- Email: [email protected]
CAM-guided surgical template implant surgery on human cadavers: Part I
Accuracy of CAD/CAM-guided surgical template implant surgery on human cadavers: Part I Andreas Pettersson, BSc,a Timo Kero, MSc,b Luc Gillot, DDS,c Bernard Cannas, DDS,d Jenny Fäldt, PhD,e Rikard Söderberg, PhD,f and Karin Näsström, PhD, DDSg Karolinska Institute, Huddinge, Sweden; Chalmers University of Technology, Göteborg, Sweden; Odontological Faculty, University Paris Descartes, Paris, France Statement of problem. An optimal method for approaching the clinical surgical situation, when using preoperatively, virtually planned implant positioning, is to transfer data to a CAD/CAM-guided surgical template with the definitive position of the implant placed after surgery. The accuracy of CAD/CAM-guided surgeries must be determined to provide safe treatment. Purpose. The purpose of this study was to compare the deviation between the position of virtually planned implants and the position of implants placed with a CAD/CAM-guided surgical template in the mandible and the maxilla in human cadavers. Material and methods. Ten maxillae and 7 mandibles, from completely edentulous cadavers, were scanned with CT, and 145 implants (Brånemark RP Groovy) were planned with software and placed with the aid of a CAD/CAM-guided surgical template. The preoperative CT scan was matched with the postoperative CT scan using voxel-based registration. The positions of the virtually planned implants were compared with the actual positions of the implants. Data were analyzed with a t test (α=.05). Results. The mean measurement differences between the computer-planned implants and implants placed after surgery for all implants placed were 1.25 mm (95% CI: 1.13-1.36) for the apex, 1.06 mm (95% CI: 0.97-1.16) for the hex, 0.28 mm (95% CI: 0.18-0.38) for the depth deviation, 2.64 degrees (95% CI: 2.41-2.87) for the angular deviation, and 0.71 mm (95% CI: 0.61-0.81 mm) for the translation deviation. Conclusions. The results demonstrated a statistically significant difference between mandibles and maxillae for the hex, apex, and depth measurements in the variation between the virtually planned implant positions and the positions of the implants placed after surgery with a CAD/CAM-guided surgical template. (J Prosthet Dent 2010;103:334-342)
Clinical Implications
A statistically significant difference in measurements was found when comparing the positions of virtually planned implants to the positions of implants placed with a CAD/CAM-guided surgical template on human cadavers. These results can be used to ensure safer patient treatment and to provide a better understanding of the deviations that can occur in CAD/CAM-guided template surgeries. PhD student, Section for Image and Functional Odontology, Department of Dental Medicine, Karolinska Institute. PhD student, Department of Product and Production Development, Chalmers University of Technology. c Clinical Instructor, Laboratory of Anatomy, Odontological Faculty, University Paris Descartes. d Clinical Instructor, Laboratory of Anatomy, Odontological Faculty, University Paris Descartes. e Senior Scientist, Early Development, Nobel Biocare AB, Göteborg, Sweden. f Professor, Head of Department of Product and Production Development, and Director, Wingquist Laboratory, Chalmers University of Technology. g Department Chair, Section for Image and Functional Odontology, Department of Dental Medicine, Karolinska Institute. a
b
The Journal of Prosthetic Dentistry
Pet tersson et al
335
June 2010
1 Overview of surgical procedure. From left to right: patient examination; radiographic guide and index; CT/CBCT scan; virtual planning; surgical template and index; and clinical procedures. The introduction of dental implants has revolutionized the oral rehabilitation of completely and partially edentulous patients.1-4 A subsequent revolution in implant dentistry involves the use of 3-dimensional (3D) planning programs to assist in accurate placement relative to anatomic and prosthodontic needs. The development of 3-D surgical planning programs, used along with computerized tomography (CT) images converted to 3-D models of clinically relevant data, has made it possible to mimic the true representation of the patient’s bone on the computer screen. This concept of implant treatment planning has been developed based on 3-D guided surgery, in which the dentist refers the patient to the radiologist, who scans the patient together with a radiographic guide and a radiographic index. The radiographic guide is scanned separately after the patient is scanned. The clinician then converts the digital imaging and communications in medicine (DICOM) files to a 3-D format of the patient’s bone and prosthesis, aligned by spherical markers. The clinician plans the patient treatment in a virtual environment, orders a surgical template according to the treatment plan, and the CAD/CAM-guided surgical template makes it possible to guide the drills and implants during the surgery. Using this method, the clinician may avoid potential surgi-
Pet tersson et al
cal mishaps, such as the placement of implants too close to significant anatomic structures, while optimizing the eventual prosthetic rehabilitation. One commercially available software used for planning the surgical and prosthetic phases is NobelGuide (Nobel Biocare AB, Göteborg, Sweden)5-7 (Fig. 1). This software enables the clinician to transfer planned implant information and positions in advance of the actual implant placement. While implant selection and placement has always been based on radiological judgments, usually determined by using 2-dimensional (2-D) radiographic images, this new method provides the clinician with radiological data for all 3 dimensions. After completing the surgical planning in a virtual environment, the data specifying the positions of the implants are used to order the components required for the surgery and to manufacture the surgical template by rapid prototyping.5,8 Custom-designed hardware components, which were developed for guiding the drills and the implants, are ordered and delivered to the clinician’s office. Using the surgical template and special hardware components, implants are placed by the surgeon at a predetermined depth and angle. The prosthesis is optimized by prosthetically driven implant placement. Studies reporting guided implant placement with the aid of a CAD/ CAM-guided surgical template have
been published, and include validation results of planned implant positions as compared to actual positions after surgery.8-11 The objective of this study was to compare the deviation between the positions of virtually planned implants and the positions of implants placed with a CAD/CAM-guided surgical template in the mandible and the maxilla in human cadavers. The research hypothesis was that there would be no difference in accuracy between the virtually planned implant positions and the actual positions of the implants placed in the mandible and maxilla with a CAD/CAM-guided surgical template. The resulting information is a first step in validating the accuracy of CAD/CAM-guided implant surgery.
