Content uploaded by Junyu Shi
Author content
All content in this area was uploaded by Junyu Shi on Jun 17, 2022
Content may be subject to copyright.
Clin Oral Impl Res. 2022;00:1–12.
|
1wileyonlinelibrary.com/journal/clr
1 | INTRODUCTIO N
Immediate implant placement has received increasing attention due
to the shor tened treatment period and predictable prognosis, which
has the potential to improve patient satisfaction (Kan et al., 2018;
Shi et al., 2015). Placement, however, utilizes mainly the socket and
the alveolar bone plate slope; bleeding from the socket interferes
with proper visualization of the implant site, thus making the sur-
gery more difficult and affecting the accuracy of implant placement
(Blanco, Carral, Argibay, & Liñares, 2019).
Suboptimal immediate implant position may result in restorative
and esthetic challenges, and may compromise the longevity of the
Received: 15 April 2022
|
Revised: 19 May 2022
|
Accepted: 29 May 2022
DOI : 10.1111/clr.13961
ORIGINAL ARTICLE
Does machine- vision- assisted dynamic navigation improve the
accuracy of digitally planned prosthetically guided immediate
implant placement? A randomized controlled trial
Shi- Min Wei1 | Yuan Li1 | Ke Deng1 | Hong- Chang Lai1 | Maurizio S. Tonetti1,2 |
Jun- Yu Shi1
© 2022 John Wiley & Sons A /S. Publishe d by John Wiley & S ons Ltd.
Shimin We i and Yuan Li auth ors are equall y contributed t o this work.
1Shanghai PerioImplant Innovation Center
and Department of Oral and Maxillo-
Facial Implantology, Shanghai Ninth
People's H ospital, C ollege of Stomatology,
Shanghai Jiao Tong University School
of Medicine, National Clinical Research
Center for Oral Disea ses, Shang hai Key
Laboratory of Stomatology & Shanghai
Research Institute of Stomatolog y,
Shanghai, China
2European Research Group on
Periodontology, Genova, Italy
Correspondence
Jun- Yu Shi and Maurizio S. Tonetti,
Depar tment of Oral and Maxillo- facial
Implantology, Shanghai Ninth People's
Hospital, School of Medicine, Shanghai
Jiaotong U niversit y, 639 Zhizaoju Road,
Shanghai 200011, China.
Emails: sjy0511@hotmail.com; maurizio.
tonetti@ergoperio.eu
Funding information
Clinical Research Program of Ninth
People’s Hospital affiliated Shanghai
Jiao Tong University School of Medicine,
Grant/Award Number: JYLJ201909;
Shanghai Clinical Research Center for
Oral Diseases, Grant/Award Number:
19411950100
Abstract
Objectives: This randomized controlled clinical trial was designed to compare the ac-
curacy of machine- vision (MV)- based dynamic navigation (DN)- assisted immediate
implant placement with the conventional freehand technique.
Material and methods: A total of 24 subjects requiring immediate implant placement
in maxillary anterior teeth were randomly assigned to either the control (freehand by
an experienced surgeon, n = 12) or the test group (MV- DN, n = 12). Implant platform,
implant apex, angular, and depth deviations with respect to prosthetically guided digi-
tal planning and differences in implant insertion torque (ITV) and implant stability
quotient (ISQ) were compared between the groups.
Results: MV- DN resulted in more accurate immediate implant position: significantly
smaller global platform deviation (1.01 ± 0.41 mm vs. 1.51 ± 0.67 mm, p = .038), plat-
form depth deviation (0.44 ± 0.46 mm vs. 0.95 ± 0.68 mm, p = .045), global apex
deviation (0.88 ± 0.43 mm vs. 1.94 ± 0.86 mm, p = .001), and lateral apex deviation
(0.68 ± 0.30 mm vs. 1.61 ± 0.88 mm, p = .004) were found in MV- DN compared to
controls. No significant intergroup differences were observed for ITV and ISQ.
Conclusions: MV- DN achieved more precise immediate implant position and compa-
rable primary stability. Further trials are necessary to assess the benefits in terms of
esthetics and tissue health/stability.
KEY WORDS
dental implants, dynamic navigation, free hand, immediate implant placement, machine vision,
randomized controlled clinical trial
2
|
WEI e t al.
implant (Tahmaseb, Wismeijer, Coucke, & Derksen, 2014). A previous
consensus report has recommended that immediate implant place-
ment should be utilized by experienced surgeons (Feine et al., 2018).
Hence, this technique- sensitive procedure requiring meticulous
execution requires additional development before its adoption by
less experienced clinicians. As such, more advanced technology
should be introduced to facilitate the immediate implant planning
and placement for surgeons. Static guides have been introduced but
poor flexibility and increased cost have limited their routine clinical
application (Joda, Derksen, Wittneben, & Kuehl, 2018).
In recent years, machine- vision (MV) enhanced and artificial
intelligence- assisted dynamic navigation (DN), a technolog y that
already has a large number of applications in industr y, is gradu-
ally being applied to image- guided and minimally invasive surgi-
cal approaches and holds great promise for safety and accuracy
(Kalfas, 2021). Thanks to the continuous development of infrared-
guided binocular stereo vision positioning technology, MV- assisted
DN systems enable visualization of the relationship between the
drills and the key anatomical structures in real- time, potentially re-
sulting in more accurate and predictable dental implant placement
(D'Haese, Ackhurst, Wismeijer, De Bruyn, & Tahmaseb, 2017; Wei
et al., 2021). A split- mouth randomized clinical trial (RCT ) evaluated
the accuracy of dynamic navigation compared with the unassisted
freehand method and concluded that utilizing a computer- guided
technique increased accuracy significantly (Aydemir & Arısan, 2020).
Although few pilot studies have introduced digital flow integrating
dynamic navigation for immediate implant surgery (Kuo, Lin, Hung,
Chiu, & Kuo, 2022; Pozzi, Arcuri, Carosi, Nardi, & Kan, 2021), there is
a paucity of adequately powered clinical trials to elucidate the effi-
cacy of MV- DN- assisted immediate implant placement.
