Manufacturing splints for orthognathic surgery
using a three-dimensional printer
Marc Christian Metzger, MD, DDS, Bettina
Hohlweg-Majert, MD, DDS,
Uli Schwarz, MD, DDS,
Matthias Teschner, Beat Hammer, MD, DDS,
Rainer Schmelzeisen, MD, DDS,
Freiburg, Germany; and Aarau,
Switzerland Albert-Ludwig University
Hirslanden Medical Center
From the Winter 2009 AADMRT Newsletter
Objective. A new technique for producing splints for
orthognathic surgery using a 3D printer is presented.
Study design. After 3-dimensional (3D) data acquisition by
computerized tomography (CT) or cone-beam computerized tomography (CBCT) from
patients with orthognathic deformations, it is possible to perform virtual
repositioning of the jaws. To reduce artifacts, plaster models were scanned
either simultaneously with the patient during the 3D data acquisition or
separately using a surface scanner. Importing and combining these data into
the preoperative planning situation allows the transformation of the planned
repositioning and the ideal occlusion. Setting a virtual splint between the
tooth rows makes it possible to encode the repositioning. After performing a
boolean operation, tooth impressions are subtracted from the virtual
splint. The "definitive" splint is then printed out by a 3D printer.
Conclusion. The presented technique combines the advantages of
conventional plaster models, precise virtual 3D planning, and the possibility
of transforming the acquired information into a dental splint. (Oral Surg
Oral Med Oral Pathol Oral Radiol Endod 2008;105:e1-e7)
The success of orthognathic surgery depends on the possibility of exact
planning based on precise diagnosis. Besides the clinical examination, the
acquisition and analysis of images determines the surgical
Computerized tomography (CT) or modern techniques with less radiation
exposure, such as cone-beam computerized tomography (CBCT), are increasing in
importance in preoperative planning compared with conventional radiologic
imaging techniques, owing to the 3-dimensionality of the acquired
Digital orthodontic study models (e-models) have been shown to provide a
valid alternative to traditional plaster study models in treatment planning
for malocclusion patients.5-7 However, satisfactory
transformation of the evaluated virtual information is only possible with
great technical effort.8
Dental splints are used in orthognathic surgery with the traditional
plaster study models.9 Despite being an established
and accepted method, orthognathic model surgery theoretically suffers from
several sources of error and inaccuracy, as detailed analysis of the technique
There is insufficient control of movements such as rotation and translation
with regard to the whole cranial situation in plaster model
surgery.10,11 The purpose of the present technical
note is to present a new procedure for producing orthognathic splints
combining the advantages of conventional techniques, such as plaster models,
with modern virtual 3-dimensional (3D) planning. By using virtual blanks
within the preoperative 3D planning, the information obtained from the
repositioning can be transformed into a dental splint after printing or
In the present case, a patient with a class III anomaly is used to
demonstrate the virtual planning and surgical implementation.
After 3D data acquisition using CT or CBCT, the data are imported into
the 3D modeling software Voxim (IVS Solutions, Chemnitz, Germany;
Fig. 1). This software, commonly used for computer-aided surgery (CAS)
procedures, offers the possibility of planning surgical procedures in multiplanar
and 3D views. Therefore, segmentation, measurement, repositioning,
and importing tools are incorporated.
Fig. 1: CT scan showing the 3D skull of a patient with class
III dysgnathia. The patient is wearing brackets which results in
artifacts around the teeth.