MATERIAL AND METHODS Seventeen cadaver jaws (10 maxillae and 7 mandibles) were used in the present study. Cadaver jaws were consecutively collected by the Anatomy Department, University Paris Descartes (formerly Paris 5 University). The use of the human-derived materials was approved by the Laboratory of Anatomy Department, according to regulations in France. The criterion for inclusion was that the experimental material should be representative of the clinical situation; thus, each jaw was intact and completely edentulous. A pilot project was first undertaken using 3 jaws to define the protocols and methods to be used in the main phase of the study. A radiographic guide that recorded the soft tissue geometry and the future occlusal scheme was fabricated for each of the edentulous residual ridges. The radiographic guide was equivalent to the conventional prosthesis for patients and was fabricated by a dental laboratory technician with clear acrylic resin (ProBase Cold; Ivoclar Vivadent AG, Schaan, Lichtenstein) and contained 6-10 gutta-percha spherical markers with a diameter of 1.5 mm and a depth of 1 mm (Roeko Guttapercha Points GT; Coltène/Whaledent GmbH, Lan-
336
Volume 103 Issue 6 genau, Germany). The markers allowed matching of the radiographic guide with the aid of the markers to the patient volume from the CT scan and created a 3-D virtual model in dedicated software. A radiographic index was produced before the first CT scan of the patient (Regisil; Dentsply Caulk, Milford, Del). The radiographic guide was positioned onto the edentulous residual ridge, together with the radiographic index, and rubber bands were used to fixate the jaws. The cadaver jaw was then fixed to a plastic box with foam and imaged with a CT scanner (SOMATOM Sensation 10; Siemens AG, Erlangen, Germany). The CT scans were performed using a 0-degree gantry tilt, 120 kV, 80 mAs, a slice width of 0.75 mm, and a reconstruction increment of 0.5 mm. A second CT scan was then performed of the radiographic guide alone, with the same settings as for the previous scan. The data was exported from the CT scanner by the radiologist and provided electronically. The resulting DICOM files were converted with software (Procera Software Clinical Design Premium, version 1.5; Nobel Biocare AB) to form a 3-D virtual model of the matched bone and the radiographic guide. The DICOM files were loaded into the software’s CT-scan file converter application, and the files for the bone and prosthesis were selected. The prosthesis and bone were matched as the gutta-percha markers were found and registered from the 2
separate scans. The software eventually created a single file containing the 3-D shape of the radiographic guide and bone combined. To maximize the number of implants in the study to be compared, a surgical plan was developed with a maximum number of implants (Brånemark System MK III Groovy RP; Nobel Biocare AB) in each jaw. Some of the implants were to be used to evaluate guiding in complex areas, such as close to the mandibular canal, incisive canal, or cortex. The planning software implant coordinates were exported to CAD design software (Procera Software Clinical Design Premium, version 1.5; Nobel Biocare AB). CAD/CAM-guided surgical templates, which use specifically designed sleeves to guide the drills and the implants during surgical placement, were ordered, together with clinically relevant components, from Nobel Biocare AB. All implants were placed in the cadaver jaws using a CAD/CAM-guided surgical template and according to the treatment plan. The surgical protocol has been described previously.5-7 The surgical protocol also included information on the drilling sequence, torque value, and details on the placement of each implant. The drilling sequence for the cadaver surgeries simulated the actual clinical situation. During the implant placement procedure, only visual guidance, provided by the soft tissue in contact with the intaglio surface of the
surgical template, and manual pressure were used while positioning the anchor pins (Guided Anchor Pin, diameter 1.5 mm; Nobel Biocare AB). The anchor pins were placed to secure the position of the surgical template to the bone and soft tissue. Three to 5 anchor pins were placed in each jaw during this procedure. With the aid of an attached template abutment, the first implant was placed to secure the position of the surgical template. The implant was attached to the surgical template using a template abutment (Guided Template Abutment Brånemark System RP; Nobel Biocare AB) which connected to the implant and engaged in the sleeve of the CAD/CAMguided surgical template to secure the position after tightening the screw. A second implant was placed to fix the second template abutment. In some situations in which posterior implants were placed, a third template abutment was placed to stabilize the surgical template in the best possible manner. Standard components were used during the surgical procedure, according to the NobelGuide protocol.5,6,7 A total of 145 Brånemark Groovy RP implants were placed, 67 implants in the mandible and 78 in the maxilla. Immediately after the surgical phase was completed, a postoperative CT scan of the jaw was performed with the same settings as for the preoperative scans. In addition, a dissection of some jaws was performed to view the anatomy and the
2 A, Virtually planned implant position. B, Actual placement of implants after dissection.
The Journal of Prosthetic Dentistry
Pet tersson et al
337
June 2010 position of the implant in comparison to the planned position (Fig. 2). The postoperative CT scan was matched against the preoperative CT scan using a 3-D voxel-based matching method, as previously described.13 The 2 different data sets from the pre- and postoperative CT scans were registered into a single coordinate system by calculation of mutual information between the corresponding voxels in the 2 data sets. The voxelbased software searched for corresponding gray values between the 2 data sets and aligned them. With the aligned data sets, the actual positions of the implants could be compared with the virtually planned positions. Linear and angular discrepancies between the planned and actual positions of each implant were analyzed in 3-D. The Euclidean distance between the planned and actual implant position was measured at the center of the apex and center of the hex of the implant. The apex refers to the tip of the implant and the hex refers to the center of the prosthetic connection of the implant. The angular deviation between the main axes of the planned and actual implant positions was computed as well. The matching and calculation procedure used was similar to that previously described.8,9,12,13
Statistical calculations were performed using statistical software (SAS Enterprise 4; SAS Institute, Inc, Cary, NC). Calculations were further reproduced using a second statistical program (STATISTICA 7.0; StatSoft, Inc, Tulsa, Okla) to verify results. Deviation apex, deviation hex, translation deviation, and angle deviation were log (e) transformed to have approximately normal distributed data for use in the statistical analyses. To test for deviation equal to zero, a 1-sample t test was used. Deviations were summarized using median, minimum, maximum, mean, and standard deviation and the corresponding 95% confidence interval. Differences between the mandible and maxilla were tested using the 2-sample t test. All tests were 2-sided, and α=.05 was considered as statistically significant.
