Computers in Biology and Medicine 43 (2013) 16531660

Contents lists available at ScienceDirect

Computers in Biology and Medicine

journal homepage: www.elsevier.com/locate/cbm

Factors affecting accuracy of implant placement with

mucosa-supported stereolithographic surgical guides in
edentulous mandibles
Megumi Ochi a, Manabu Kanazawa a,n, Daisuke Sato b, Shohei Kasugai b, Shigezo Hirano a,
Shunsuke Minakuchi a
a
b

Complete Denture Prosthodontics, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, Tokyo, Japan
Oral Implantology and Regenerative Dental Medicine, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, Tokyo, Japan

art ic l e i nf o

a b s t r a c t

Article history:
Received 30 March 2013
Accepted 28 July 2013

This study aimed to evaluate the accuracy of implant placement with mucosa-supported surgical guides
in edentulous mandibles and to determine the factors affecting accuracy. Implant placement was
simulated on the preoperative CT image and mucosa-supported surgical guides were fabricated for six
edentulous mandible models and 15 patients with edentulous mandibles, using CAD/CAM technology.
Deviations of the actual implant position from the planned position were calculated by comparing the
planned image and the postoperative image. Based on the results, it was concluded that mucosasupported surgical guides have high accuracy and that bone density and mucosal thickness could affect
accuracy.
& 2013 Elsevier Ltd. All rights reserved.

Keywords:
Dental implant
Edentulous mandible
Accuracy
Mucosa-supported surgical guide
Stereolithographic
Computed tomography

1. Introduction
Stereolithographic surgical guides can simplify the techniquesensitive and operator-dependent surgical procedures in implantsupported restorations, beneting both patient and dentist [1].
When stereolithographic surgical guides were used, variations in
accuracy of implant placement within surgeons and between
surgeons were reduced, compared to surgery using conventional
guides [2]. Shorter surgery duration and less discomfort after
surgery were recorded when mucosa-supported surgical guides
were used [3]. Therefore, the use of mucosa-supported stereolithographic surgical guides in edentulous patients will increase
along with the demand for implant-supported restorations in
edentulous patients.
Although many researchers have evaluated the accuracy of
implant placement with mucosa-supported surgical guides [417],
few studies have assessed factors affecting the accuracy of
mucosa-supported surgical guides. Errors during implant placement have often been explained by instability of the surgical guide
[4]. Considering a report that alveolar ridge resorption was
associated with denture stability in edentulous mandibles [18],
alveolar ridge shape seems to inuence surgical guide accuracy,

n
Correspondence to: Complete Denture Prosthodontics, Graduate School
of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45,
Yushima, Bunkyo-ku, Tokyo 113-8549, Japan. Tel.: 81 3 5803 5563;
fax: 81 3 5803 5586.
E-mail address: m.kanazawa.ore@tmd.ac.jp (M. Kanazawa).

0010-4825/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.compbiomed.2013.07.029

which has not been examined. A previous study reported greater

accuracy in the edentulous mandible, which has higher bone
density, than in edentulous maxilla [5]. However, in another study,
the surgical guide showed higher accuracy in edentulous maxilla
because it covered a larger area than the edentulous mandible [6].
There are some reports that thicker supporting mucosa of edentulous maxilla might decrease accuracy [7,8]. Therefore, bone
density, mucosal thickness, and area of supporting mucosa seem
to affect the accuracy of implant placement with mucosasupported surgical guides in edentulous mandible. Previous
papers have not investigated the inuence of these three factors
together on accuracy.
In this paper, the authors performed a model study and a
clinical study, in which the accuracy of implant placement with
mucosa-supported surgical guides was evaluated by matching the
images of planned and placed implants and calculating the
positional deviations between them. The purpose of the model
study was to determine the system-related accuracy using edentulous mandible models and the inuence of alveolar ridge shape
on accuracy. The purpose of the clinical study was to evaluate the
procedure-related accuracy in patients with edentulous mandible
and determine the factors affecting accuracy. The null hypothesis
for the model study was that there would be no difference
between the accuracy of implant placement in two types of
models with different alveolar ridge shapes. The null hypothesis
for the clinical study was that there would be no correlation
between the accuracy and bone density, mucosal thickness, and
area of supporting mucosa.