All planning steps in CAS are based on virtual segmentation procedures, which
are necessary for performing repositioning. By using predefined Le Fort I and
Obwegeser-Dal Pont osteotomy lines, the upper and lower jaws can be segmented
Fig. 2: Dataset after segmentation of the upper (blue) and lower
(red) jaws with regard to the Le Fort I and Obwegeser-Dal Pont
Artifacts are not always avoidable owing to preoperative orthodontic treatment,
including dental brackets or dental restorations. This prevents an accurate
adjustment of the upper and lower tooth rows to determine the planned new
occlusion. To solve this problem, 2 possible solutions are presented:
Tomography. One solution involves scanning the plaster model of the upper and
lower jaws during the patient scan. A CBCT scan is shown in Fig. 3. Both models
are put together in the desired occlusion and placed over the patient's head
during the CT scan. The whole dataset is imported into the software, where a
tool is available for separating this dataset back into 2 separate sets (skull
dataset and dental model dataset), which is necessary for the further planning
steps. After segmentation of the upper and lower tooth rows of the model
dataset, a special export tool allows the generation of three Standard
Tessellation Language (STL) files of the segmented parts. Two files contain the
single tooth rows, and the third contains the determined new occlusion (Fig.
Fig. 3: Cone-beam CT dataset of plaster models of the upper and
lower tooth rows before (top left) and after (top right)
segmentation. A special export tool allows the generation of 3 STL
files: one of the segmented lower jaw (bottom left), one of the
upper jaw (bottom right), and one of the determined new
occlusion (top right).
Surface scan. Another solution consists of scanning the plaster models
using a dental surface scanner. By scanning each single model (lower and upper
jaws) and the situation of both models in the determined new occlusion, 3
scanning STL files are generated (Fig. 4).
Fig. 4: Surface scans (STL files) of the plaster models,
including the upper jaw (bottom right), the lower jaw (bottom
left), and the models in the determined new occlusion (top).
In the following step, the acquired STL files are imported into the 3D
modeling software. By defining at least 3 reference points for each virtual
tooth row and their corresponding points in the skull dataset, a precise overlap
of the generated STL files and their corresponding anatomic structures in the
skull dataset is achieved. Note that, at this planning step, the file containing
the determined new occlusion is not required, because the repositioning of the
upper jaw will be encoded by producing a preliminary dental splint (Fig. 5).
Fig. 5:The situation of the scanned models of the upper and the
lower jaws after virtual alignment to the skull dataset. Note that the
third model representing the determined occlusion is not yet included,
because the repositioning of the upper jaw will be encoded by producing
a preliminary dental splint.
An important function for further planning is the possibility of attaching the
virtually aligned models to the segmentations of the upper and lower jaws in the
skull dataset. This guarantees that movements made for the planned repositioning
in the skull dataset (upper and lower jaw in the skull dataset) will be followed
by the aligned scan models.
Definition of symmetry planes
The whole skull dataset is manually aligned to the symmetry planes using the
multiplanar view. Automatic procedures would have the disadvantage of including
asymmetrical parts of the skull in the calculations, which would lead to
inaccuracy. The midsagittal plane was used for vertical orientation, and the
Frankfurter horizontal plane was used for transversal orientation.
Repositioning: upper jaw
First, the occlusion plane of the upper jaw was defined by setting 3
points. Two points were set at the first molar teeth, and the third
was placed at the contact point between the 2 middle front teeth. By
attaching these points to the segmentation of the upper jaw, following
of the movements carried out during the repositioning of the plane is
Using the multiplanar view, it is possible to reposition the upper jaw
in each single direction. The upper jaw was rotated until the occlusion
plane was parallel to the horizontal plane in the coronal view. In the
sagittal view, an inclination was performed until it was parallel to the
Frankfurter horizontal plane. In the axial view, the upper jaw can be
moved in anterior and posterior directions until the required profile
projection is achieved (Fig. 6).
Fig. 6: Adjustment of the upper jaw (blue) considering the
occlusion plane with regard to the horizontal plane (top left),
the inclination (top right), and the required anterior
movement (bottom left). 3D visualization of the situation after
repositioning (bottom right).
It should be noted that the axial rotation of the upper jaw should not be
considered at this point, because this requires orientation of the lower jaw
(see next section).
Repositioning: lower jaw
To transform the actual position of the upper jaw to the lower, the third STL
file containing the planned occlusion is used (Fig. 7, A). By aligning the
virtual model to the repositioned upper jaw in the same way as described above,
it is now possible to move the lower jaw fragment into an adequate position
(Fig. 7, B). In the presented case, controlling the new position of the lower
jaw led to a large deviation of the osteotomy lines, which is recognizable in
the axial view (Fig. 8).