RESULTS The differences between the mandible and maxilla were statistically significant for apex, hex, and depth deviation (Table I). The differences between the virtually planned implants and the actual positions of the implants were statistically significant for all 5 outcome variables: apex, hex, depth, translation deviation, and an-
gle (Tables II, III, and IV). A box plot analysis was used to show the deviations in the mandible and maxilla; the differences between the mandible and the maxilla are illustrated for the deviation between the planned implant compared to the placed implant in the apex, hex, angle, depth, and translation measurements, in different panels (Fig. 3). The translation deviation graph illustrates the parallel deviation between the virtually planned implant and the implant placed after surgery (Fig. 4). The translation error difference between placed implants (greenblue color) and the planned implants (gray implants with threads) can be observed by visualizing the segmented implants placed after surgery and the virtually planned implants after matching (Fig. 5). The greatest random variation observed in the data was observed in the mandible “3Mand,” with the maximum deviation (Table II). All implants were positioned in the same direction in terms of the repositioning error. The maximum range deviations for both the mandible and maxilla occurred in the same cadaver head (3Mand and 13Max) (Table II).
Table I. Summary statistics for tests between mandible and maxilla. Numbers of implants: 67 for mandible and 78 for maxilla. Deviations in millimeters. Note: Negative value for depth deviation indicates that implant did not reach planned position. Positive value indicates that implant was placed deeper than planned position. LL: lower level, UL: upper level Mandible Range Variable
Mean Min Max
SD
Maxilla Range 95% CI 95% CI LL UL Mean Min Max
SD
95% CI 95% CI LL UL
P
Depth
0.48
–0.07
1.46
0.52
0.36
0.61
0.1
–0.03
1.61
0.60
0.03
0.24
<.001
Apex
1.24
0.13
3.63
0.58
1.08
1.43
0.96
0.12
2.43
0.50
0.86
1.08
.01
Hex
1.05
0.41
3.13
0.47
0.94
1.18
0.83
0.07
2.78
0.57
0.73
0.94
.01
Angle
2.46
0.26
7.44
0.67
2.09
2.9
2.02
0.08
5.38
0.66
1.74
2.34
.08
Translation
0.49
0.01
2.87
1.12
0.37
0.64
0.45
0.00
2.24
1.07
0.35
0.57
.63
Pet tersson et al
338
Volume 103 Issue 6
Table II. Summary of total deviations at the apex and hex and per cadaver and jaw type. Deviations in millimeters. LL: lower level, UL: upper level Cadaver
Deviation Apex
Deviation Hex
Number of 95% CI 95% CI Implants Median Mean SD LL UL 1Mand
P
Median Mean SD
95% CI 95% CI LL UL
P
9
1.17
1.07
0.46
0.77
1.17
.001
0.69
0.71
0.16
0.61
0.74
.001
2Mand
9
0.96
1.03
0.36
0.79
1.11
.001
0.95
0.79
0.27
0.61
0.85
.001
3Mand
11
3.14
2.52
1.16
1.83
2.73
.001
2.46
2.19
0.90
1.66
2.35
.001
4Mand
11
1.6
1.47
0.84
0.97
1.62
.001
1.36
1.26
0.41
1.02
1.33
.001
5Mand
8
1.47
1.60
0.56
1.21
1.74
.001
1.18
1.81
0.09
1.75
1.83
.001
6Mand
11
1.16
1.32
0.54
1.00
1.42
.001
0.91
0.92
0.28
0.75
0.97
.001
7Mand
8
0.87
0.91
0.24
0.74
0.97
.001
1.06
1.04
0.21
0.89
1.09
.001
8Max
8
1.17
1.15
0.23
0.99
1.20
.001
0.91
0.99
0.20
0.85
1.03
.001
9Max
6
0.51
0.55
0.19
0.40
0.61
.001
0.54
0.56
0.19
0.41
0.62
.001
10Max
8
1.13
1.07
0.42
0.78
1.17
.001
0.85
0.81
0.19
0.68
0.86
.001
11Max
8
1.02
0.97
0.33
0.74
1.05
.001
0.66
0.62
0.15
0.51
0.65
.001
12Max
10
1.00
1.01
0.22
0.88
1.06
.001
0.57
0.57
0.25
0.42
0.62
.001
13Max
8
1.00
1.30
0.64
0.86
1.46
.001
1.78
1.90
0.62
1.47
2.05
.001
14Max
8
0.67
0.67
0.22
0.52
0.73
.001
0.86
0.81
0.14
0.71
0.84
.001
15Max
6
1.26
1.32
0.52
0.90
1.49
.002
0.95
1.10
0.35
0.82
1.21
.001
16Max
9
1.38
1.46
0.30
1.27
1.53
.001
1.46
1.48
0.12
1.40
1.50
.001
17Max
7
1.30
1.07
0.50
0.70
1.21
.001
0.79
0.67
0.37
0.40
0.77
.001
145
1.12
1.25
0.68
1.13
1.36
<.001
0.93
1.06
0.58
0.97
1.16
<.001
Total
Table III. Summary of total depth and translation deviations and per cadaver and jaw type. Deviations in millimeters. Note: Negative value for depth deviation indicates that implant did not reach planned position. Positive value indicates that implant was placed deeper than planned position. LL: lower level, UL: upper level Cadaver
Deviation Depth
Translation Deviation
Number of 95% CI 95% CI Implants Median Mean SD LL UL
P
Median Mean SD
95% CI 95% CI LL UL
P
1Mand
9
0.43
0.45
0.13
0.37
0.48
.001
0.49
0.56
0.36
0.32
0.64
.002
2Mand
9
0.55
0.6
0.27
0.42
0.66
.001
0.60
0.57
0.29
0.38
0.63
.001
3Mand
11
0.66
0.64
0.34
0.44
0.70
.001
1.43
1.52
0.81
1.04
1.66
.001
4Mand
11
0.84
0.57
0.78
0.11
0.71
.036
0.66
0.97
0.82
0.49
1.12
.003
5Mand
8
0.66
0.35
0.72
–0.15
0.53
.212
0.25
0.40
0.52
0.04
0.53
.066
6Mand
11
0.33
0.13
0.63
–0.24
0.24
.509
0.57
0.75
0.71
0.33
0.88
.006
7Mand
8
0.69
0.67
0.17
0.55
0.71
.001
0.44
0.39
0.18
0.27
0.43
.001
8Max
8
0.70
0.65
0.43
0.36
0.76
.003
0.35
0.38
0.32
0.16
0.46
.011
9Max
6
–0.21
–0.19
0.10
–0.27
–0.15
.007
0.35
0.40
0.16
0.27
0.45
.002
10Max
8
0.32
0.32
0.16
0.21
0.36
.001
0.16
0.18
0.15
0.08
0.22
.010
11Max
8
0.39
0.33
0.19
0.20
0.38
.002
0.57
0.60
0.28
0.41
0.67
.001
12Max
10
0.31
0.35
0.19
0.23
0.39
.001
0.43
0.50
0.29
0.32
0.56
.001
13Max
8
0.70
0.88
0.40
0.60
0.97
.001
0.80
1.08
0.59
0.67
1.23
.001
14Max
8
–0.54
–0.41
0.40
–0.69
–0.31
.022
0.43
0.46
0.22
0.30
0.51
.001
15Max
6
–0.2
–0.29
0.23
–0.47
–0.21
.026
0.93
1.01
0.49
0.61
1.17
.004
16Max
9
–0.92
–0.92
0.28
–1.10
–0.86
.001
1.28
1.29
0.24
1.14
1.35
.001
17Max
7
0.15
0.20
0.27
0.01
0.28
.089
0.80
0.60
0.51
0.22
0.74
.022
145
0.39
0.28
0.59
0.18
0.38
<.001
0.56
0.71
0.59
0.61
0.81
<.001
Total
The Journal of Prosthetic Dentistry
Pet tersson et al
339
June 2010
Table IV. Summary of total angular deviations and per cadaver and jaw type. Deviations in degrees. LL: lower level, UL: upper level Cadaver
Angle Deviation
Number of Implants Median
7 6 5 4 3 2 1 0 –1 –2
* *
1.16
P
9
2.54
2.82
1.54
1.81
3.16
.001
9
2.29
2.20
1.46
1.25
2.52
.002
3Mand
11
2.62
2.66
1.20
1.95
2.87
.001
4Mand
11
4.02
4.40
1.98
3.23
4.75
.001
5Mand
8
2.67
2.71
1.70
1.53
3.13
.003
6Mand
11
3.39
3.40
1.50
2.51
3.67
.001
7Mand
8
2.24
2.04
0.84
1.46
2.25
.001
8Max
8
1.60
1.84
1.18
1.02
2.13
.003
9Max
6
1.34
1.39
0.38
1.09
1.52
.001
10Max
8
2.26
2.01
1.27
1.13
2.32
.003
11Max
8
2.12
2.47
1.14
1.67
2.75
.001
12Max
10
2.64
2.67
0.67
2.26
2.80
.001
13Max
8
4.61
4.29
0.92
3.65
4.51
.001
14Max
8
2.74
2.66
0.68
2.18
2.82
.001
15Max
6
1.72
1.82
1.00
1.02
2.14
.007
16Max
9
2.05
1.88
0.70
1.42
2.03
.001
17Max
7
2.44
2.22
0.97
1.49
2.49
.001
145
2.48
2.64
1.42
2.41
2.87
<.001
Deviation Hex
* *
* * * * *
1.02
1.06
Translation Deviation
*
*
0.56
0.56
* * 0.55
Maxilla
Angle Deviation
* * *
0.87
Deviation Depth
* * *
Mandible
95% CI 95% CI LL UL
2Mand
Deviation Apex
3 2 1 0 –1 –2
SD
1Mand
Total
7 6 5 4
Mean
Mandible
2.68
2.31
Mandible
Maxilla
*
0.21
Maxilla
3 Box plot of deviations in mandible and maxilla. Angle deviations are in degrees, all other deviations are in millimeters. Note: Length of box corresponds to interquartile range. Horizontal line and numbers within box correspond to median. Plus sign indicates mean. Crosses indicate outliers.
Pet tersson et al
340
Volume 103 Issue 6 28
Number of Implants
24 20 16 12 8 4 0
0 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9
2.5
3.0
Translational Deviation 4 Histogram of translational deviation. Measurements in millimeters.
5 Translation deviation. Figure from 3Mand (Table III), with largest error in group. Planned implants are gray with threads. Postoperative implants are blue-green color. Note all implants are positioned in parallel direction after surgery compared to planned position.