1654

M. Ochi et al. / Computers in Biology and Medicine 43 (2013) 16531660

2. Materials and methods

2.1. Presurgical preparation
The study protocol was approved by the Institutional Ethics
Committee of Tokyo Medical and Dental University (#531). For the
model study, six edentulous mandible test models were fabricated,
which consisted of a stone cast (Zo-Dental Plaster, Shimomura
Gypsum, Saitama, Japan) and a 2-mm-thick layer of silicone
impression material (Examixne Regular Type, GC, Tokyo, Japan)
as articial mucosa. They were divided equally into two groups
according to the alveolar ridge shape (Absorbed group and Nonabsorbed group). The three models in each group were fabricated
identically. A radiographic guide was fabricated for each group
using multipurpose pourable resin (Procast DSP clear, GC) (Fig. 1a).
Eight gutta-percha markers (Temporary Stopping, GC) as radiopaque ducials were placed in the radiographic guide.
The clinical study population consisted of 15 patients with edentulous mandibles, who were scheduled to be treated with twoimplant overdenture at the University Hospital of Dentistry, Tokyo
Medical and Dental University, between July 2010 and January 2013.
Patients with systemic diseases, poor oral hygiene, uncontrolled
diabetes, current irradiation to the head or neck, psychological
disorders, or administration of bisphosphonates were excluded. All
patients were informed of the study protocol and signed an informed
consent form. The patient's complete denture with an optimal shape,
occlusion and mucosal t was used as a radiographic guide (Fig. 1b),
and eight gutta-percha markers were placed.
Preoperative planning and manufacturing of the surgical guide
(NobelGuide, Nobel Biocare, Gothenburg, Sweden) were performed
by the double scanning technique [19]. For every CT scan, conebeam CT (CBCT) (Finecube, Yoshida, Tokyo, Japan; voxel size

0.157 mm, slice thickness 0.146 mm, acquisition time 19 s, FOV

82 mm  75.1 mm, 90 kV, 4 mA) was used. In the model study,
the radiographic guide and the test model were connected using
rubber bands and were scanned for the rst CT scan. In the clinical
study, the patient was instructed to clench the radiographic guide
with maximal bite force and the rst CT scan was performed. The
radiographic guide alone was then scanned for the second CT scan.
After a virtual 3D image of the combined radiographic guide and
the model or bone was created by the planning software (Procera,
Nobel Biocare), two implants were virtually located at their optimal
position in the intraforaminal area (Fig. 2a and b). Individually
customized surgical guides which contained two guided sleeves
and three anchor pin sleeves were manufactured by rapid prototyping (Fig. 3a and b).
2.2. Surgical procedure
The implant placement in the model study was performed by a
prosthodontist with Z7 years of experience in implant-placement
surgery. During surgery, the surgical guide was attached to the model
using rubber bands. The implant placement in the clinical study was
performed under intravenous sedation and local anesthesia, by an
implantologist with Z12 years of experience in implant-placement
surgery at the University Hospital of Dentistry, Tokyo Medical and
Dental University. The local anesthetic injection was administered via
the sleeves, with the surgical guide strongly compressed against the
ridge so that mucosa swelled by the anesthetic would not disturb the
stability of the surgical guide. Until the implant insertion was over, the
surgical guide was kept in place by strongly compressing its bilateral
molar regions. The implant (Nobel Speedy Groovy RP 4  11.5
18 mm2, Nobel Biocare) was inserted according to the manufacturer's
protocol for a apless surgical procedure (Fig. 4a and b) [9].

Fig. 1. Radiographic guides for the (a) model study and (b) clinical study. (a) Radiographic guide (1) was fabricated on stone cast (2) with articial mucosa (3). (b) The
patient's complete denture made of acrylic resin was used as a radiographic guide, in which gutta-percha markers were placed as reference points (arrow).

Fig. 2. Planning on the Procera software in the (a) model study and (b) clinical study. Placement of two implants (1) and three anchor pins (2) was planned presurgically,
according to anatomy and prosthetic design.

M. Ochi et al. / Computers in Biology and Medicine 43 (2013) 16531660

1655

Fig. 3. Stereolithographic surgical guides for the (a) model study and (b) clinical study.

Fig. 4. Surgical procedure. During the operation, the surgical guide was stabilized on (a) the model or (b) the mucosa, using anchor pins.

Fig. 5. Matching procedure. (a) Positions of the planned and placed implants were compared by matching the planned (blue) and postoperative (red) images. (b) The x-axis was
dened as the line connecting the neck of the right (blue) and left (light blue) implant of the planned image, while the z-axis was dened as the central axis of the right implant. Each
deviation was calculated at the implant neck and apex. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