Fig. 7:A, Situation after import and adjustment of the
virtual model (cyan) representing the ideal occlusion to the
repositioned upper jaw (blue). B, Adjustment of the lower
jaw (red) using the information from the virtual model showing
the ideal occlusion.
Fig. 8:Situation of the lower jaw (red) after adjustment
with regard to the virtual model (green) showing the ideal
occlusion. A deviation of the axial rotation is visible at the mandible
angle in the region of the osteotomy (right to left).
To correct this, both segmented parts (upper and lower jaw) were linked and
rotated axially. The rotation center was placed at the front teeth, because the
front teeth appeared to be well positioned in the midsagittal plane.
First splint. To produce the first splint, the repositioned upper jaw
and the original lower jaw with their aligned STL files were used to transform
the necessary information into a virtual splint. This was selected from a
database including several different sizes of virtual "blank splints"
(Fig. 9). A correctly sized splint was selected and placed between the tooth
Fig. 9: Two virtual "blank splints," which can be used
for planning. Left: plane splint; right: cuneiform
Second splint. To produce the second splint, the repositioned upper and
lower jaws with their aligned STL files were used. A second virtual
splint was selected and also placed between the tooth rows (Fig. 10).
Fig. 10: Planned final situation after performing a slight
rotation of the upper and lower jaws. The virtual "blank
splint" (green) has been imported and set between the tooth
After performing the splint placement, the software performs a boolean
operation by subtracting the impressions of the virtual models from the virtual
"blank splint," resulting in final splints (Fig. 11).
Fig. 11: Transformation of the virtual "blank splint"
into a splint including the repositioning information of the jaws by
performing a boolean operation. Left: subtracting the virtual
upper jaw; middle: subtracting the virtual lower
jaw; right: final virtual splint with tooth impressions.
After these steps, it is possible to export the virtual splints (STL files)
to a 3D printer or a drilling device to produce the final splints.
In the presented case, a postoperative control CT scan was performed.
Using the automatic image fusion procedure, the preoperative planning
situation and the clinical outcome could be evaluated (Fig. 12).
Fig. 12: Postoperative control performed by image fusion of the
preoperative (brown) and postoperative (green) datasets,
including the postoperative planning of the upper (blue) and
lower (red) jaws.
A new technique for producing dental splints using the information
obtained by virtual preoperative 3D planning has been presented. It
combines the advantages of the commonly used splint technique and
the possibility of precise virtual 3D planning with regard to the whole
Despite the great benefits of going digital in several areas, the human sense
of touch is not yet replaceable without great effort when fine adjustment of
objects is required.12,13 By continuing to use common
plaster models to determine the ideal occlusion, the surgeon need not abandon
haptic involvement. Thereby, important details of interdigitation, occlusal
anatomy, and wear facets can be precisely transformed into the virtual
situation. This prevents the false appearance of crossbite or overjet on the
monitor screen, which is challenging and time consuming when planning is only
carried out virtually.6
On the other hand, the additional advantages of the digital planning devices
and software tools, such as accuracy, efficiency and ease of measurement, and
diagnostic procedures, are included. Simulating the operation on plaster models
is difficult owing to the lack of a real link between the cephalometric analysis
and the commonly performed plaster model surgery. Rotational and translational
movements of the plaster models are insufficiently controlled during the model
surgery stage. Finally, the splint, which transfers the final relative position
of the maxilla to the mandible, summates all of the errors of the previous
stages.10 By considering the whole skull asymmetry and
not only the situation presented by plaster models, a large anatomic region can
be included in the planning process.
Laser surface scanning and rapid prototyping is commonly used in industry and
medicine as a noninvasive alternative for generating 3D computerized images
and models. The accuracy of dental surface scanners and the rapid prototyping
procedures for orthognathic surgery is now beyond all
question.14,15 Despite the ability to only scan
visible surfaces, the advantages of ease of use, self-calibration, and
automatic image distortion correction make generating 3D computerized images
very convenient.16 Additionally, the plaster models
can be discarded after the digitalization is done. The patient records are
instantly available on the computer, and large and expensive storage areas are
no longer necessary.6 When models or splints are
needed, modern 3D printers allow fast, easy, and cheap (€0.20/cc) production
in different materials.