The hypothesis of this RCT was that the global platform discrep-
ancy between digitally planned prosthetically guided implant po-
sition and actual position could be decreased by applying MV- DN
compared to freehand placement by an experienced surgeon.
2 | MATERIALS AND METHODS
2.1 | Experimental design
This was a double- blind, randomized, controlled, clinical trial to
compare the discrepancy between digitally planned prosthetically
guided implant position and actual position following MV- DN or
freehand immediate implant position. The primar y outcome was
platform deviation between the planned and actual implant position.
Secondary outcomes included apex deviation, angular deviation, im-
plant insertion torque (ITV) and implant stability quotient (ISQ), and
implant survival rate.
The study protocol was approved by the ethics committee of
Shanghai Ninth People's Hospital, School of Medicine, Shanghai
Jiao Tong University (approval number: SH9H- 2020- T122- 2) and
prospec tively registered on Clini calTr ials.gov (NCT04999956). The
study was performed in compliance with the Helsinki Declaration,
and written informed consent was obtained from all subjects prior
to the implementation of the study.
Consecutive subjects scheduled to receive immediate implant
placement were invited to participate. The study was conducted at
the Shanghai Perio- Implant Innovation Center, Department of Oral
and Maxillo- facial Implantology of Shanghai Ninth People's Hospital
between January 2021 and December 2021. The CONSORT flow-
chart is presented in Figure 1.
2.2 | Subject criteria
One calibrated investigator (SMW) was responsible for patients
screening and enrollment if patients fulfilled the following inclusion
and exclusion criteria.
The inclusion criteria were as follows:
(i) ≥18 years old and in good health;
(ii) the maxillary incisor that cannot be retained due to non-
periodontitis, including fracture, endodontic failure, and root
resorption;
(iii) the buccal bone plate is complete (via CBCT and bone sounding);
(iv) no acute infection;
(v) the extraction socket has at least 3– 5 mm apical bone.
The exclusion criteria were as follows:
(i) general contraindications of oral implant surgery (such as immu-
nodeficiency, long- term use of corticosteroids);
(ii) treatments or diseases that may affect bone tissue metabo-
lism (for example, taking bisphosphonates or receiving local
radiotherapy);
(iii) periodontitis history or uncontrolled periodontitis.
(iv) heavy smokers or previous heavy smoking histor y (quit smoking
time <5 years or >20 cigarettes per day);
(v) refuse to participate in this trial.
2.3 | Calculation of sample size
A commercial software (GPower version 3.1, Dusseldorf, Germany)
was used to estimate the required sample size with 80% of study
power and significant level (α) of 0.05. Based on the accuracy of dy-
namic navigation and the freehand method reported in a previous
randomized clinical trial (Aydemir & Arısan, 2020; 1.01 ± 0.07 mm
vs. 1.70 ± 0.13 mm), 4 subject per group are required to detect a
clinically important difference of global platform deviation with 80%
power and alpha set at 0.05. To compensate for potential loss to
follow- up or incomplete observations, this study recruited 12 sub-
jects per group.
|
3
WEI et al.
2.4 | Control of bias
Subjects were randomized based on a computer- generated list with
a permuted block of four (SAS v9.4; SAS Institute Inc., Car y, NC,
USA). Allocation was concealed with opaque envelopes that were
opened after tooth extraction and before immediate implant inser-
tion. The examiners and the patients were blind to allocation until
completion of the study. A single experienced implantologist (JYS)
performed all surgical procedures. While clinician blinding was not
possible at the surgery time, it was possible during the data collec-
tion and analysis stages.
2.5 | Prosthetically guided digital implant planning
A registration device (u- shaped tube with silicon nitride markers) was
fixed onto the patient's implant site with a silicone elastomer (DMG
Chemisch- Pharmazeutische, Hamburg, Germany) before CBCT scan-
ing. The implant position was digitally planned superimposing Dicom
data (i- CAT, Imaging Sciences International, Hatfield, PA, USA) and STL
data (TRIOS® 3 W, 3Shape, Copenhagen, Denmark) using a dedicated
software (Dcarer Dental Implant Navigation System, Suzhou, China).
In cases in which the crown was still in place at baseline, the implant
crown was designed according to the original crown. When only the
residual root was present, the crown was designed according to the
morphology of the adjacent and contralateral teeth.
Implant planning was prosthetically guided according to the fol-
lowing parameters:
(i) the buccal and lingual emergence angle of the crown less than
60°;
(ii) implant at the center of the predetermined mesio- distal width
of the final restoration with a minimal distance of 2 mm from the
adjacent teeth;
FIGURE 1 Consolidated standards of reporting trails (CONSORT) flowchart
4
|
WEI e t al.
(iii) neck of the implant 0.5 to 1 mm below the palatal bone crest and
4 mm apical to the target soft tissue level;
(iv) implant placed along the palatal wall of the extraction socket;
(v) space (buccal gap) of at least 2 mm between the buccal bone wall
and the implant;
(vi) at the cervical level, the long axis of the implant protrudes from
the cingulum of the designsed crown if possible.
2.6 | Extraction and control implant placement
After anesthesia, the periodontal fiber was cut by a tunneling knife
(Hu- Friedy, Chicago, USA), and the target tooth was extracted
carefully using a flapless approach to avoid the buccal plate bone
damage. The alveolus was carefully curetted and irrigated with
saline. The osteotomy site was prepared according to the manu-
facturer’s standard protocol and a tapered implant was installed
attempting to reproduce the digitally planned position (12– 14 mm
length, 3.3 mm shoulder diameter, narrow CrossFit NC, bone
level; Straumann, Waldenburg, Switzerland). A 4 mm healing cap
was placed. Given that ITV less than 15 N/cm, cover screw was
placed. The buccal gap was filled with Bio- Oss Collagen (Geistlisch
Pharmaceutical, Wolhusen, Switzerland). Then, gelatin sponges
(Jinling Pharmaceutical, Jiangsu, China) were placed to cover the
alveolus and the wounds were sutured using figure- eight suture
with an absorbable suture (Vycryl Plus 4– 0; Ethicon, Sommerville,
USA) (Figure 2).
