bodies (Fig 10). These values were also much higher than those found in an earlier study15,16 that utilized an earlier version of the Core3D scan bodies. The older Core3D scan body was entirely constructed with PEEK, but the newer version used in this study consisted of a lower metal segment that sits on the coronal flat plane of each implant. Possible im-proper mating of PEEK to metal during the scan-body fabrication process may have resulted in a noncylindri-cal scan body, and consequently, poorer coaxiality with respect to the corresponding implant. This may also ex-plain the counterintuitive trends in scan-body height for the Core3D scan body as torque magnitude in-creased (Fig 11). Amann Girrbach and Sirona scan-body heights decreased as torque magnitude increased, indicating compression of scan bodies. Interestingly, this compression was not observed for the Core3D scan bodies; mean scan-body height decreased when torque application increased from 5 to 10 Ncm, but in-creased when torque application was increased from 10 to 15 Ncm. These findings hint at innate scan-body distortions that occur and highlight obvious differ-ences between scan-body brands. The contribution of the scan body toward overall accuracy of the acquired virtual models, therefore, should be taken into account.

In addition, two studies demonstrated that the ma-chining tolerance of implant components is measurable

and should be considered in all accuracy studies. Ma et al25 evaluated two generations of Nobel Biocare im-plant components, and found that tolerances ranged from 22 to 100 μm. Braian et al26 utilized external- and internal-hexagon Biomet 3i implants. Horizontal ma-chining tolerance on prefabricated gold abutments for external-hexagon implants was 44 ± 9 μm, while that for internal-hexagon implants was 58 ± 28 μm. When castable plastic abutments were used, horizon-tal tolerance for the external-hexagon implants was 12 ± 89 μm, while that for internal-hexagon implants was 86 ± 47 μm.

It appears that the dental laboratory scanner sys-tems studied are appropriate for the derivation of implant models in the fully edentulous patient. The conventional impression provides excellent 3D ac-curacy and precision, and errors attributable to the subsequent digital laboratory scanning step are minor. Dental laboratory scanner systems provide the option for a digital workflow, with the most important advan-tages being fabrication consistency and the possibility of using zirconia and other CAD/CAM materials.

True Definition exhibited the poorest accuracy and does not appear to be appropriate for the investigated clinical scenario. The overall scan procedure for True Definition was found to be more difficult to achieve compared with TRIOS, which may reflect limitations in system capability. Use of a powder spray for True Defini-tion may also contribute to errors, as it is impossible to evenly coat all areas of interest. TRIOS was not found to be significantly different from other test groups. How-ever, for dR, the results indicate slightly greater distor-tion of TRIOS compared with Impregum and the dental laboratory scanner systems. Intraoral scanner systems may still be utilized if there are clear clinical contraindi-cations to the use of the conventional impression sys-tem, such as patient intolerance toward conventional impression materials. In this situation, factors such as the presence of unique soft tissue morphology may in-crease the accuracy of intraoral scanning, and attempts should be made to compensate for noncharacteristic morphology. Lee27 has recently suggested that unique

5 10 15

140

120

100

80

60

40

20

0

Mea

n co

axia

lity

(µm

)

Torque magnitude (Ncm)

Fig 10 Mean coaxiality of scan bodies (µm) by system, varia-tion with torque magnitude. Error bars = SD.

Core 3DAmann GirrbachSirona

5 10 15

605040302010

0–10–20–30–40

Devi

atio

n fro

m n

omin

al

scan

bod

y he

ight

(µm

)

Torque magnitude (Ncm)

Fig 11 Deviation from nominal scan body height (µm) by sys-tem, variation with torque magnitude. Error bars = SD.

Core 3DAmann GirrbachSirona



Tan et al

shapes drawn on the palate with pressure-indicating paste may enhance the stitching of images.

LimitationsThe in vitro nature of this study eliminated a number of patient- and operator-related factors that could have affected results. All impressions (conventional or intra-oral scans) were made on master models with no sur-rounding saliva, soft tissue, and oral musculature, and thus, there was uninhibited access and visibility. While the relative ease of the impression step was equally ap-plicable to all systems in this study, the increase in dif-ficulty in vivo is unlikely to be of equal extent.

Factors such as patient comfort also could not be evaluated. Wismeijer et al28 suggested that intraoral scanners were preferred due to the unpleasant taste of the conventional impression material and the prepara-tion required for the conventional impression. Howev-er, patients of the same study perceived the duration of the intraoral scan more negatively. Moreover, some intraoral scanner systems such as the 3M True Defini-tion Scanner require use of a surface scanning spray, application of which increases time needed and dis-comfort experienced.

Parallel implants were chosen over angulated ones, which have been reported to affect scan accuracy.16 This allowed for a reduction of experimental variables in this study, thus focusing on the role of interimplant dis-tance on impression system accuracy. However, clinical-ly, it is likely that implants in the maxilla are angulated, and this should be a variable to be further investigated.