2.3. Postsurgical procedure

After surgery, the postoperative CT scan was performed. The model
was scanned with the surgical guide still xed on the model after the
surgery. The patient was scanned within 3 days after the surgery,
clenching the radiographic guide with maximal bite force. The resultant
DICOM data were converted into standard triangle language (n.stl) les
in 3D CAD software (Mimics, Materialise, Leuven, Belgium) and a 3D
postoperative image was obtained. Two implants were connected to
the surgical guide using specic pins (Guided Cylinder with Pin Unigrip
Branemark System, Nobel Biocare) through the guided sleeves. In this
combination, the positional relationship between the implants and the
surgical guide was regarded as equal to that planned in the planning
software. Therefore, the planned image was obtained by scanning the
combination. Planned image and postoperative image were then
superimposed with reference to the surface of radiographic guides
and surgical guides (Fig. 5a) with a matching method using an iterative
closest point algorithm. In this method, the apices of triangles that

compose the 3D objects in n.stl les were used as the closest points and
an algorithm that minimized the distance between corresponding
apices was repeated. For exclusion of outliers and accurate matching, a
distance threshold was set and decreased from 3 mm to 0.01 mm
during the repeated algorithm. Global deviation between the planned
and actually inserted implant positions was measured. Mesio-distal,
bucco-lingual, and depth deviation were also taken as measurement
parameters by setting the 3D space coordinates (Fig. 5b). Each
deviation was calculated at the implant neck and apex. When the
coordinates of the measurement sites were (x1, y1, z1) in the planned
image and (x2, y2, z2) in the postoperative image, each deviation was
calculated by the following formulas:
Global deviation mm fx2 x1 2 y2 y1 2 z2 z1 2 g1=2
Mesio  distal deviation mm x2 x1
Bucco  lingual deviation mm y2 y1
Depth deviation mm z2 z1

1656

M. Ochi et al. / Computers in Biology and Medicine 43 (2013) 16531660

Fig. 6. Measurement of bone density. On a cross-section in the planning software, ve equally spaced points (red points in the right image) were located along the major axis
of the planned implant site. The mean gray value at the points was calculated and regarded as bone density at the implant site. (For interpretation of the references to color
in this gure legend, the reader is referred to the web version of this article.)

Bone density was determined from the gray value obtained by

using a specic tool in the planning software for measurement of
density values [20]. The mean gray value at 5 equally spaced 2-mm
diameter points along the major axis of the planned implant site was
measured (Fig. 6). For evaluation of mucosal thickness, the distance
between the alveolar crest and the radiographic guide was measured
with the 3D CAD software at three points on the left and right rst
molars (Fig. 7). In the cross-sections at the mesial and distal contact
points and middle point of the rst molar, the minimum distance
between the surface of the alveolar crest and the base of the radiographic guide was measured at the midpoint of the buccal shelf.
The mean distance at these six points was dened as the mucosal
thickness. The area of tissue surface of the radiographic guide was
measured with design and meshing software (3-matic, Materialise),
which was regarded as the area of supporting mucosa (Fig. 8).

2.4. Statistical analysis

After the KolmogorovSmirnov test, the SteelDwass multiple
comparison procedure was performed to compare the global deviations at the implant neck and apex between the Absorbed group, Nonabsorbed group of the model study, and in the clinical study. In the
clinical study, correlations between each deviation (i.e., global, mesiodistal, bucco-lingual, and depth) and bone density, mucosal thickness,
and area of supporting mucosa were tested with the Spearman's rank
correlation coefcient. Correlations were tested between each deviation and the bone density at every implant site. Correlations were
tested between each deviation with higher absolute value of the two
implants in the same patient and the mucosal thickness and area of
supporting mucosa. Data were analyzed using the statistical analysis
software (JMP 10.0, SAS Institute, NC, USA; and Microsoft Excel 2011,
Microsoft Corporation, WA, USA). The level of signicance was set at
P 0.05.

Fig. 7. Measurement of mucosal thickness. On a cross-section at the rst molar of

the radiographic guide, the distance between the surface of the alveolar crest (red)
and the base of the radiographic guide (blue) was measured and regarded as the
mucosal thickness. (For interpretation of the references to color in this gure
legend, the reader is referred to the web version of this article.)

3. Results
Twelve implants were placed on six models, and 30 implants
were placed in 15 patients (seven males and eight females, mean
age 67.1 years) with edentulous mandibles. A total of 42 implants
were analyzed by comparing the preoperatively planned and actually
placed positions.
The global, mesio-distal, bucco-lingual, and depth deviations at the
implant neck and apex in the model study and the clinical study are
summarized in Table 1. Global deviations in the Absorbed group, Nonabsorbed group and the clinical study were 0.4170.11 mm, 0.337

M. Ochi et al. / Computers in Biology and Medicine 43 (2013) 16531660

0.09 mm and 0.8970.44 mm at the implant neck and 0.537

0.11 mm, 0.4570.12 mm and 1.0870.47 mm at the implant apex,
respectively. No statistically signicant difference was observed when
comparing global deviations in the Absorbed and Non-absorbed
groups of the model study. Global deviations at the implant neck
were signicantly higher in the clinical study than in the Absorbed
and Non-absorbed groups. Global deviations at the implant apex were
signicantly higher in the clinical study than in the Absorbed and
Non-absorbed groups.
Spearman's rank correlation coefcients between each deviation
and bone density, mucosal thickness, and area of supporting mucosa
are shown in Tables 24. There were signicant negative correlations
between bone density and depth deviations at the implant neck and
apex (Table 2, Figs. 9 and 10). Of the 30 implants, 21 were placed more
supercially and nine were placed more deeply compared to the
planned depths. There was a signicant positive correlation between
mucosal thickness and the global deviation at implant apex (Table 3
and Fig. 11). No statistically signicant correlation was found between
the area of supporting mucosa and any deviation (Table 4).