2.7 | Experimental intervention: Machine vision-
assisted dynamic navigation
Before surgery, calibration (Figure 3a) and registration (Figure 3b)
were performed. Calibration determined the relationship between
the surgical handpiece and the patient tracking ar ray and allowed the
computer to identify the relationship between drill position and real
anatomical structures. Under accurate tracking by stereo cameras,
handpiece with long and short ball drills were placed on tracking
array successively for calibration. Registration was determined to
match real anatomical structures with CBCT image via u- shaped
tube. The tracking array was fixed on the posterior teeth of max-
illa by bis- acryl composite resins (DMG Chemisch- Pharmazeutische,
Hamburg, Germany). The short ball drill was placed on six fiducial
markers in u- shaped tube pits in turn for registration. The u- shaped
tube was removed and surgeon placed the drill tip on the tooth ana-
tomical markers to confirm the position was correct.
A binocular stereo vision- based navigator (Dcarer, Suzhou,
China) was used. In brief, two cameras were calibrated to obtain the
relative positional relationship of the osteotomy drills, the surgical
area and overlay it on the digital planning. Determination of the
global position of the drills within the digital planning was based on
the epipolar constraints algorithm. Osteotomy site preparation and
implant insertion were performed looking at the screen (Figure 3c),
and other procedures were performed looking at the patient mouth
as the same to control implant placement (Figure 2).
2.8 | Outcome assessment
The actual implant position was assessed 1 week after surgery with a
CBCT (i- CAT, Imaging Sciences International, Hatfield, PA, USA). The
position was compared to the digital plan using the Accuracy analy-
sis of the Dcarer dynamic navigation software by two independent
calibrated examiners who were blind to the group allocation (YL and
KD). The inter- examiner agreement was analyzed by intra- class cor-
relation coefficient (ICC) for global platform deviation.
The measurement in software were proceeded as follow
(Figure 4):
(i) at least four feature point s (such as tooth cusps or bone pits)
were selected in preoperative and postoperative CBCT for rough
registration.
(ii) feature surface circles in preoperative CBCT were selected, and
then a mathematical algorithm displayed a similar feature surface
circle in postoperative CBCT. The algorithm registered 1000 of
FIGURE 2 Surgical procedure (a)
before surgery operation; (b) the target
tooth was extracted carefully using a
flapless approach to assure the buccal
plate was complete. (c) an implant was
installed palatally, and the jumping gap
was filled with bio- Oss collagen; (d) gelatin
sponges were placed on the top, and the
wounds were sutured with resorbable
sutures
(a) (b)
(c) (d)
|
5
WEI et al.
points in these two circles via conventional iso- surface thresh-
olding technology.
(iii) the planned and actual implant were identified, and the deviation
was be automatically calculated.
The following parameters were measured (Figure 5):
• global platform deviation
• lateral platform deviation
• platform depth deviation
• vector platform depth deviation
• global apex deviation
• lateral apex deviation
• apex depth deviation
• vector apex depth deviation
• angular deviation
Insertion torque value (ITV) and implant stability quotient (ISQ)
were measured as primar y stabilit y. After the implant was installed
using dynamic navigation, the ITV were evaluated by a dynamometric
wrench (Straumann, Waldenburg, Swit zerland). Then, a Smartpeg was
screwed into each implant, and resonance frequency analysis (RFA)
was performed using Osstell Mentor (Osstell/Integration Diagnostics,
Goteborg, Sweden). ISQ were recorded in the buccal and the palatal
directions three times and averaged.
2.9 | Statistical analysis
The statistical analyses were performed by one independent inves-
tigator who was blind to the group allocation (YL). Statistical analy-
sis was conducted using SPSS v22 (SPSS Inc., Chicago, USA). After
verification of the normality assumption with the Shapiro– Wilk test,
FIGURE 3 Dynamic navigation (a) calibration; (b) registration; (c) drilling and implant procedures looking at the screen
(a)
(c)
(b)
6
|
WEI e t al.
the equality of variance was assessed with the F test. Significance
of intergroup differences for continuous variables was assessed with
the Student's two sample t tests when distribution was normal and
variances were equal. The Welch two sample t tests were performed
when distribution was normal but variances were not equal, and
Mann– Whitney U tests were performed when distribution was not
normal. Chi- Square test was per formed to test group differences for
categorial variables. The significance level α was set at .05 for all tests.
3 | RESULTS
3.1 | Implant survival
A total of 24 patients were recruited and randomly allocated into
either treatment group. All patients completed the study (7 male pa-
tients and 5 female patients in the control group, 8 male patients and
4 female patients in the test group). The total study population had
a mean ± SD age of 38 ± 16 y (control group, 40 ± 17 y; test group,
36 ± 15 y). When crown delivery, the survival rate was 100% in both
groups. Basic information was displayed in Table 1.
3.2 | Accuracy
Table 2 and Figures 6, 7 showed the platform, apex, and angular de-
viations between planned and actual impl ant s for two groups. Most
accuracy- related outcomes were in favor of computer- assisted im-
mediate implant surgery. The control group had almost double the
deviation in platform depth deviation, global apex deviation, lateral
apex deviation, and angular deviation observed in the experimen-
tal group (0.95 ± 0.68 mm vs. 0.44 ± 0.46 mm, 1.94 ± 0.86 mm vs.