DISCUSSION Based upon the results of this study, the research hypothesis cannot be accepted, as there was a statistically significant difference between the results of the mandible and the maxilla for depth deviation, apex, and hex, with the greater variation registered in the mandible. One explanation for the variation could be that the surgical template in the mandible is less stable, as it covers a smaller area compared to the maxilla. The results demonstrated a significant difference between the virtually planned implants and the actually placed implants for all 5 variables, apex, hex, depth, translation deviation, and angle. However, although the values were statistically significant,
it was not clear whether the results are clinically significant. CAD/CAM-guided implant surgery offers the clinician another method to ensure accurate and prosthetically driven implant placement. As with most new advancements, there may be limitations and risks. While the most accurate assessments will come from clinical use, this study on cadavers might add valuable information about the accuracy of guided surgery. The present study provides information about the results of the specific treatment performed, with limitations such as the disadvantage of not using the surgical index. The results could be used to improve instructions to clinicians and enable clinicians to provide safer patient treatment. In a review article including results from
The Journal of Prosthetic Dentistry
8 published studies, Schneider et al11 presented results with a mean deviation at the hex of 1.07 mm (95% CI: 0.76-1.22 mm) and a mean deviation at the apex of 1.63 mm (95% CI: 1.262 mm). The present study presents a mean deviation at the hex of 1.06 mm (95% CI: 0.97-1.16 mm) and a mean deviation at the apex of 1.25 mm (95% CI: 1.13-1.36 mm) and, thus, demonstrates similar or better results. The positioning of the radiographic guide calls for the patient to occlude onto a radiographic index and radiographic guide during the imaging procedure. In this study, the radiographic guides could not be positioned in the conventional manner. The guides had to be manually placed with the help of rubber bands. This method produced a risk of positioning the radiographic
Pet tersson et al
341
June 2010 guide in a less than ideal manner. In 12 out of 17 cadaver jaws, a space was visible between the soft tissue and radiographic guide. The traditional clinical procedure to position the CAD/CAM-guided surgical template onto the registered position of the radiographic guide could, for obvious reasons, not be performed, due to the inability to make an occlusal registration. Therefore, the surgeon manually placed the CAD/CAM-guided surgical template, which may have resulted in less accurate positioning of the implants. Implant planning was performed in an experimental manner, placing as many implants as possible with regard to the anatomical situation, considering the availability of the bone volume and the technical limitations within the system. The limitation of determining the position for the radiographic guide and surgical template in the experimental model introduced errors that should not be present clinically. To preserve the specimens, each cadaver was frozen and then thawed up to 4 times, which may have caused dehydration and a change in the size and shape of the soft tissue. These alterations of the soft tissue probably caused minor changes in the positioning of the radiographic guide and surgical template from the planned positioning. In an attempt to mimic the clinical setting as closely as possible, formaldehyde was not used. Another issue was the flexibility of the soft tissue in the cadavers, which might add further variation to the results. All of these minor differences between a living patient’s tissue and the cadaver tissue may have contributed to the errors in implant placement, in comparison to the planned implant positions. One cadaver in the present study demonstrated the maximum errors for both the mandible and the maxilla (Fig. 5). This indicates that the radiographic guide and surgical template were likely malpositioned. When reviewing the variation between the planned and the placed
Pet tersson et al
implants in the mandible (3Mand) visually (Fig. 5), it appeared that a parallel movement (translation) occurred with the placed implants compared to the planned implants. Malpositioning of the radiographic guide and the surgical template might result in errors, such as rotation and misplacement. additional research is required to obtain more information about the deviation from misplacements of the radiographic guide and the surgical template. In this study, the variations between the virtually planned implants and the implants placed after surgery were compared using matching methods similar to those presented in a study by Van Assche et al.9 This study 9 was performed on formalintreated cadavers and presented results with a range of 0.3-2.3 mm with a mean value of 1.1 mm for the hex and a range of 0.7-2.4 mm with a mean value of 2.0 mm for the apex. The current study presented results with a range of 0.07-3.13 mm and a mean value of 1.06 mm for the hex and a range of 0.12-3.63 mm with a mean value of 1.25 mm for the apex. One reason for the differences between this study and the study by Van Assche may be the number of experimental surgeries performed and included in each of the different studies. The study by Van Assche included only 4 subjects and 12 implants, while this study included 17 subjects and 145 implants. Another reason might be that tooth-supported CAD/ CAM surgical templates were used in the study by Van Assche, while in the current study, only completely edentulous CAD/CAM surgical templates were used. Tooth-supported CAD/ CAM surgical templates are probably more stable compared to those for totally edentulous subjects, adding a possible disadvantage when excluding the surgical index and, thus, causing larger deviations. The current study presents the variation obtained in completely edentulous specimens that were frozen and thawed, and without the option of using a surgi-
cal index. The results could be used to provide a better understanding of possible deviations that could occur when performing CAD/CAM-guided surgeries. This information can also be useful for clinicians, in improving this specific treatment method. In future studies, if positioning and repositioning errors are reduced, the results will depend on the surgical system limitations and errors, as well as on variations between surgeons. Additional studies are needed to learn more about the various conditions affecting virtual planning and guided surgery, and how accurate CAD/ CAM-guided surgery is compared to the freehand placement of implants. CONCLUSIONS Within the limitations of this study, the results demonstrate a statistically significant difference between mandibles and maxillae for the hex, apex, and depth measurements between the virtually planned implant positions and the positions of the implants placed after surgery with a CAD/CAM-guided surgical template on human cadavers. REFERENCES 1. Brånemark PI, Adell R, Breine U, Hansson BO, Lindström J, Ohlsson A. Intra-osseous anchorage of dental prostheses. I. Experimental studies. Scand J Plast Reconstr Surg 1969;3:81-100. 2. Jemt T, Johansson J. Implant treatment in the edentulous maxillae: a 15-year followup study on 76 consecutive patients provided with fixed prostheses. Clin Implant Dent Relat Res 2006;8:61-9. 3. Adell R, Eriksson B, Lekholm U, Brånemark PI, Jemt T. Long-term follow-up study of osseointegrated implants in the treatment of totally edentulous jaws. Int J Oral Maxillofac Implants 1990;5:347-59. 4. Lekholm U, Gunne J, Henry P, Higuchi K, Lindén U, Bergström C, et al. Survival of the Brånemark implant in partially edentulous jaws: a 10-year prospective multicenter study. Int J Oral Maxillofac Implants 1999;145:639-45.
342
Volume 103 Issue 6 5. van Steenberghe D, Glauser R, Blomback U, Andersson M, Schutyser F, Pettersson A, et al. A computed tomographic scan-derived customized surgical template and fixed prosthesis for flapless surgery and immediate loading of implants in fully edentulous maxillae: a prospective multicenter study. Clin Implant Dent Relat Res 2005;7 Suppl 1:S111-20. 6. Marchack CB. An immediately loaded CAD/ CAM-guided definitive prosthesis: a clinical report. J Prosthet Dent 2005;931:8-12. 7. Parel SM, Triplett RG. Interactive imaging for implant planning, placement, and prosthesis construction. J Oral Maxillofac Surg 2004;62 (9 Suppl 2):41-7. 8. van Steenberghe D, Naert I, Andersson M, Brajnovic I, van Cleynenbreugel J, Suetens P. A custom template and definitive prosthesis allowing immediate implant loading in the maxilla: a clinical report. Int J Oral Maxillofac Implants 2002;17:663-70.
9. Van Assche N, van Steenberghe D, Guerrero ME, Hirsch E, Schutyser F, Quirynen M, et al. Accuracy of implant placement based on pre-surgical planning of three-dimensional cone-beam images: a pilot study. J Clin Periodontol 2007;34:816-21. 10.Di Giacomo GA, Cury PR, de Araujo NS, Sendyk WR, Sendyk CL. Clinical application of stereolithographic surgical guides for implant placement: preliminary results. J Periodontol 2005;76:503-7. 11.Schneider D, Marquardt P, Zwahlen M, Jung R E. A systematic review on the accuracy and the clinical outcome of computerguided template-based implant dentistry. Clin Oral Implants Res 2009;20 Suppl 4:73-86. 12.Rade L, Westergen B. Mathematics handbook for science and engineering. 5th ed. New York: Springer; 2004. p. 85. 13.Maes F, Collignon A, Vandermeulen D, Marchal G, Suetens P. Multimodality image registration by maximization of mutual information. IEEE Trans Med Imaging 1997;16:187-98.
Corresponding author: Mr Andreas Pettersson Karolinska Institute Section for Image and Functional Odontology Department of Dental Medicine Alfred Nobels Allé 8 Box 4064 SE-14104 Huddinge SWEDEN Fax: +46 87795301 E-mail: [email protected] Acknowledgments The authors thank Matts Andersson for his contribution of ideas throughout the study, the department of Early Development at Nobel Biocare AB, Göteborg, Sweden, for their support of this study, and Filip Schutyser for his expertise in 3-D software design and function and the matching process. The authors also thank Prof Jean-François Gaudy for his support and continuous help during the study at the Laboratory of Anatomy, Odontological Faculty, University Paris Descartes, Paris, France. Copyright © 2010 by the Editorial Council for The Journal of Prosthetic Dentistry.