4. Discussion
In the model study, the present results suggest that the alveolar
ridge shape of the model does not inuence the accuracy, as no
signicant difference was observed in global deviations between the
Absorbed group and Non-absorbed group. When the alveolar ridge is
absorbed in an edentulous mandible, the surgical guide will be lifted
up by the mouth oor or moved horizontally by the tongue and buccal

Fig. 8. Measurement of the area of supporting mucosa. On the 3D image of the

radiographic guide, the border between tissue surface (green) and polished surface
(pink) was marked and the area of tissue surface was measured. (For interpretation
of the references to color in this gure legend, the reader is referred to the web
version of this article.)

1657

mucosa. In the model, the surgical guide is far more stable compared
to the actual surgical environment, as the model has no mouth oor,
tongue, or buccal mucosa that may cause instability of the guide.
Therefore, deviations that result from instability of the surgical guide
caused by ridge resorption may seldom occur on the model. Movement of the surgical guide related to the resilience of articial mucosa
[10,11] was minimized by using anchor pins and the same rubber
band that xed the radiographic guide during CT scan. Therefore, the
deviations in the model study were presumably system-related
deviations caused by errors during CT image acquisition, data processing, and manufacturing of the surgical guides using rapid prototyping
[21,22], and mechanical errors caused by the bur-sleeve gap [12]. The
error during image acquisition and data processing could be almost
0.5 mm [23]. Although the accuracy of rapid prototyping is reported to
range from 0.1 to 0.2 mm [13], incorrect setting in the software prior
to rapid prototyping can result in gross deformation of surgical guides
[22]. Horwitz et al. [24] suggested that deviations related to examiner
errors or errors during CT image acquisition ranged from 0.32 to
0.49 mm, which accounted for a major portion of whole deviations. In
the present study, errors during the matching procedure could occur
from incorrect threshold setting during n.stl conversion. Another factor
for examiner error would be the difference between the 3D images
of radiographic guides and surgical guides, which were used as
reference objects during matching. The mean global deviations in
the present model study were 0.3770.11 mm at the implant neck and
0.4970.12 mm at the apex, which were within the range of systemrelated deviations so far reported. Combining the ndings of Horwitz
et al. [24], if there were examiner errors during measurement of
deviations in the present study, the actual deviation would possibly be
around 00.97 mm at the neck and 01.1 mm at the apex. When the
system-related deviations compensate each other, the total deviation
will be minimized. However, as the deviations are theoretically
generated from the cumulative sum of all errors, it is important to
be aware of the largest deviation possible.
Signicantly higher deviations were observed in the clinical
study compared with the model study. Various procedure-related
factors, such as patients' movement during CT scan, instability of
the surgical guides while drilling, limited mouth opening, and
bone density, seem to have inuenced the accuracy and further
added to the system-related deviations.
Cassetta et al. suggested that higher bone density resulted in a
more supercial implant position when single-type guides with
depth control were used [14], and higher global deviations at the
implant neck and apex when multiple-type guides without depth
control were used [15]. In the present clinical study, single-type
guides with depth control were used and signicant negative
correlations between bone density and depth deviations at the
implant neck and apex were observed. Of the 30 implants placed,
21 were placed more supercially and nine were placed more

Table 1
Global, mesio-distal, bucco-lingual, and depth deviations in the model study and the clinical study (n42).
Deviations

Model study (n 12)

Clinical study (n30)

Absorbed group (n 6)

Neck (mm)

Apex (mm)

Global
Mesio-distal
Bucco-lingual
Depth
Global
Mesio-distal
Bucco-lingual
Depth

Non-absorbed group (n 6)

Max

Min

Mean

SD

Max

Min

Mean

SD

Max

Min

Mean

SD

0.59
0.22
0.08
 0.21
0.67
0.26
0.00
 0.20

0.29
 0.09
 0.28
 0.59
0.39
 0.46
 0.42
 0.59

0.41a
0.02
 0.08
 0.35
0.53c
 0.20
 0.21
 0.35

0.11
0.10
0.12
0.13
0.11
0.22
0.13
0.14

0.44
0.20
0.14
 0.20
0.63
0.56
0.36
 0.20

0.21
 0.09
 0.08
 0.42
0.32
 0.22
 0.31
 0.42

0.33a
0.04
0.02
 0.30
0.45c
0.03
0.04
 0.30

0.09
0.09
0.08
0.08
0.12
0.27
0.20
0.08

2.66
0.74
0.70
1.22
2.21
1.47
0.91
1.17

0.21
 0.40
 2.50
 1.16
0.20
 1.96
 1.92
 1.19

0.89b
0.14
 0.29
 0.28
1.08d
0.11
 0.29
 0.30

0.44
0.34
0.55
0.61
0.47
0.70
0.59
0.61

Global deviations of each row with a different superscript letters were signicantly different (p o 0.05).