0.88 ± 0.43 mm, 1.61 ± 0.88 mm vs. 0.68 ± 0.30 mm, 5.97 ± 5.37°
vs. 2.51 ± 1.50°). Significantly smaller platform deviations in global
FIGURE 4 Accuracy measurement in software (a- b) rough registration: Four feature points were selected both in preoperative and in
postoperative CBCT; (c- d) precise registration: Six feature surface circles in preoperative CBCT were selected, and then a mathematical
algorithm displayed similar feature surface circles in postoperative CBCT. The algorithm registered 1000 of points in these two circles via
conventional iso- surface thresholding technolog y; (e- f ) accuracy measurement: Preoperative and postoperative CBCT were overlapped. The
planned and actual implant were identified, and the deviation was automatically calculated
(a) (b)
(c) (d)
(e) (f)
|
7
WEI et al.
platform deviation, platform depth deviation, global apex deviation,
and lateral apex deviation were found in the MV- DN group com-
pared to the controls (p = .038, .045, .001 and .004, respectively).
According to vector depth, platform depth deviation in MV- DN
group was more significantly apical than that in control group
(p = .004).
Average intraclass correlation coefficient (ICC) for global plat-
form deviation was 0.96 (95% CI 0.90– 0.98). The measurement of
two examiners showed excellent reliability.
3.3 | Primary stability
Table 3 showed the primar y stabilit y for the two groups. No signifi-
cant differences in ISQ immediately after implant placement was
found between the two groups (control group: 59.07 ± 13.41 vs. test
group: 59.67 ± 10.61, p > .05). No significant difference in ITV dis-
tribution (≥35, 15– 35 and ≤15 N/cm) was found between the two
groups (p > .05).
4 | DISCUSSION
The main findings of the present clinical trial were that MV- DN-
assisted immediate implant placement significantly reduced the
discrepancy with the digitally planned prosthetic- driven position
compared with implants placed with freehand surgery by an expe-
rience surgeon, and the observed increase in accuracy of implant
position was obtained without compromise in the ability to achieve
primary implant stability.
Specifically, MV- DN group showed significantly better accuracy
in regards of global platform deviation (95% CI 0.03– 0.98 mm), plat-
form depth deviation (95% CI 0.01– 1.05 mm), global apex deviation
(95% CI 0.47– 1.65 mm) and lateral apex deviation (95% CI 0.36–
1.51 mm). The level of observed discrepancy in MV- DN assisted im-
mediately placed implants were in line with previous comparisons
in model study (Chen et al., 2018) and in healed ridges (Aydemir &
Arısan, 2020). This indicates that MV- DN could effectively improve
the accuracy of surgical implant placement, even in technically chal-
lenging immediate cases. In industry, MV technology utilizes ma-
chines instead of the human eye to locate, judge, and guide, which
has the advantages of accurate positioning, reduced labor costs,
and long- lasting working time (Guo et al., 2021). When performing
image- guided and minimally invasive surgeries, it is particularly im-
portant for the MV- assisted DN system to provide accurate global
spatial positioning and real- time human– machine interaction feed-
back. In this context, the employed active infrared- guided binocular
stereo vision positioning technology showed promising results for
navigated implant placement. More research and development are
required to further optimize the work flow and improved the human–
computer interaction.
Decomposing the deviation may better understand implant
placement in fresh sockets. Global deviation is a comprehensive
indicator combined of lateral deviation and depth deviation and
angular deviation is a synthesis of global platform deviation and
global apex deviation (Figure 5). The present study showed the con-
ventional freehand method mainly had unsatisfactory accuracy in
FIGURE 5 Accuracy measurement
TABLE 1 Demographics and clinical data of patients
Group Freehand
Dynamic
navigation
Number 12 12
Age (year)
Mean ± SD 40 ± 17 36 ± 15
M i n – M a x 18– 69 1 8 – 6 7
Gender
Male 7 (58.3) 8 (66.7)
Female 5 (41.7) 4 (33.3)
Incisor
Central 10 (83 .4) 8 (66.7)
Lateral 1 (8.3) 4 (33.3)
Canine 1 (8.3) 0 (0.0)
Reason for extraction
Crown/root Fracture 6 (50.0) 5 (41.7)
Endodontic failure 6 (50.0) 7 (58.3)
Root resorption 0 (0.0) 0 (0.0)
SRP classification
Class I 11 (91.67) 9 (75.00)
Class II 1 (8.33) 2 (16.67)
Class III 0 (0.0) 0 (0.0)
Class IV 0 (0.0) 1 (8.33)
Socket (mm)
Width (mean ± SD) 6.03 ± 0.64 6.05 ± 0.63
Length (mean ± SD) 10.3 ± 0.99 9. 8 ± 1.26
Implant length
12 mm 8 (66.7) 8 (66.7)
14 mm 4 (33.3) 4 (33.3)
Note: Data are presented as either mean ± SD or n (%).
Abbreviations: SD, st andard deviation; SRP, sagitt al root position
classification.
8
|
WEI e t al.
platform depth and lateral apex, eventually resulting in higher global
platform deviation, global apex deviation and angular deviation. That
can explain higher platform depth error in control group there is a
visual error in eye observation on the palatal alveolar ridge due to
flapless method and blood filling, probably causing more deviation
(Arisan, Karabuda, Mumcu, & Özdemir, 2013). More significant lat-
eral apex deviation may because the apex position and direction
cannot be observed (Brief, Edinger, Hassfeld, & Eggers, 2005; Chen,
Yuh, et al., 2018; Kaewsiri, Panmekiate, Subbalekha, Mattheos, &
Pimkhaokham, 2019). Of note, the average depth deviation in con-
trol group was more coronal and that in DN was more apical. The
numeric discrepancy is that visual errors from top to bottom may
cause the implant to be slightly shallower than expected. When the
screen displays warning signs in navigation, human beings need a
few seconds to react before stopping (Thompson et al., 1992), re-
sulting in a slightly deeper position.
Interestingly, in this study, global apex deviation was smaller than
the global platform deviation (0.88 ± 0.43 mm vs. 1.01 ± 0.41 mm).
Our result was in contrast to the both recent static navigation meta-
analysis data in healed ridges (1.33 mm; 95% CI: 1.17– 1.50 mm vs.