Access to The Journal of Prosthetic Dentistry Online is reserved for print subscribers! Full-text access to The Journal of Prosthetic Dentistry Online is available for all print subscribers. To activate your individual online subscription, please visit The Journal of Prosthetic Dentistry Online. Point your browser to http://www.journals. elsevierhealth.com/periodicals/ympr/home, follow the prompts to activate online access here, and follow the instructions. To activate your account, you will need your subscriber account number, which you can find on your mailing label (note: the number of digits in your subscriber account number varies from 6 to 10). See the example below in which the subscriber account number has been circled. Sample mailing label
This is your subscription account number
*********AUTO**SCH 3-DIGIT 001 1 V97-3 J010 12345678-9 J. H. DOE 531 MAIN ST CENTER CITY, NY 10001-001
Personal subscriptions to The Journal of Prosthetic Dentistry Online are for individual use only and may not be transferred. Use of The Journal of Prosthetic Dentistry Online is subject to agreement to the terms and conditions as indicated online.
The Journal of Prosthetic Dentistry
Pet tersson et al
Clinical Implications
A statistically significant difference in measurements was found when comparing the positions of virtually planned implants to the positions of implants placed with a CAD/CAM-guided surgical template on human cadavers. These results can be used to ensure safer patient treatment and to provide a better understanding of the deviations that can occur in CAD/CAM-guided template surgeries. PhD student, Section for Image and Functional Odontology, Department of Dental Medicine, Karolinska Institute. PhD student, Department of Product and Production Development, Chalmers University of Technology. c Clinical Instructor, Laboratory of Anatomy, Odontological Faculty, University Paris Descartes. d Clinical Instructor, Laboratory of Anatomy, Odontological Faculty, University Paris Descartes. e Senior Scientist, Early Development, Nobel Biocare AB, Göteborg, Sweden. f Professor, Head of Department of Product and Production Development, and Director, Wingquist Laboratory, Chalmers University of Technology. g Department Chair, Section for Image and Functional Odontology, Department of Dental Medicine, Karolinska Institute. a
b
The Journal of Prosthetic Dentistry
Pet tersson et al
335
June 2010
1 Overview of surgical procedure. From left to right: patient examination; radiographic guide and index; CT/CBCT scan; virtual planning; surgical template and index; and clinical procedures. The introduction of dental implants has revolutionized the oral rehabilitation of completely and partially edentulous patients.1-4 A subsequent revolution in implant dentistry involves the use of 3-dimensional (3D) planning programs to assist in accurate placement relative to anatomic and prosthodontic needs. The development of 3-D surgical planning programs, used along with computerized tomography (CT) images converted to 3-D models of clinically relevant data, has made it possible to mimic the true representation of the patient’s bone on the computer screen. This concept of implant treatment planning has been developed based on 3-D guided surgery, in which the dentist refers the patient to the radiologist, who scans the patient together with a radiographic guide and a radiographic index. The radiographic guide is scanned separately after the patient is scanned. The clinician then converts the digital imaging and communications in medicine (DICOM) files to a 3-D format of the patient’s bone and prosthesis, aligned by spherical markers. The clinician plans the patient treatment in a virtual environment, orders a surgical template according to the treatment plan, and the CAD/CAM-guided surgical template makes it possible to guide the drills and implants during the surgery. Using this method, the clinician may avoid potential surgi-
Pet tersson et al
cal mishaps, such as the placement of implants too close to significant anatomic structures, while optimizing the eventual prosthetic rehabilitation. One commercially available software used for planning the surgical and prosthetic phases is NobelGuide (Nobel Biocare AB, Göteborg, Sweden)5-7 (Fig. 1). This software enables the clinician to transfer planned implant information and positions in advance of the actual implant placement. While implant selection and placement has always been based on radiological judgments, usually determined by using 2-dimensional (2-D) radiographic images, this new method provides the clinician with radiological data for all 3 dimensions. After completing the surgical planning in a virtual environment, the data specifying the positions of the implants are used to order the components required for the surgery and to manufacture the surgical template by rapid prototyping.5,8 Custom-designed hardware components, which were developed for guiding the drills and the implants, are ordered and delivered to the clinician’s office. Using the surgical template and special hardware components, implants are placed by the surgeon at a predetermined depth and angle. The prosthesis is optimized by prosthetically driven implant placement. Studies reporting guided implant placement with the aid of a CAD/ CAM-guided surgical template have
been published, and include validation results of planned implant positions as compared to actual positions after surgery.8-11 The objective of this study was to compare the deviation between the positions of virtually planned implants and the positions of implants placed with a CAD/CAM-guided surgical template in the mandible and the maxilla in human cadavers. The research hypothesis was that there would be no difference in accuracy between the virtually planned implant positions and the actual positions of the implants placed in the mandible and maxilla with a CAD/CAM-guided surgical template. The resulting information is a first step in validating the accuracy of CAD/CAM-guided implant surgery.
MATERIAL AND METHODS Seventeen cadaver jaws (10 maxillae and 7 mandibles) were used in the present study. Cadaver jaws were consecutively collected by the Anatomy Department, University Paris Descartes (formerly Paris 5 University). The use of the human-derived materials was approved by the Laboratory of Anatomy Department, according to regulations in France. The criterion for inclusion was that the experimental material should be representative of the clinical situation; thus, each jaw was intact and completely edentulous. A pilot project was first undertaken using 3 jaws to define the protocols and methods to be used in the main phase of the study. A radiographic guide that recorded the soft tissue geometry and the future occlusal scheme was fabricated for each of the edentulous residual ridges. The radiographic guide was equivalent to the conventional prosthesis for patients and was fabricated by a dental laboratory technician with clear acrylic resin (ProBase Cold; Ivoclar Vivadent AG, Schaan, Lichtenstein) and contained 6-10 gutta-percha spherical markers with a diameter of 1.5 mm and a depth of 1 mm (Roeko Guttapercha Points GT; Coltène/Whaledent GmbH, Lan-
336
Volume 103 Issue 6 genau, Germany). The markers allowed matching of the radiographic guide with the aid of the markers to the patient volume from the CT scan and created a 3-D virtual model in dedicated software. A radiographic index was produced before the first CT scan of the patient (Regisil; Dentsply Caulk, Milford, Del). The radiographic guide was positioned onto the edentulous residual ridge, together with the radiographic index, and rubber bands were used to fixate the jaws. The cadaver jaw was then fixed to a plastic box with foam and imaged with a CT scanner (SOMATOM Sensation 10; Siemens AG, Erlangen, Germany). The CT scans were performed using a 0-degree gantry tilt, 120 kV, 80 mAs, a slice width of 0.75 mm, and a reconstruction increment of 0.5 mm. A second CT scan was then performed of the radiographic guide alone, with the same settings as for the previous scan. The data was exported from the CT scanner by the radiologist and provided electronically. The resulting DICOM files were converted with software (Procera Software Clinical Design Premium, version 1.5; Nobel Biocare AB) to form a 3-D virtual model of the matched bone and the radiographic guide. The DICOM files were loaded into the software’s CT-scan file converter application, and the files for the bone and prosthesis were selected. The prosthesis and bone were matched as the gutta-percha markers were found and registered from the 2
separate scans. The software eventually created a single file containing the 3-D shape of the radiographic guide and bone combined. To maximize the number of implants in the study to be compared, a surgical plan was developed with a maximum number of implants (Brånemark System MK III Groovy RP; Nobel Biocare AB) in each jaw. Some of the implants were to be used to evaluate guiding in complex areas, such as close to the mandibular canal, incisive canal, or cortex. The planning software implant coordinates were exported to CAD design software (Procera Software Clinical Design Premium, version 1.5; Nobel Biocare AB). CAD/CAM-guided surgical templates, which use specifically designed sleeves to guide the drills and the implants during surgical placement, were ordered, together with clinically relevant components, from Nobel Biocare AB. All implants were placed in the cadaver jaws using a CAD/CAM-guided surgical template and according to the treatment plan. The surgical protocol has been described previously.5-7 The surgical protocol also included information on the drilling sequence, torque value, and details on the placement of each implant. The drilling sequence for the cadaver surgeries simulated the actual clinical situation. During the implant placement procedure, only visual guidance, provided by the soft tissue in contact with the intaglio surface of the
surgical template, and manual pressure were used while positioning the anchor pins (Guided Anchor Pin, diameter 1.5 mm; Nobel Biocare AB). The anchor pins were placed to secure the position of the surgical template to the bone and soft tissue. Three to 5 anchor pins were placed in each jaw during this procedure. With the aid of an attached template abutment, the first implant was placed to secure the position of the surgical template. The implant was attached to the surgical template using a template abutment (Guided Template Abutment Brånemark System RP; Nobel Biocare AB) which connected to the implant and engaged in the sleeve of the CAD/CAMguided surgical template to secure the position after tightening the screw. A second implant was placed to fix the second template abutment. In some situations in which posterior implants were placed, a third template abutment was placed to stabilize the surgical template in the best possible manner. Standard components were used during the surgical procedure, according to the NobelGuide protocol.5,6,7 A total of 145 Brånemark Groovy RP implants were placed, 67 implants in the mandible and 78 in the maxilla. Immediately after the surgical phase was completed, a postoperative CT scan of the jaw was performed with the same settings as for the preoperative scans. In addition, a dissection of some jaws was performed to view the anatomy and the