1658

M. Ochi et al. / Computers in Biology and Medicine 43 (2013) 16531660

Table 2
Spearman's rank correlation coefcients between bone density and each deviation (n 30).
Deviation at the neck

Bone density

r
95% CI
p

Deviation at the apex

Global

Mesio-distal

Bucco-lingual

Depth

Global

Mesio-distal

Bucco-lingual

Depth

0.10
 0.27, 0.44
0.62

0.37
0.01, 0.64
0.05

0.03
 0.33, 0.39
0.88

 0.57
 0.77,  0.26
o0.01

0.07
 0.30, 0.42
0.72

0.37
0.01, 0.64
0.06

0.18
 0.19, 0.51
0.37

 0.56
 0.77,  0.25
o 0.01

rSpearman's rank correlation coefcients; 95% CI95% condence interval; and pp-value.
Table 3
Spearman's rank correlation coefcients between mucosal thickness and each deviation (n 15).
Deviation at the neck

Mucosal thickness

r
95% CI
p

Deviation at the apex

Global

Mesio-distal

Bucco-lingual

Depth

Global

Mesio-distal

Bucco-lingual

Depth

0.01
 0.51, 0.52
0.96

 0.05
 0.55, 0.47
0.85

0.08
 0.45, 0.57
0.79

0.21
 0.34, 0.65
0.44

0.73
0.35, 0.90
o0.01

0.08
 0.45, 0.57
0.77

0.16
 0.38, 0.62
0.56

0.21
 0.34, 0.65
0.44

rSpearman's rank correlation coefcients; 95% CI95% condence interval; and pp-value.
Table 4
Spearman's rank correlation coefcients between area of the supporting mucosa and each deviation (n 15).
Deviation at the neck

r
95% CI
p

Area

Deviation at the apex

Global

Mesio-distal

Bucco-lingual

Depth

Global

Mesio-distal

Bucco-lingual

Depth

0.20
 0.35, 0.65
0.47

0.06
 0.47, 0.56
0.82

0.14
 0.40, 0.61
0.63

0.42
 0.12, 0.77
0.12

0.01
 0.51, 0.52
0.97

0.06
 0.47, 0.56
0.83

 0.15
 0.62, 0.39
0.59

0.42
 0.12, 0.77
0.12

rSpearman's rank correlation coefcients; 95% CI95% condence interval; and pp-value.

Depth deviation at the apex (mm)

Depth deviation at the neck (mm)

1.5
1
0.5
0
-0.5
-1
-1.5

200

400

600

800

1000

1200

Bone density

1.5
1
0.5
0
-0.5
-1
-1.5

200

400

600

800

1000

1200

Bone density

Fig. 9. Correlation between bone density and depth deviation at the implant neck
(n30). The line is an approximate line.

Fig. 10. Correlation between bone density and depth deviation at the implant apex
(n 30). The line is an approximate line.

deeply, although there was depth control of the drills. In this

clinical study, when the insertion torque value was insufcient, the
implant was torqued manually after the surgical guide was
removed, until it achieved a sufcient torque level. The implant
might have been inserted deeper than planned due to this
procedure. Based on the results of the previous studies [14,15]
and the present study, it can be assumed that when surgical
guides are used, the implants tend to be placed more supercially
than planned and they come close to the planned position when
inserted more deeply at bone sites with lower bone density.
A signicant positive correlation between mucosal thickness
and the global deviation at the implant apex implied that an
increased mucosal thickness led to higher global deviation at the
implant apex. Thick mucosa may increase the freedom of

movement of the guide and lead to positional discrepancy

between the radiographic guide during CT scan and surgical guide
during surgery. In the clinical study by D'haese et al. [7], higher
deviations were observed in patients with thicker mucosa. Vasak
et al. [8] reported a positive correlation between buccal mucosal
thickness and the bucco-lingual deviation in patients with edentulous
maxilla. These previous studies focused on mucosa-supported
surgical guides for edentulous maxillae. The present study suggests that mucosal thickness also affects the deviation in edentulous mandibles.
In edentulous cadavers, greater deviations were found in the
mandible than in the maxilla [6]. The authors of the paper
explained that the surgical guide in the mandible was less stable

Global deviation at the apex (mm)

M. Ochi et al. / Computers in Biology and Medicine 43 (2013) 16531660

the results differed between the two studies because different

reference objects were used. Although this means the reference
object affects the results, many papers do not provide a detailed
account of the reference object used.
Accessibility is probably also an important factor affecting
accuracy of surgical guides. In this study, accessibility was not
taken into account because the drilling and insertion of implants
were performed in the intraforaminal area. Future studies are
required to determine the accuracy of implant placement in molar
areas of edentulous mandible with mucosa-supported surgical
guides, and to identify more factors affecting accuracy.