1.03 mm; 95% CI: 0.88– 1.18 mm) (Siqueira et al., 2020), a cadaver
study in fresh sockets (0.93 ± 0.34 mm vs. 0.85 ± 0.38 mm; Chen
TABLE 2 Accuracy measurements
Group Freehand
Dynamic
navigation p Value
95% CI for
differences
Platform deviation
(mm)
Global Mean ± SD 1 .51 ± 0.67 1.01 ± 0.41 .038 (0.03, 0.98)
Median (IQR) 1.64 (0.87, 2.04) 0.94 (0.71, 1.26)
M i n – M a x 0. 34– 2.5 4 0 . 4 9 – 1 . 7 5
Lateral Mean ± SD 1.02 ± 0. 59 0.83 ± 0.33 .514 (−0.23, 0.70)
Median (IQR) 0.93 (0.56, 1.54) 0.74 (0.64, 0.94)
M i n – M a x 0.09– 2.00 0 . 4 8 – 1 . 7
Depth Mean ± SD 0.95 ± 0.68 0.44 ± 0.46 .0 45 (0.01, 1.05)
Median (IQR) 0.87 (0.30, 1.47) 0.33 (0.07, 0.63)
M i n – M a x 0 . 1 7 – 2 . 3 7 0.04– 1.62
Vector depth Mean ± SD −0 .61 ± 1.02 0.43 ± 0.48 .004 (−1 .73,
−0.34)
Median (IQR) −0.52 (−1.47, −0.15) 0.33 (0.07, 0.63)
M i n – M a x −2.37 to 1.21 −0.05- 1.62
Apex deviation (mm) Global Mean ± SD 1.94 ± 0.86 0.88 ± 0.43 .001 (0.47, 1.65)
Median (IQR) 1.89 (1.50, 2.57) 0.87 (0.48, 1.17)
M i n – M a x 0.38– 3.39 0 . 2 9 – 1 . 7 8
Lateral Mean ± SD 1.61 ± 0.88 0.68 ± 0.30 .004 (0.36 , 1.51)
Median (IQR) 1.78 (0.74, 2.14) 0.70 (0.40, 0.90)
M i n – M a x 0.36– 3.38 0. 21– 1. 13
Depth Mean ± SD 0.83 ± 0.69 0.45 ± 0.57 .248 (−0.15, 1.02)
Median (IQR) 0.76 (0.166, 1.45) 0.34 (0.08, 0.63)
M i n – M a x 0.09– 2.07 0.02– 1.62
Vector depth Mean ± SD −0.26 ± 1.08 0.44 ± 0.48 .0 51 (−1.43, 0.02)
Median (IQR) −0.20 (−1.28, 0.64) 0.34 (0.08, 0.63)
M i n – M a x −2.07 to 1.36 −0.04 to 1.62
Angular deviation (°) Mean ± SD 5.97 ± 5.37 2.51 ± 1.50 .078 (−0.13, 5.50)
Median (IQR) 4.06 (2.3, 10.54) 2.30 (1.39, 3.65)
M i n – M a x 0 . 2 7 – 1 7 . 5 9 (0 .45– 5 .29)
Note: Data are presented as mean ± SD, median (Interquartile range, IQR), Min– Max.
Vector depth: a positive number means actual implants is more apical than expected; vice versa. CI, confidence intervals; SD, standard deviation;
IQR, Interquartile range.
To test group dif ferences for continuous variables, Student's t wo sample t te sts were performed when distribution was normal and variances were
equal (global platform deviation), Welch two sample t test were performed when distribution was normal but variances were not equal (vector
platform depth deviation, global apex deviation, lateral apex deviation, vector apex depth deviation), and Mann– Whitney U tests were performed
when distribution was not normal (lateral platform deviation, platform depth deviation, apex depth deviation, and angular deviation). Bold values
denote st atistical significance at the p < 0.05 level.
|
9
WEI et al.
et al., 2018), and a model study in fresh sockets (1.19 ± 0.35 mm vs.
0.74 vs. 0.15 mm; Chen et al., 2022). A possible explanation for this
observation could be that the osteotomy site of immediately placed
implants would show a slightly buccal shift due to the corticalization
of the palatal bone wall of the alveolus but template guide could
not control the apex (Wei et al., 2022). Another explanation may be
that when navigating in real time, the deviation of the apical position
displayed on the screen will receive more attention to obtain better
accuracy. Hence, dynamic navigation technology may allow better
control of apex lateral parameter than templates in immediate im-
plant placement. Indeed a potential benefit of dynamic navigation
compared to the use of a stat ic guide is allowing th e surgeon a degree
of flexibility in making small, yet potentially significant corrections in
FIGURE 6 Box plots of platform deviation and apex deviation. To test group differences for continuous variables, Student's two sample
t tests were performed for global platform deviation, Welch two sample t tests were used for vector platform depth deviation, global apex
deviation, lateral apex deviation, vector apex depth deviation, and Mann– Whitney U tests were performed for lateral platform deviation,
platform depth deviation, apex depth deviation, angular deviation. *p < .05
FIGURE 7 Box plot of angular deviation. Mann– Whitney U test
was performed to test the group difference (angular deviation).
*p < .05
TABLE 3 Primary stability
Group Freehand
Dynamic
navigation p Value
ISQ 59.07 ± 13.41 59.67 ± 10. 61 >.05
ITV ≤15 N/cm 2 (16.67%) 1 (8.33%) >.05
15– 35 N/cm 4 (33.33%) 5 (41.67%)
≥35 N/cm 6 (50%) 6 (50%)
Note: Data are presented as either mean ± SD or n (%).
Student 's two sample t test was per formed to te st group dif ferences for
continuous variables (ISQ). Chi- Square test was performed to test group
differences for categorial variables (ITV).
Abbreviations: ISQ, implant stability quotient; IT V, insertion torque
value.
10
|
WEI e t al.
implant placement (Wismeijer et al., 2018). Nonetheless, much more
powered clinical trials should be conducted to elucidate this issue.