2 A, Virtually planned implant position. B, Actual placement of implants after dissection.
The Journal of Prosthetic Dentistry
Pet tersson et al
337
June 2010 position of the implant in comparison to the planned position (Fig. 2). The postoperative CT scan was matched against the preoperative CT scan using a 3-D voxel-based matching method, as previously described.13 The 2 different data sets from the pre- and postoperative CT scans were registered into a single coordinate system by calculation of mutual information between the corresponding voxels in the 2 data sets. The voxelbased software searched for corresponding gray values between the 2 data sets and aligned them. With the aligned data sets, the actual positions of the implants could be compared with the virtually planned positions. Linear and angular discrepancies between the planned and actual positions of each implant were analyzed in 3-D. The Euclidean distance between the planned and actual implant position was measured at the center of the apex and center of the hex of the implant. The apex refers to the tip of the implant and the hex refers to the center of the prosthetic connection of the implant. The angular deviation between the main axes of the planned and actual implant positions was computed as well. The matching and calculation procedure used was similar to that previously described.8,9,12,13
Statistical calculations were performed using statistical software (SAS Enterprise 4; SAS Institute, Inc, Cary, NC). Calculations were further reproduced using a second statistical program (STATISTICA 7.0; StatSoft, Inc, Tulsa, Okla) to verify results. Deviation apex, deviation hex, translation deviation, and angle deviation were log (e) transformed to have approximately normal distributed data for use in the statistical analyses. To test for deviation equal to zero, a 1-sample t test was used. Deviations were summarized using median, minimum, maximum, mean, and standard deviation and the corresponding 95% confidence interval. Differences between the mandible and maxilla were tested using the 2-sample t test. All tests were 2-sided, and α=.05 was considered as statistically significant.
RESULTS The differences between the mandible and maxilla were statistically significant for apex, hex, and depth deviation (Table I). The differences between the virtually planned implants and the actual positions of the implants were statistically significant for all 5 outcome variables: apex, hex, depth, translation deviation, and an-
gle (Tables II, III, and IV). A box plot analysis was used to show the deviations in the mandible and maxilla; the differences between the mandible and the maxilla are illustrated for the deviation between the planned implant compared to the placed implant in the apex, hex, angle, depth, and translation measurements, in different panels (Fig. 3). The translation deviation graph illustrates the parallel deviation between the virtually planned implant and the implant placed after surgery (Fig. 4). The translation error difference between placed implants (greenblue color) and the planned implants (gray implants with threads) can be observed by visualizing the segmented implants placed after surgery and the virtually planned implants after matching (Fig. 5). The greatest random variation observed in the data was observed in the mandible “3Mand,” with the maximum deviation (Table II). All implants were positioned in the same direction in terms of the repositioning error. The maximum range deviations for both the mandible and maxilla occurred in the same cadaver head (3Mand and 13Max) (Table II).
Table I. Summary statistics for tests between mandible and maxilla. Numbers of implants: 67 for mandible and 78 for maxilla. Deviations in millimeters. Note: Negative value for depth deviation indicates that implant did not reach planned position. Positive value indicates that implant was placed deeper than planned position. LL: lower level, UL: upper level Mandible Range Variable
Mean Min Max
SD
Maxilla Range 95% CI 95% CI LL UL Mean Min Max
SD
95% CI 95% CI LL UL
P
Depth
0.48
–0.07
1.46
0.52
0.36
0.61
0.1
–0.03
1.61
0.60
0.03
0.24
<.001
Apex
1.24
0.13
3.63
0.58
1.08
1.43
0.96
0.12
2.43
0.50
0.86
1.08
.01
Hex
1.05
0.41
3.13
0.47
0.94
1.18
0.83
0.07
2.78
0.57
0.73
0.94
.01
Angle
2.46
0.26
7.44
0.67
2.09
2.9
2.02
0.08
5.38
0.66
1.74
2.34
.08
Translation
0.49
0.01
2.87
1.12
0.37
0.64
0.45
0.00
2.24
1.07
0.35
0.57
.63
Pet tersson et al
338
Volume 103 Issue 6
Table II. Summary of total deviations at the apex and hex and per cadaver and jaw type. Deviations in millimeters. LL: lower level, UL: upper level Cadaver
Deviation Apex
Deviation Hex
Number of 95% CI 95% CI Implants Median Mean SD LL UL 1Mand
P
Median Mean SD
95% CI 95% CI LL UL
P
9
1.17
1.07
0.46
0.77
1.17
.001
0.69
0.71
0.16
0.61
0.74
.001
2Mand
9
0.96
1.03
0.36
0.79
1.11