The time required to master each implant impression system may vary within the same operator and among different operators. Therefore, no limitations were im-posed on the number of tries allowed for each impres-sion system in this study. The single operator in charge of performing all conventional impressions, intraoral scans, and dental laboratory scanner scans underwent training provided by the respective suppliers, and was guided by other operators more experienced and familiar with each impression system. Conventional impressions were only accepted if they were of reasonable quality, in that there was proper fusion of impression material, complete capture of the area of interest, and absence of gross irregularities, bubbles, and separation of impres-sion material from the tray. The digital impressions were evaluated using different criteria. Virtual models were checked using the respective instantaneous previews to ensure that scan images were properly stitched. Further, the reference scan bodies had to be captured in their en-tirety with good detail reproduction; scans with incom-plete or distorted scan bodies were rejected. The results of this study reflect the accuracy of implant impression systems based on impressions or scans that would have been accepted by any operator, regardless of familiarity

or experience, using widely accepted and reasonable criteria. Operator familiarity thus did not pose a true limi-tation to this study.

Centroid position determination depended on the accuracy of measured geometric features, and there-fore, strict tolerance criteria were implemented. It may be argued that this level of scanning accuracy is unnec-essary, as the various software are still able to determine implant position via a “best-fit” algorithm. Unfortunately, the accuracy of the algorithms present in each software cannot be evaluated, as the proprietary data manipula-tion steps are not reported by any manufacturer.

The present study simulated six-implant and eight-implant situations in the edentulous maxilla. These two scenarios were chosen for their relatively common indication in completely edentulous patients request-ing reconstruction via fixed prostheses. Future studies evaluating an alternative number of implants (eg, two or four implants) in the edentulous maxilla will further aid clinicians who wish to attempt digital impressions for the fabrication of a removable prosthesis assisted by two or more implants, or for the fabrication of an All-on-4 fixed prosthesis.

CONCLUSIONS

This in vitro study examined the effect of implant im-pression system and interimplant distance on the 3D accuracy of implant positions in the resultant stone or virtual models. Within the limitations of this study, the following conclusions were made. Impregum consis-tently exhibited the lowest or second-lowest dR at all implant locations. True Definition exhibited the poorest accuracy for all linear distortions (dx, dy, dz, and dR) and 3D reference distance distortion (ΔR) parameters in both models A and B, but not for absolute angular distortions (Absdθx and Absdθy). Excluding True Definition, there was no significant difference among the remaining five impression systems for linear distortion parameters (dx, dy, dz, and dR) in both models A and B. The six impression systems could not be consistently ranked for angular distortions. Reducing interimplant distance may de-crease global linear distortions (dR) for intraoral scanner systems, but had no effect on Impregum and the dental laboratory scanner systems.

ACKNOWLEDGMENTS

The authors thank Mr Chan Swee Heng and Mr Tok Wee Wah for technical assistance. The authors also thank Strau-mann, Core3D Centres, 3Shape, 3M ESPE, Amann Girrbach, and Sirona for providing system-specific training and sup-port. The authors reported no conflicts of interest related to this study.



Tan et al

REFERENCES

1. Hämmerle CH, Cordaro L, van Assche N, et al. Digital technologies to support planning, treatment, and fabrication processes and outcome assessments in implant dentistry. Summary and consen-sus statements. The 4th EAO consensus conference 2015. Clin Oral Implants Res 2015(suppl 11);26:97–101.

2. Grant GT, Campbell SD, Masri RM, Andersen MR, American College of Prosthodontists Digital Dentistry Glossary Development Task Force. Glossary of digital dental terms: American College of Prosth-odontists. J Prosthodont 2016;25(suppl 2):s2–s9.

3. Lee H, So JS, Hochstedler JL, Ercoli C. The accuracy of im-plant impressions: A systematic review. J Prosthet Dent 2008;100:285–291.

4. Wee AG. Comparison of impression materials for direct multi-implant impressions. J Prosthet Dent 2000;83:323–331.

5. Hoods-Moonsammy VJ, Owen P, Howes DG. A comparison of the accuracy of polyether, polyvinyl siloxane, and plaster impressions for long-span implant-supported prostheses. Int J Prosthodont 2014;27:433–438.

6. Gherlone E, Capparé P, Vinci R, Ferrini F, Gastaldi G, Crespi R. Con-ventional versus digital impressions for “all-on-four” restorations. Int J Oral Maxillofac Implants 2016;31:324–330.

7. Papaspyridakos P, Gallucci GO, Chen CJ, Hanssen S, Naert I, Van-denberghe B. Digital versus conventional implant impressions for edentulous patients: Accuracy outcomes. Clin Oral Implants Res 2016;27:465–472.