The leading five steps in dynamic navigation workflow include
image processing, virtual implant position design, calibration, reg-
istration, and surgery operation. The errors may occur at each step
and accumulate and influence the final result.
CBCT is the primary method to acquire navigation images. It
has been confirmed to gain higher accuracy than conventional
multi- slice computed tomography (MSCT) due to its small voxel
size (Tao et al., 2020). In CBCT, layer thickness, voxel size, thresh-
old segmentation process, patients' movement, metal restoration,
and operator's proficiency, all these factors influence accuracy
(Wei et al., 2021). Dental magnetic resonance imaging (dMRI) is
also a promising imaging technology due to its nonionizing tech-
nique, which is currently used to create static guides and achieves
comparable accuracy to CBCT (Flügge et al., 2020). However,
dMRI tooth surface reconstruction accuracy is still relatively low
(Hilgenfeld et al., 2019), while it is crucial for dynamic navigation.
Limited by technology, no one has reported this technology to dy-
namic navigation yet. More technologies and algorithms should be
improved in this regard.
Calibration and registration are also influencing factors of ac-
curacy. Bone landmark registration (titanium screws) is the gold
standard while it is invasive (Ramezanzade et al., 2021). Occlusal
splint registration and U- shaped tube registration are more conve-
nient methods, but they still increase the chair time (Jorba- Garcia,
Figueiredo, Gonzalez- Barnadas, Camps- Font, & Valmaseda-
Castellon, 2019). In the scientific literature, the accuracy of bone
landmark registration (0.720 ± 0.322 mm, 1.168 ± 0.313 mm,
3.103 ± 1.019°), occlusal splint registration (1.0 ± 0.07 mm,
1.83 ± 0.12 mm and 5.59 ± 0.39°), and U- shaped tube registration
(1.36 ± 0.65 mm, 1.48 ± 0.65 mm, and 3.71 ± 1.32°) are comparable
(Aydemir & Arısan, 2020; Stefanelli et al., 2020; Wu et al., 2020),
which are in accordance with our results (1.01 ± 0.41 mm,
0.88 ± 0.43 mm and 2.51 ± 1.50°). Different commercially available
navigation systems suggest registration of anatomical landmarks
to reduce the calibration time, that remains a signific ant barrier
to adoption. Recently, two main approaches have been repor ted:
(i) Navident fuses the intraoral scanning data with the CBCT data;
while (ii) X- Guide uses the X- Mark algorithm to improve the accu-
racy of recognition of sur face anatomical structures. However, no
studies have investigated the accuracy of these emerging registra-
tion modalities.
Previous research has shown that the surgeon's proficiency in
navigation affects accuracy (Golob Deeb et al., 2019). The use of
navigation requires the unnatural learned skill of performing the sur-
gery while looking at the screen rather than the operating field; a
learning curve is necessary (Block, Emery, Lank, & Ryan, 2016). In
this study, all procedures were performed by a surgeon experienced
in surgical navigation procedures. No information is available about
the effect of the specific experience on the results.
Another error worth mentioning, but not routinely reported, is
the process of deviation measurement. When performing deviation
analysis, it is necessar y to overlap the preoperative CBCT and the
postoperative CBCT. This process requires selecting several same
marker points in two CBCT images and has technical sensitivity.
The software automatically calculates the overlapping error based
on the difference between the three- dimensional positions of the
points on the coordinate system. In our study, two trained inves-
tigators with excellent intra- examiner reproducibility completed
this process.
The present study shows several shor tcomings. First, the lim-
ited sample size is insufficient to assess potential predictors or
confounders. In addition, the long- term clinical, radiographic and
esthetic benefits of the increased accuracy of implant placement
with MV- DN needs to be fully assessed. Finally, the application
of MV- DN systems may increase the operation time, affect the
patient's experience, and increase the patient's operation cost.
Therefore, the operation time, patient- centered outcomes, and
socioeconomic benefits should be reported in future, larger, mul-
ticenter studies.
5 | CONCLUSION
This study reported proof of principle evidence that MV- DN in-
creased the accuracy of global implant placement in immediate im-
plant placement compared with a freehand approach. Furthermore,
the level of accuracy obtained is comparable to the one obtained
with static or dynamically guided implant placement in healed
ridges, thereby supporting the potential applicability of the tech-
nology to overcome the specific surgic al challenges of immediate
implant placement. Further studies are necessary to confirm gener-
alizability, clinical relevance and cost- ef fectiveness of the obser ved
benefits.
AUTHOR CONTRIBUTIONS
Conception and design: S.M. Wei, J.Y. Shi; Performing the surger-
ies: J.Y. Shi; Data collection and analysis: Y. Li, K . Deng; Manuscript
preparation: S.M. Wei, Y. Li, J.Y. Shi; Reviewing and editing draft: K.
Deng, H.C. Lai, M.S. Tonetti, J.Y. Shi; All authors critically reviewed
the manuscript and approved the final manuscript for publication.
CONFLICT OF INTERESTS
None of the authors has a financial interest in any of the products,
devices, or drugs mentioned in this manuscript.
DATA AVAIL ABILI TY STATEMENT
The data that support the findings of this study are available from
the corresponding author upon reasonable request.
ORCID
Hong- Chang Lai https://orcid.org/0000-0002-0133-7516
Ju n- Yu Sh i https://orcid.org/0000-0001-7482-1766
|
11
WEI et al.