.001
0.95
0.79
0.27
0.61
0.85
.001
3Mand
11
3.14
2.52
1.16
1.83
2.73
.001
2.46
2.19
0.90
1.66
2.35
.001
4Mand
11
1.6
1.47
0.84
0.97
1.62
.001
1.36
1.26
0.41
1.02
1.33
.001
5Mand
8
1.47
1.60
0.56
1.21
1.74
.001
1.18
1.81
0.09
1.75
1.83
.001
6Mand
11
1.16
1.32
0.54
1.00
1.42
.001
0.91
0.92
0.28
0.75
0.97
.001
7Mand
8
0.87
0.91
0.24
0.74
0.97
.001
1.06
1.04
0.21
0.89
1.09
.001
8Max
8
1.17
1.15
0.23
0.99
1.20
.001
0.91
0.99
0.20
0.85
1.03
.001
9Max
6
0.51
0.55
0.19
0.40
0.61
.001
0.54
0.56
0.19
0.41
0.62
.001
10Max
8
1.13
1.07
0.42
0.78
1.17
.001
0.85
0.81
0.19
0.68
0.86
.001
11Max
8
1.02
0.97
0.33
0.74
1.05
.001
0.66
0.62
0.15
0.51
0.65
.001
12Max
10
1.00
1.01
0.22
0.88
1.06
.001
0.57
0.57
0.25
0.42
0.62
.001
13Max
8
1.00
1.30
0.64
0.86
1.46
.001
1.78
1.90
0.62
1.47
2.05
.001
14Max
8
0.67
0.67
0.22
0.52
0.73
.001
0.86
0.81
0.14
0.71
0.84
.001
15Max
6
1.26
1.32
0.52
0.90
1.49
.002
0.95
1.10
0.35
0.82
1.21
.001
16Max
9
1.38
1.46
0.30
1.27
1.53
.001
1.46
1.48
0.12
1.40
1.50
.001
17Max
7
1.30
1.07
0.50
0.70
1.21
.001
0.79
0.67
0.37
0.40
0.77
.001
145
1.12
1.25
0.68
1.13
1.36
<.001
0.93
1.06
0.58
0.97
1.16
<.001
Total
Table III. Summary of total depth and translation deviations and per cadaver and jaw type. Deviations in millimeters. Note: Negative value for depth deviation indicates that implant did not reach planned position. Positive value indicates that implant was placed deeper than planned position. LL: lower level, UL: upper level Cadaver
Deviation Depth
Translation Deviation
Number of 95% CI 95% CI Implants Median Mean SD LL UL
P
Median Mean SD
95% CI 95% CI LL UL
P
1Mand
9
0.43
0.45
0.13
0.37
0.48
.001
0.49
0.56
0.36
0.32
0.64
.002
2Mand
9
0.55
0.6
0.27
0.42
0.66
.001
0.60
0.57
0.29
0.38
0.63
.001
3Mand
11
0.66
0.64
0.34
0.44
0.70
.001
1.43
1.52
0.81
1.04
1.66
.001
4Mand
11
0.84
0.57
0.78
0.11
0.71
.036
0.66
0.97
0.82
0.49
1.12
.003
5Mand
8
0.66
0.35
0.72
–0.15
0.53
.212
0.25
0.40
0.52
0.04
0.53
.066
6Mand
11
0.33
0.13
0.63
–0.24
0.24
.509
0.57
0.75
0.71
0.33
0.88
.006
7Mand
8
0.69
0.67
0.17
0.55
0.71
.001
0.44
0.39
0.18
0.27
0.43
.001
8Max
8
0.70
0.65
0.43
0.36
0.76
.003
0.35
0.38
0.32
0.16
0.46
.011
9Max
6
–0.21
–0.19
0.10
–0.27
–0.15
.007
0.35
0.40
0.16
0.27
0.45
.002
10Max
8
0.32
0.32
0.16
0.21
0.36
.001
0.16
0.18
0.15
0.08
0.22
.010
11Max
8
0.39
0.33
0.19
0.20
0.38
.002
0.57
0.60
0.28
0.41
0.67
.001
12Max
10
0.31
0.35
0.19
0.23
0.39
.001
0.43
0.50
0.29
0.32
0.56
.001
13Max
8
0.70
0.88
0.40
0.60
0.97
.001
0.80
1.08
0.59
0.67
1.23
.001
14Max
8
–0.54
–0.41
0.40
–0.69
–0.31
.022
0.43
0.46
0.22
0.30
0.51
.001
15Max
6
–0.2
–0.29
0.23
–0.47
–0.21
.026
0.93
1.01
0.49
0.61
1.17
.004
16Max
9
–0.92
–0.92
0.28
–1.10
–0.86
.001
1.28
1.29
0.24
1.14
1.35
.001
17Max
7
0.15
0.20
0.27
0.01
0.28
.089
0.80
0.60
0.51
0.22
0.74
.022
145
0.39
0.28
0.59
0.18
0.38
<.001
0.56
0.71
0.59
0.61
0.81
<.001
Total
The Journal of Prosthetic Dentistry
Pet tersson et al
339
June 2010
Table IV. Summary of total angular deviations and per cadaver and jaw type. Deviations in degrees. LL: lower level, UL: upper level Cadaver
Angle Deviation
Number of Implants Median
7 6 5 4 3 2 1 0 –1 –2
* *
1.16
P
9
2.54
2.82
1.54
1.81
3.16
.001
9
2.29
2.20
1.46
1.25
2.52
.002
3Mand
11
2.62
2.66
1.20
1.95
2.87
.001
4Mand
11
4.02
4.40
1.98
3.23
4.75
.001
5Mand
8
2.67
2.71
1.70
1.53
3.13
.003
6Mand
11
3.39
3.40
1.50
2.51
3.67
.001
7Mand
8
2.24
2.04
0.84
1.46
2.25
.001
8Max
8
1.60
1.84
1.18
1.02
2.13
.003
9Max
6
1.34
1.39
0.38
1.09
1.52
.001
10Max
8
2.26
2.01
1.27
1.13
2.32
.003
11Max
8
2.12
2.47
1.14
1.67
2.75
.001
12Max
10
2.64
2.67
0.67
2.26
2.80
.001
13Max
8
4.61
4.29
0.92
3.65
4.51
.001
14Max
8
2.74
2.66
0.68
2.18
2.82
.001
15Max
6
1.72
1.82
1.00
1.02
2.14
.007
16Max
9
2.05
1.88
0.70
1.42
2.03
.001
17Max
7
2.44
2.22
0.97
1.49
2.49
.001
145
2.48
2.64
1.42
2.41
2.87
<.001
Deviation Hex
* *
* * * * *
1.02
1.06
Translation Deviation
*
*
0.56
0.56
* * 0.55
Maxilla
Angle Deviation
* * *
0.87
Deviation Depth
* * *
Mandible
95% CI 95% CI LL UL
2Mand
Deviation Apex
3 2 1 0 –1 –2
SD
1Mand
Total
7 6 5 4
Mean
Mandible
2.68
2.31
Mandible
Maxilla
*
0.21
Maxilla
3 Box plot of deviations in mandible and maxilla. Angle deviations are in degrees, all other deviations are in millimeters. Note: Length of box corresponds to interquartile range. Horizontal line and numbers within box correspond to median. Plus sign indicates mean. Crosses indicate outliers.
Pet tersson et al
340
Volume 103 Issue 6 28
Number of Implants
24 20 16 12 8 4 0
0 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9
2.5
3.0
Translational Deviation 4 Histogram of translational deviation. Measurements in millimeters.
5 Translation deviation. Figure from 3Mand (Table III), with largest error in group. Planned implants are gray with threads. Postoperative implants are blue-green color. Note all implants are positioned in parallel direction after surgery compared to planned position.