8. Vandeweghe S, Vervack V, Dierens M, De Bruyn H. Accuracy of digital impressions of multiple dental implants: An in vitro study. Clin Oral Implants Res 2017;28:648–653.

9. 3Shape. 2013. 3Shape TRIOS Color Digital Impression Solution. Available at: https://www.3shape.com/en/knowledge-center/news-and-press/press-releases/2013/3shape-launches-trios-colour. (Accessed 18 March 2018).

10. 3M. 2015. 3M True Definition Scanner Technical Specifications. Available at: http://multimedia.3m.com/mws/media/919861O/3m-true-definition-scanner-technical-specifications.pdf. (Accessed 10 January 2017).

11. Amann Girrbach. 2017. Amann Girrbach - Products - Ceramill Map400. Available at: https://www.amanngirrbach.com/en/prod-ucts/scan/ceramill-map400/. (Accessed 10 January 2017).

12. Dental Compare. 2016. Ceramill Map400 Scanner from AmannGir-rbach GmbH. Available at: http://www.dentalcompare.com/4723-Dental-Laboratory-3D-Scanning-Systems/4459992-Ceramill-Map-400-Scanner/. (Accessed 10 January 2017).

13. Sirona. 2013. The inEos X5: A multi-talent for the dental lab. Available at: http://www.sirona.com/en/news-events/news-press/digital-dentistry-news-detail/27118/. (Accessed 10 Janu-ary 2017).

14. Dental Equipment Store. 2017. 3Shape D900 Series. Available at: http://www.dentalequipmentcenter.com/3shape-dental.html. (Accessed 10 January 2017).

15. Chew AA, Esguerra RJ, Teoh KH, Wong KM, Ng SD, Tan KB. Three-dimensional accuracy of digital implant impressions: Effects of different scanners and implant level. Int J Oral Maxillofac Implants 2017;32:70–80.

16. Chia VA, Esguerra RJ, Teoh KH, Teo JW, Wong KM, Tan KB. In vitro three-dimensional accuracy of digital implant impressions: The effect of implant angulation. Int J Oral Maxillofac Implants 2017;32:313–321.

17. Teo JW, Tan KB, Nicholls JI, Wong KM, Uy J. Three-dimensional accuracy of plastic transfer impression copings for three implant systems. Int J Oral Maxillofac Implants 2014;29:577–584.

18. Chandran DT, Jagger DC, Jagger RG, Barbour ME. Two- and three-dimensional accuracy of dental impression materials: Effects of storage time and moisture contamination. Biomed Mater Eng 2010;20:243–249.

19. Del Corso M, Abà G, Vazquez L, Dargaud J, Dohan Ehrenfest DM. Optical three-dimensional scanning acquisition of the position of osseointegrated implants: An in vitro study to determine method accuracy and operational feasibility. Clin Implant Dent Relat Res 2009;11:214–221.

20. Baig MR, Tan KB, Nicholls JI. Evaluation of the marginal fit of a zirconia ceramic computer-aided machined (CAM) crown system. J Prosthet Dent 2010;104:216–227.

21. Ender A, Mehl A. Accuracy of complete-arch dental impressions: A new method of measuring trueness and precision. J Prosthet Dent 2013;109:121–128.

22. Tan KB, Rubenstein JE, Nicholls JI, Yuodelis RA. Three-dimensional analysis of the casting accuracy of one-piece, osseointegrated implant-retained prostheses. Int J Prosthodont 1993;6:346–363.

23. Örtorp A, Jemt T, Bäck T, Jälevik T. Comparisons of precision of fit between cast and CNC-milled titanium implant frameworks for the edentulous mandible. Int J Prosthodont 2003;16:194–200.

24. Stimmelmayr M, Güth JF, Erdelt K, Edelhoff D, Beuer F. Digital evaluation of the reproducibility of implant scanbody fit—an in vitro study. Clin Oral Investig 2012;16:851–856.

25. Ma T, Nicholls JI, Rubenstein JE. Tolerance measurements of various implant components. Int J Oral Maxillofac Implants 1997;12:371–375.

26. Braian M, De Bruyn H, Fransson H, Christersson C, Wennerberg A. Tolerance measurements on internal- and external-hexagon implants. Int J Oral Maxillofac Implants 2014;29:846–852.

27. Lee JH. Improved digital impressions of edentulous areas. J Pros-thet Dent 2017;117:448–449.

28. Wismeijer D, Mans R, van Genuchten M, Reijers HA. Patients’ pref-erences when comparing analogue implant impressions using a polyether impression material versus digital impressions (Intraoral Scan) of dental implants. Clin Oral Implants Res 2014;25:1113–1118.


Documents

Comparison of Three-Dimensional Accuracy of Digital and