REFERENCES
Arisan, V., Karabuda, C. Z., Mumcu, E., & Özdemir, T. (2013). Implant po-
sitioning errors in freehand and computer- aided placement meth-
ods: A single- blind clinical comparative study. The International
Journal of Oral & Maxillofacial Implant s, 28(1), 190– 204. h t t p s :// do i .
org/10.11607/ jomi.2691
Aydemir, C. A., & Arısan, V. (2020). Accuracy of dental implant placement
via dynamic navigation or the freehand method: A split- mouth
randomized controlled clinical trial. Clinical Oral Impla nts Research,
31(3), 255– 263. h ttp s://doi. org /10.1111/clr.13563
Blanco, J., Carral, C ., Argibay, O., & Liñares, A. (2019). Implant placement
in fresh extraction sockets. Periodontology 2000, 79(1), 151– 167.
https://doi .org /10.1111/prd.1 2253
Block, M., Emery, R., Lank, K., & Ryan, J. (2016). Implant placement ac-
curacy using dynamic navigation. The International Journal of Oral
& Maxillofacial Implants, 32(1), 92– 99. https://doi.org/10.11607/
jomi.5004
Brief, J., Edinger, D., Hassfeld, S., & Eggers, G. (2005). Accuracy of image-
guided implantology. Clinical Oral Implants Research, 16(4), 495–
501. https://doi.org/10.1111/j.1600- 0501.2005.01133.x
Chen, Z., Li, J., Ceolin Meneghetti, P., Galli, M., Mendonça, G., & Wang,
H. L. (2022). Does guided level (fully or partially) influence implant
placement accuracy at post- extraction sockets and healed sites? An
in vitro study. Clinical Oral Investigations. https://doi.org/10.1007/
s 0 0 7 8 4 - 0 2 2 - 0 4 5 1 2 - y
Chen, Z., Li, J., Sinjab, K., Mendonca, G., Yu, H., & Wang, H. L. (2018).
Accuracy of flapless immediate implant placement in anterior max-
illa using computer- assisted versus freehand surgery: A cadaver
study. Clinical Oral Implants Research, 29(12), 1186– 1194. h t t p s: //
doi .org/10.1111/clr.133 82
Chen, C . K., Yuh, D. Y., Huang, R . Y., Fu, E., Tsai, C. F., & Chiang, C. Y.
(2018). Accuracy of implant placement with a navigation system,
a laboratory guide, and freehand drilling. The International Journal
of Oral & Maxillofacial Implants, 33(6), 1213– 1218. h t t p s :// do i .
org/10.11607/ jomi.6585
D'Haese, J., Ackhurst, J., Wismeijer, D., De Bruyn, H., & Tahmaseb, A .
(2017). Current state of the art of computer- guided implant sur-
ger y. Periodontology 2000, 73(1), 121– 133. https: //doi .org/10.1111/
prd.12175
Feine, J., Abou- Ayash, S., al Mardini, M., de Santana, R. B., Bjelke-
Holtermann, T., Bornstein, M. M., Braegger, U., Cao, O., Cordaro, L.,
Eycken, D., Fillion, M., Gebran, G., Huynh- Ba, G., Joda, T., Levine,
R., Mattheos, N., Oates, T. W., Abd- Ul- Salam, H., Santosa, R.,
… Zubiria, J. P. V. (2018). Group 3 ITI consensus report: Patient-
reported outcome measures associated with implant dentistry.
Clinical Oral Implant s Research, 29(Suppl 16), 270– 275. h t t p s: //d o i .
org /10.1111/clr.1329 9
Flügge, T., Ludwig, U., Winter, G., Amrein, P., Kernen, F., & Nelson, K.
(2020). Fully guided implant surgery using magnetic resonance im-
aging - an in vitro study on accuracy in human mandibles. Clinical
Oral Implants Research, 31(8), 737– 746. https://doi .org /10.1111/
clr.13622
Golob Deeb, J., Bencharit, S., Carrico, C. K., Lukic, M., Hawkins, D.,
Rener- Sitar, K., & Deeb, G. R. (2019). Exploring training dental im-
plant placement using computer- guided implant navigation system
for predoctoral students: A pilot study. European Journal of Dental
Education: official journal of the Association for Dental Education in
Europe, 23 (4), 415– 423. ht tps ://doi.org/10.1111/eje.12447
Guo, N., Zhao, Q., Fan, H., Yu, H., Cui, X., Lv, M., & Li, T. (2021). Binocular
stereo vision calibration based on accurate ellipse detection algo-
rithm of direct calculation and gr ating conversion check. Optik, 242,
16 626 9. https://doi.org/10.1016/j.ijleo.2021.166269
Hilgenfeld, T., Juerchott, A., Deisenhofer, U. K., Weber, D., Rues, S.,
Rammelsberg, P., Heiland, S., Bendszus, M., & Schwindling, F. S.
(2019). In vivo accuracy of tooth surface reconstruction based on
CBCT and dental MRI- A clinical pilot study. Clinical Oral Implants
Research, 30(9), 920– 927. htt ps://doi .org/10.1111/clr.13498
Joda, T., Derksen, W., Wittneben, J. G ., & Kuehl, S. (2018). Static
computer- aided implant surgery (s- CAIS) analysing patient-
reported outcome measures (PROMs), economics and surgical
complications: A systematic review. Clinical Oral Implants Research,
29(Suppl 16), 359– 373. htt ps://doi.org /10.1111/clr.13136
Jorba- Garcia, A ., Figueiredo, R., Gonzalez- Barnadas, A., C amps- Font, O.,
& Valmaseda- Castellon, E. (2019). Accuracy and the role of expe-
rience in dynamic computer guided dental implant surgery: An in-
vitro study. Medicina Oral Patologia Oral Y Cirugia Bucal, 24 (1), E76–
E83. https://doi.org/10.4317/medor al.22785
Kaewsiri, D., Panmekiate, S., Subbalekha, K., Mattheos, N., &
Pimkhaokham, A. (2019). The accuracy of static vs. dynamic
computer- assisted implant surgery in single tooth space: A random-
ized controlled trial. Clinical Oral Implants Research, 30(6), 505– 514.