DISCUSSION Based upon the results of this study, the research hypothesis cannot be accepted, as there was a statistically significant difference between the results of the mandible and the maxilla for depth deviation, apex, and hex, with the greater variation registered in the mandible. One explanation for the variation could be that the surgical template in the mandible is less stable, as it covers a smaller area compared to the maxilla. The results demonstrated a significant difference between the virtually planned implants and the actually placed implants for all 5 variables, apex, hex, depth, translation deviation, and angle. However, although the values were statistically significant,
it was not clear whether the results are clinically significant. CAD/CAM-guided implant surgery offers the clinician another method to ensure accurate and prosthetically driven implant placement. As with most new advancements, there may be limitations and risks. While the most accurate assessments will come from clinical use, this study on cadavers might add valuable information about the accuracy of guided surgery. The present study provides information about the results of the specific treatment performed, with limitations such as the disadvantage of not using the surgical index. The results could be used to improve instructions to clinicians and enable clinicians to provide safer patient treatment. In a review article including results from
The Journal of Prosthetic Dentistry
8 published studies, Schneider et al11 presented results with a mean deviation at the hex of 1.07 mm (95% CI: 0.76-1.22 mm) and a mean deviation at the apex of 1.63 mm (95% CI: 1.262 mm). The present study presents a mean deviation at the hex of 1.06 mm (95% CI: 0.97-1.16 mm) and a mean deviation at the apex of 1.25 mm (95% CI: 1.13-1.36 mm) and, thus, demonstrates similar or better results. The positioning of the radiographic guide calls for the patient to occlude onto a radiographic index and radiographic guide during the imaging procedure. In this study, the radiographic guides could not be positioned in the conventional manner. The guides had to be manually placed with the help of rubber bands. This method produced a risk of positioning the radiographic
Pet tersson et al
341
June 2010 guide in a less than ideal manner. In 12 out of 17 cadaver jaws, a space was visible between the soft tissue and radiographic guide. The traditional clinical procedure to position the CAD/CAM-guided surgical template onto the registered position of the radiographic guide could, for obvious reasons, not be performed, due to the inability to make an occlusal registration. Therefore, the surgeon manually placed the CAD/CAM-guided surgical template, which may have resulted in less accurate positioning of the implants. Implant planning was performed in an experimental manner, placing as many implants as possible with regard to the anatomical situation, considering the availability of the bone volume and the technical limitations within the system. The limitation of determining the position for the radiographic guide and surgical template in the experimental model introduced errors that should not be present clinically. To preserve the specimens, each cadaver was frozen and then thawed up to 4 times, which may have caused dehydration and a change in the size and shape of the soft tissue. These alterations of the soft tissue probably caused minor changes in the positioning of the radiographic guide and surgical template from the planned positioning. In an attempt to mimic the clinical setting as closely as possible, formaldehyde was not used. Another issue was the flexibility of the soft tissue in the cadavers, which might add further variation to the results. All of these minor differences between a living patient’s tissue and the cadaver tissue may have contributed to the errors in implant placement, in comparison to the planned implant positions. One cadaver in the present study demonstrated the maximum errors for both the mandible and the maxilla (Fig. 5). This indicates that the radiographic guide and surgical template were likely malpositioned. When reviewing the variation between the planned and the placed
Pet tersson et al
implants in the mandible (3Mand) visually (Fig. 5), it appeared that a parallel movement (translation) occurred with the placed implants compared to the planned implants. Malpositioning of the radiographic guide and the surgical template might result in errors, such as rotation and misplacement. additional research is required to obtain more information about the deviation from misplacements of the radiographic guide and the surgical template. In this study, the variations between the virtually planned implants and the implants placed after surgery were compared using matching methods similar to those presented in a study by Van Assche et al.9 This study 9 was performed on formalintreated cadavers and presented results with a range of 0.3-2.3 mm with a mean value of 1.1 mm for the hex and a range of 0.7-2.4 mm with a mean value of 2.0 mm for the apex. The current study presented results with a range of 0.07-3.13 mm and a mean value of 1.06 mm for the hex and a range of 0.12-3.63 mm with a mean value of 1.25 mm for the apex. One reason for the differences between this study and the study by Van Assche may be the number of experimental surgeries performed and included in each of the different studies. The study by Van Assche included only 4 subjects and 12 implants, while this study included 17 subjects and 145 implants. Another reason might be that tooth-supported CAD/ CAM surgical templates were used in the study by Van Assche, while in the current study, only completely edentulous CAD/CAM surgical templates were used. Tooth-supported CAD/ CAM surgical templates are probably more stable compared to those for totally edentulous subjects, adding a possible disadvantage when excluding the surgical index and, thus, causing larger deviations. The current study presents the variation obtained in completely edentulous specimens that were frozen and thawed, and without the option of using a surgi-
cal index. The results could be used to provide a better understanding of possible deviations that could occur when performing CAD/CAM-guided surgeries. This information can also be useful for clinicians, in improving this specific treatment method. In future studies, if positioning and repositioning errors are reduced, the results will depend on the surgical system limitations and errors, as well as on variations between surgeons. Additional studies are needed to learn more about the various conditions affecting virtual planning and guided surgery, and how accurate CAD/ CAM-guided surgery is compared to the freehand placement of implants. CONCLUSIONS Within the limitations of this study, the results demonstrate a statistically significant difference between mandibles and maxillae for the hex, apex, and depth measurements between the virtually planned implant positions and the positions of the implants placed after surgery with a CAD/CAM-guided surgical template on human cadavers. REFERENCES 1. Brånemark PI, Adell R, Breine U, Hansson BO, Lindström J, Ohlsson A. Intra-osseous anchorage of dental prostheses. I. Experimental studies. Scand J Plast Reconstr Surg 1969;3:81-100. 2. Jemt T, Johansson J. Implant treatment in the edentulous maxillae: a 15-year followup study on 76 consecutive patients provided with fixed prostheses. Clin Implant Dent Relat Res 2006;8:61-9. 3. Adell R, Eriksson B, Lekholm U, Brånemark PI, Jemt T. Long-term follow-up study of osseointegrated implants in the treatment of totally edentulous jaws. Int J Oral Maxillofac Implants 1990;5:347-59. 4. Lekholm U, Gunne J, Henry P, Higuchi K, Lindén U, Bergström C, et al. Survival of the Brånemark implant in partially edentulous jaws: a 10-year prospective multicenter study. Int J Oral Maxillofac Implants 1999;145:639-45.
342
Volume 103 Issue 6 5. van Steenberghe D, Glauser R, Blomback U, Andersson M, Schutyser F, Pettersson A, et al. A computed tomographic scan-derived customized surgical template and fixed prosthesis for flapless surgery and immediate loading of implants in fully edentulous maxillae: a prospective multicenter study. Clin Implant Dent Relat Res 2005;7 Suppl 1:S111-20. 6. Marchack CB. An immediately loaded CAD/ CAM-guided definitive prosthesis: a clinical report. J Prosthet Dent 2005;931:8-12. 7. Parel SM, Triplett RG. Interactive imaging for implant planning, placement, and prosthesis construction. J Oral Maxillofac Surg 2004;62 (9 Suppl 2):41-7. 8. van Steenberghe D, Naert I, Andersson M, Brajnovic I, van Cleynenbreugel J, Suetens P. A custom template and definitive prosthesis allowing immediate implant loading in the maxilla: a clinical report. Int J Oral Maxillofac Implants 2002;17:663-70.
9. Van Assche N, van Steenberghe D, Guerrero ME, Hirsch E, Schutyser F, Quirynen M, et al. Accuracy of implant placement based on pre-surgical planning of three-dimensional cone-beam images: a pilot study. J Clin Periodontol 2007;34:816-21. 10.Di Giacomo GA, Cury PR, de Araujo NS, Sendyk WR, Sendyk CL. Clinical application of stereolithographic surgical guides for implant placement: preliminary results. J Periodontol 2005;76:503-7. 11.Schneider D, Marquardt P, Zwahlen M, Jung R E. A systematic review on the accuracy and the clinical outcome of computerguided template-based implant dentistry. Clin Oral Implants Res 2009;20 Suppl 4:73-86. 12.Rade L, Westergen B. Mathematics handbook for science and engineering. 5th ed. New York: Springer; 2004. p. 85. 13.Maes F, Collignon A, Vandermeulen D, Marchal G, Suetens P. Multimodality image registration by maximization of mutual information. IEEE Trans Med Imaging 1997;16:187-98.
Corresponding author: Mr Andreas Pettersson Karolinska Institute Section for Image and Functional Odontology Department of Dental Medicine Alfred Nobels Allé 8 Box 4064 SE-14104 Huddinge SWEDEN Fax: +46 87795301 E-mail: [email protected] Acknowledgments The authors thank Matts Andersson for his contribution of ideas throughout the study, the department of Early Development at Nobel Biocare AB, Göteborg, Sweden, for their support of this study, and Filip Schutyser for his expertise in 3-D software design and function and the matching process. The authors also thank Prof Jean-François Gaudy for his support and continuous help during the study at the Laboratory of Anatomy, Odontological Faculty, University Paris Descartes, Paris, France. Copyright © 2010 by the Editorial Council for The Journal of Prosthetic Dentistry.
Access to The Journal of Prosthetic Dentistry Online is reserved for print subscribers! Full-text access to The Journal of Prosthetic Dentistry Online is available for all print subscribers. To activate your individual online subscription, please visit The Journal of Prosthetic Dentistry Online. Point your browser to http://www.journals. elsevierhealth.com/periodicals/ympr/home, follow the prompts to activate online access here, and follow the instructions. To activate your account, you will need your subscriber account number, which you can find on your mailing label (note: the number of digits in your subscriber account number varies from 6 to 10). See the example below in which the subscriber account number has been circled. Sample mailing label
This is your subscription account number
*********AUTO**SCH 3-DIGIT 001 1 V97-3 J010 12345678-9 J. H. DOE 531 MAIN ST CENTER CITY, NY 10001-001
Personal subscriptions to The Journal of Prosthetic Dentistry Online are for individual use only and may not be transferred. Use of The Journal of Prosthetic Dentistry Online is subject to agreement to the terms and conditions as indicated online.
The Journal of Prosthetic Dentistry
Pet tersson et al