2.50
2.00
1.50
1.00
0.50
0.00
1.00

1659

1.50

2.00

2.50

3.00

Mucosal thickness (mm)

Fig. 11. Correlation between mucosal thickness and global deviation at the implant
apex (n 15). The line is an approximate line.

because it covered a smaller area. However, in the present clinical

study, no statistically signicant correlation was found between
the area of supporting mucosa and any deviation. The reason for
this could be that, except for the guided sleeves and anchor pin
sleeves, the surgical guide was an exact replica of the complete
denture fabricated by an experienced prosthodontist. Because the
surgical guide had an optimal shape, occlusion and mucosal t,
difference in area of supporting mucosa did not affect the deviation. In addition, the difference in area among mandibles was
much smaller than the difference between maxilla and mandible.
In this study, bone density and mucosal thickness seemed to be
the procedure-related factors that affected the accuracy of
mucosa-supported stereolithographic surgical guides in edentulous mandibles. These factors should be taken into account when
the clinicians decide the position of implants presurgically.
There were three more factors specic to the present study that
were not directly responsible for the accuracy of the surgical guide
but may have affected the results. First, in the clinical study, the
way of compressing the guide onto the ridge during CT scan and
during surgery was different. During CT scan, the patient clenched
the radiographic guide with maximal bite force. During the
surgery, although an occlusal index should be used according to
the manufacturer's protocol, it was not possible for the patient to
bite the index in this study because the surgery was performed
under intravenous sedation. Therefore, the assistant stabilized the
surgical guide by strongly compressing the bilateral molar regions
during the surgery. This could have made a difference in magnitude and direction of force on the guides. Secondly, torquing the
implant manually after the surgical guide was removed could have
introduced an additional depth deviation. However, as sufcient
insertion torque value is essential for attaining initial stability of
the implant, this process was unavoidable. Finally, superimposing
the planned image and postoperative image with reference to the
surface of radiographic guide and surgical guide could have
resulted in errors in the matching procedure, because the positional relationship between the implant and radiographic guide
was not rigid due to the presence of mucosa between them. Even if
the surgical guide moved during the operation, the actual deviation could not be detected when the radiographic guide moved
similarly during the postoperative CT scanning. Errors in the
matching procedure were reduced when the images were superimposed on the basis of the bone [9]. However, because end users
were not allowed to convert the planning data, including the bone,
to 3D data, the planned image was obtained by scanning the
combination of surgical guide and implants in the present study.
In the previous studies assessing accuracy, various reference
objects have been applied for matching the preoperative and
postoperative images [9,16,17]. Pettersson et al. [9] and Komiyama
et al. [16] analyzed the same clinical cases to assess accuracy, but

5. Conclusions
Within the limitations of this study, the following conclusions
can be drawn:
1. System-related deviations are slight even if errors occur during
measurement of deviations.
2. Implants are more likely to be placed deeply at bone sites with
lower bone density. In the surgery using surgical guides, the
implants tend to be placed more supercially compared to the
planned position.
3. The global deviations at the implant apex are more likely to
increase when the mucosa is thicker.
From these ndings, it can be concluded that mucosa-supported
surgical guides have high accuracy and that factors such as bone
density and mucosal thickness could affect accuracy.

6. Summary
Stereolithographic surgical guides manufactured by CAD/CAM
technology enable transfer of the software plan to the surgical eld
more accurately. Though the accuracy of implant placement with
mucosa-supported surgical guides has been reported previously, few
studies have assessed factors affecting accuracy. This study aimed to
evaluate the accuracy of implant placement with mucosa-supported
surgical guides in edentulous mandibles and determine the factors
affecting accuracy.
Six edentulous mandible models with articial mucosa were
fabricated for the model study and divided into two groups
according to the alveolar ridge shape (Absorbed group and Nonabsorbed group). Fifteen patients with edentulous mandibles were
enrolled in the clinical study. A preoperative CT scan was taken
and mucosa-supported surgical guides were prepared by virtual
treatment planning in which two implants were located in the
intraforaminal area. After the implants were inserted using the
surgical guides, a postoperative CT scan was taken and the planned
image and postoperative image were superimposed with a matching method using an iterative closest point algorithm. The global,
mesio-distal, bucco-lingual, and depth deviations at implant neck
and apex were determined and compared between the groups
with the SteelDwass multiple comparisons. In the clinical study,
correlations between each deviation and bone density, mucosal
thickness, and area of supporting mucosa were tested with the
Spearman's rank correlation coefcient.
Global deviations in the Absorbed group, Non-absorbed group and
the clinical study were 0.4170.11 mm, 0.3370.09 mm and 0.897
0.44 at the implant neck and 0.5370.11 mm, 0.4570.12 mm and
1.0870.47 mm at the implant apex respectively. No signicant
difference was observed in any deviation when comparing the
Absorbed and Non-absorbed groups of the model study. Global
deviations at the implant neck and apex were signicantly higher in