https://doi .org /10.1111/clr.134 35
Kalfas, I. H. (2021). Machine vision navigation in spine surgery. Front
Surg, 8, 640554. https://doi.org/10.3389/fsurg.2021.640554
Kan, J. Y. K., Rungcharassaeng, K., Deflorian, M., Weinstein, T., Wang,
H.- L., & Testori, T. (2018). Immediate implant placement and pro-
visionalization of maxillary anterior single implants. Periodontology
2000, 77(1), 197– 212. https://doi.org /10.1111/prd.1 2212
Kuo, P. J., Lin, C. Y., Hung, T. F., Chiu, H. C., & Kuo, H. Y. (2022). A novel
application of dynamic guided navigation system in immediate im-
plant placement. Journal of Dental Sciences, 17(1), 354– 360. h t t p s ://
doi.org/10.1016/j.jds.2021.09.030
Pozzi, A., Arcuri, L ., Carosi, P., Nardi, A ., & Kan, J. (2021). Clinical and
radiological outcomes of novel digital workflow and dynamic nav-
igation for single- implant immediate loading in aesthetic zone: 1-
year prospective case series. Clinical Oral Implants Research, 32(12) ,
1 3 9 7 – 1 4 1 0 . ht tps: //doi .org/10.1111 /clr.138 39
Ramezanzade, S., Keyhan, S. O., Tuminelli, F. J., Fallahi, H. R., Yousefi, P.,
& Lopez- Lopez, J. (2021). Dynamic- assisted navigational system in
zygomatic implant surgery: A qualitative and quantitative system-
atic review of current clinical and cadaver studies. Journal of Oral
and Maxillofacial Surgery, 79(4), 799– 812. ht tps://doi.org/10.1016/j.
joms.2020.10.009
Shi, J. Y., Wang, R., Zhuang, L . F., Gu, Y. X., Qiao, S. C., & Lai, H. C. (2015).
Esthetic outcome of single implant crowns following type 1 and
type 3 implant placement: A systematic review. Clinica l Oral Implant s
Research, 26(7), 768– 774. htt ps://doi.or g/10.1111 /clr.123 34
Siqueira, R., Chen, Z., Galli, M., Saleh, I., Wang, H. L., & Chan, H. L. (2020).
Does a fully digital workflow improve the accuracy of computer-
ass ist ed im pla nt surgery in partially edentulous patients? A sys tem -
atic review of clinical trials. Clinical Implant Dentistry and Related
Research, 22(6), 660– 671. ht tps://doi.org/10.1111/cid.12937
Stefanelli, L. V., Mandelaris, G. A ., Franchina, A., Pranno, N., Pagliarulo,
M., Cera, F., Maltese, F., Angelis, F., & Carlo, S. D. (2020). Accuracy
of dynamic navigation system workflow for implant supported full
arch pros thesis. A Case S eries. International Journal of Environmental
Research and Public Health, 17(14), 5038. https://doi.org/10.3390/
i j e r p h 1 7 1 4 5 0 3 8
Tahmaseb, A., Wismeijer, D., Coucke, W., & Derksen, W. (2014).
Computer technology applications in surgical implant dentistry: A
systematic review. The International Journal of Oral & Ma xillofacial
Implants, 29(Suppl), 25– 42. https://doi.org/10.11607/ jomi.2014s
uppl.g1.2
Tao, B., Shen, Y., Sun, Y., Huang, W., Wang, F., & Wu, Y. (2020).
Comparative accuracy of cone- beam CT and conventional multis-
lice computed tomography for real- time navigation in zygomatic
implant surgery. Clinical Implant Dentistry and Related Research,
22(6), 747– 755. https://doi .org/10.1111/cid.12958
Thompson, P. D., Colebatch, J. G ., Brown, P., Rothwell, J. C., Day, B.
L., Obeso, J. A., & Marsden, C. D. (1992). Voluntary stimulus-
sensitive jerks and jumps mimicking myoclonus or pathological
12
|
WEI e t al.
star tle syndromes. Movement Disorders, 7(3), 257– 262. ht t p s : //d o i.
org/10.1002/mds.87007 0312
Wei, S. M., Shi, J. Y., Qiao, S. C., Zhang, X., Lai, H. C., & Zhang, X. M.
(2022). Accuracy and primary st ability of t apered or straight im-
plants placed into fresh extraction socket using dynamic navigation:
A randomized controlled clinical trial. Clinical Oral Investigations,
26(3), 2733– 2741. h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / s 0 0 7 8 4 - 0 2 1 - 0 4 2 4 7 - 2
Wei, S. M., Zhu, Y., Wei, J. X., Zhang, C. N., Shi, J. Y., & Lai, H. C. (2021).
Accurac y of dynamic navigation in implant surger y: A systematic
review and meta- analysis. Clinical Oral Implants Research, 32(4),
383– 393. https://doi.org/10.1111/clr.13719
Wismeijer, D., Joda, T., Flügge, T., Fokas, G., Tahmaseb, A ., Bechelli, D.,
Bohner, L., Bornstein, M., Burgoyne, A., Caram, S., Carmichael, R.,
Chen, C . Y., Coucke, W., Derksen, W., Donos, N., el Kholy, K., Evans,
C., Fehmer, V., Fickl, S., … Wu, V. (2018). Group 5 ITI consensus re-
port: Digital technologies. Clinical Oral Implants Research, 29(Suppl
16), 436– 442. ht tps ://doi.o rg/10.1111/clr.13309
Wu, D., Zhou , L., Yang, J., Zhang, B ., Lin, Y., Chen, J., Hua ng, W., & Chen, Y.
(2020). Accuracy of dynamic navigation compared to static surgical
guide for dental implant placement. International Journal of Implant
Dentistry, 6(1), 78. h t t p s : / / d o i . o r g / 1 0 . 1 1 8 6 / s 4 0 7 2 9 - 0 2 0 - 0 0 2 7 2 - 0
SUPPORTING INFORMATION
Additional supporting information may be found in the online
version of the article at the publisher ’s website.
How to cite this article: Wei, S- M, Li, Y., Deng, K., Lai, H- C,
Tonetti, M. S., & Shi, J- Y (2022). Does machine- vision-
assisted dynamic navigation improve the accuracy of digitally
planned prosthetically guided immediate implant placement?
A randomized controlled trial. Clinical Oral Implants Research,
00, 1– 12. https://doi. org /10.1111/clr.13961