1660

M. Ochi et al. / Computers in Biology and Medicine 43 (2013) 16531660

the clinical study than in the Absorbed and Non-absorbed groups.

Signicant correlations were observed between bone density and
depth deviations and between mucosal thickness and global deviation
at the implant apex. No signicant correlation was found between the
area of supporting mucosa and any deviation.
Deviations observed in the model study seemed to be systemrelated deviations, which were regarded as clinically negligible. In the
clinical study, procedure-related factors seem to have inuenced
accuracy and added further to the system-related deviations. It was
suggested that bone density at the implant site and mucosal thickness
were the procedure-related factors. The implants are likely to be
placed more supercially than planned when the bone density at the
implant site is higher, and an increased mucosal thickness can lead to
higher global deviation at the implant apex. Bone density and mucosal
thickness, which seem to be procedure-related factors that affect the
accuracy of mucosa-supported stereolithographic surgical guides in
edentulous mandibles, should be taken into account when clinicians
decide the position of implants presurgically.
The following conclusions were made: (1) System-related
deviations are slight, (2) Bone density could affect depth deviations at the neck and apex, and (3) Mucosal thickness could affect
global deviation at the apex.
Conict of interest statement
None declared.
Acknowledgment
This study was supported by a Grant-in-Aid for Young Scientists
(No. 23792211) from the Ministry of Education, Culture, Sports,
Science and Technology of Japan.
References
[1] S.D. Ganz, Presurgical planning with CT-derived fabrication of surgical guides,
J. Oral Maxillofac. Surg. 63 (Suppl) (2005) 5971.
[2] D.P. Sarment, P. Sukovic, N. Clinthorne, Accuracy of implant placement with a
stereolithographic surgical guide, Int. J. Oral Maxillofac. Implants 18 (2003)
571577.
[3] V. Arisan, C.Z. Karabuda, T. Ozdemir, Implant surgery using bone- and mucosasupported stereolithographic guides in totally edentulous jaws: surgical and
post-operative outcomes of computer-aided vs. standard techniques, Clin. Oral
Implants Res. 21 (2010) 980988.
[4] N. van Assche, D. van Steenberghe, M.E. Guerrero, E. Hirsch, F. Schutyser,
M. Quirynen, R. Jacobs, Accuracy of implant placement based on pre-surgical
planning of three-dimensional cone-beam images: a pilot study, J. Clin.
Periodontol. 34 (2007) 816821.
[5] O. Ozan, K. Orhan, I. Turkyilmaz, Correlation between bone density and
angular deviation of implants placed using CT-generated surgical guides,
J. Craniofac. Surg. 22 (2011) 17551761.
[6] A. Pettersson, T. Kero, L. Gillot, B. Cannas, J. Fldt, R. Sderberg, Accuracy of
CAD/CAM-guided surgical template implant surgery on human cadavers: Part
I, J. Prosthet. Dent. 103 (2010) 334342.
[7] J. D'haese, H. De Bruyn, Effect of smoking habits on accuracy of implant
placement using mucosally supported stereolithographic surgical guides, Clin.
Implant Dent. Relat. Res. 15 (2013) 402411.
[8] C. Vasak, G. Watzak, A. Gahleitner, G. Strbac, M. Schemper, W. Zechner, Computed
tomography-based evaluation of template (NobelGuide)-guided implant positions: a prospective radiological study, Clin. Oral Implants Res. 22 (2011) 11571163.
[9] A. Pettersson, A. Komiyama, M. Hultin, K. Nsstrm, B. Klinge, Accuracy of
virtually planned and template guided implant surgery on edentate patients,
Clin. Implant Dent. Relat. Res. 14 (2012) 527537.
[10] S.K. Turbush, I. Turkyilmaz, Accuracy of three different types of stereolithographic
surgical guide in implant placement: an in vitro study, J. Prosthet. Dent. 108 (2012)
181188.
[11] M.M. Soares, N.D. Harari, E.S. Cardoso, M.C. Manso, M.B. Conz, G.M. Vidigal Jr,
An in vitro model to evaluate the accuracy of guided surgery systems, Int. J.
Oral Maxillofac. Implants 27 (2012) 824831.
[12] M. Cassetta, A. Di Mambro, M. Giansanti, L.V. Stefanelli, C. Cavallini, The
intrinsic error of a stereolithographic surgical template in implant guided
surgery, Int. J. Oral Maxillofac. Surg. 42 (2012) 264275.

[13] D. van Steenberghe, I. Naert, M. Andersson, I. Brajnovic, J. van Cleynenbreugel,

P. Suetens, A custom template and denitive prosthesis allowing immediate
implant loading in the maxilla: a clinical report, Int. J. Oral Maxillofac.
Implants 17 (2002) 663670.
[14] M. Cassetta, L.V. Stefanelli, M. Giansanti, A. Di Mambro, S. Calasso, Depth deviation
and occurrence of early surgical complications or unexpected events using a single
stereolithographic surgi-guide, Int. J. Oral Maxillofac. Surg. 40 (2011) 13771387.
[15] M. Cassetta, L.V. Stefanelli, M. Giansanti, S. Calasso, Accuracy of implant
placement with a stereolithographic surgical template, Int. J. Oral Maxillofac.
Implants 27 (2012) 655663.
[16] A. Komiyama, A. Pettersson, M. Hultin, K. Nsstrm, B. Klinge, Virtually
planned and template-guided implant surgery: an experimental model
matching approach, Clin. Oral Implants Res. 22 (2011) 308313.
[17] S.A. Al-Harbi, A.Y. Sun, Implant placement accuracy when using stereolithographic
template as a surgical guide: preliminary results, Implant Dent. 18 (2009) 4656.
[18] S. Huumonen, B. Haikola, K. Oikarinen, A.L. Sderholm, T. Remes-Lyly, K. Sipil,
Residual ridge resorption, lower denture stability and subjective complaints
among edentulous individuals, J. Oral Rehabil. 39 (2012) 384390.
[19] A.M. Sanna, L. Molly, D. van Steenberghe, Immediately loaded CADCAM manufactured xed complete dentures using apless implant placement procedures: a
cohort study of consecutive patients, J. Prosthet. Dent. 97 (2007) 331339.
[20] R. Gonzlez-Garca, F. Monje, The reliability of cone-beam computed tomography
to assess bone density at dental implant recipient sites: a histomorphometric
analysis by micro-CT, Clin. Oral Implants Res. (2012). (Epub ahead of print).
[21] P.S. Chang, T.H. Parker, C.W. Patrick Jr., M.J. Miller, The accuracy of stereolithography
in planning craniofacial bone replacement, J. Craniofac. Surg. 14 (2003) 164170.
[22] L.J. Stumpel, Deformation of stereolithographically produced surgical guides: an
observational case series report, Clin. Implant Dent. Relat. Res. 14 (2012) 442453.
[23] M.S. Reddy, T. Mayeld-Donahoo, F.J. Vanderven, M.K. Jeffcoat, A comparison
of the diagnostic advantages of panoramic radiography and computed
tomography scanning for placement of root form dental implants, Clin. Oral
Implants Res. 5 (1994) 229238.
[24] J. Horwitz, O. Zuabi, E.E. Machtei, Accuracy of a computerized tomographyguided template-assisted implant placement system: an in vitro study, Clin.
Oral Implants Res. 20 (2009) 11561162.

Megumi Ochi is a Clinical Staff in Department of Complete Denture Prosthodontics,

Tokyo Medical and Dental University in Japan. She has received her D.D.S. degree in
2008. Her research interests focus on implant overdenture (IOD) and surgical guides for
dental implant treatment.

Manabu Kanazawa is an Assistant Professor in Department of Complete Denture

Prosthodontics, Tokyo Medical and Dental University in Japan. He has received his
Ph.D. degree in 2006. His research interests focus on computer aided design/
computer aided manufacturing (CAD/CAM) and implant overdenture (IOD).

Daisuke Sato is an Adjunct Lecturer in Department of Oral Implantology and Regenerative Dental Medicine, Tokyo Medical and Dental University in Japan. He has received
his Ph.D. degree in 2004. His research interests focus on immediate loading of dental
implants.

Shohei Kasugai is a Professor in Department of Oral Implantology and Regenerative

Dental Medicine, Tokyo Medical and Dental University in Japan. He has received his
Ph.D. degree in 1983. His research interests focus on tissue regeneration.

Shigezo Hirano is an Adjunct Lecturer in Department of Complete Denture

Prosthodontics, Tokyo Medical and Dental University in Japan. He received his
Ph.D. degree in 1992. His research interests focus on implant overdenture and
masticatory performance.

Shunsuke Minakuchi is a Professor in Department of Complete Denture Prosthodontics, Tokyo Medical and Dental University in Japan. He has received his Ph.D.
degree in 1987. His research interests focus on CAD/CAM, masticatory performance
and denture base material.