Limitations of this technique
In the presented case, enough space was found between the 2 tooth rows to
allow insertion of the virtual splint without moving the lower jaw. However,
in cases where an overlap occurs owing to the repositioning of the upper jaw,
a vertical opening of the lower jaw is unavoidable. Therefore, it is possible
to define a rotational axis through both mandible joints, which approximately
simulates the movement of the lower jaw. However, this simple rotation does
not really take into consideration the individual anatomic situation of each
Another limitation of this technique is the impossibility of determining
the vertical orientation of the upper jaw. Here, conventional techniques, such
as measuring the distances between the infraorbital foramen and the canine
teeth, still have to be used to determine the accurate vertical level. Using
the virtual planning software, the length values can be easily acquired after
the accurate placement of the upper jaw has been carried out. A recently
published technique includes Kirschner wires for intraoperatively detecting
the correct placement of the upper jaw.9
Hardware components such as the dental surface scanner or the 3D printer
demand large initial costs. This is the reason an alternative method for
scanning the plaster models using CBCT or CT during data acquisition from the
patient was demonstrated. When splint production is available, the virtual
splint data could be sent to an institution offering a printing service.
1. Stamm T, Meyer U, Meier N, Ehmer U, Joos U. Public domain computer-aided surgery (CAS) in
orthodontic and maxillofacial surgery. J Orofac Orthop 2002;63:62-75.
2. Wortche R, Hassfeld S, Lux CJ, Mussig E, Hensley FW, Krempien R, et
al. Clinical application of cone beam digital volume tomography in children with
cleft lip and palate. Dentomaxillofac Radiol 2006;35:88-94.
3. Ziegler CM, Woertche R, Brief J, Hassfeld S. Clinical indications for
digital volume tomography in oral and maxillofacial surgery. Dentomaxillofac
4. Macchi A, Carrafiello G, Cacciafesta V, Norcini A. Three-dimensional
digital modeling and setup. Am J Orthod Dentofacial Orthop
5. Whetten JL, Williamson PC, Heo G, Varnhagen C, Major PW. Variations in
orthodontic treatment planning decisions of class II patients between virtual
3-dimensional models and traditional plaster study models. Am J Orthod
Dentofacial Orthop 2006; 130:485-91.
6. Stevens DR, Flores-Mir C, Nebbe B, Raboud DW, Heo G, Major PW. Validity,
reliability, and reproducibility of plaster vs digital study models: comparison
of peer assessment rating and Bolton analysis and their constituent
measurements. Am J Orthod Dentofacial Orthop 2006;129:794-803.
7. Miller RJ, Derakhshan M. Three-dimensional technology improves the range
of orthodontic treatment with esthetic and removable aligners. World J Orthod
8. Mischkowski RA, Zinser MJ, Kubler AC, Krug B, Seifert U, Zoller
JE. Application of an augmented reality tool for maxillary positioning in
orthognathic surgery-a feasibility study. J Craniomaxillofac Surg
9. Gil JN, Claus JD, Manfro R, Lima SM Jr. Predictability of maxillary
repositioning during bimaxillary surgery: accuracy of a new technique. Int J
Oral Maxillofac Surg 2007;36:296-300.
10. Olszewski R, Reychler H. [Limitations of orthognathic model surgery:
theoretical and practical implications.] Rev Stomatol Chir Maxillofac
11. Yosano A, Yamamoto M, Shouno T, Shiiki S, Hamase M, Kasahara K, et
al. Model surgery technique for Le Fort I osteotomy-alteration in occlusal
plane associated with upward trans- position of posterior maxilla. Bull Tokyo
Dent Coll 2005; 46:67-78.
12. Noroozi H. Orthodontic treatment planning software. Am J Orthod
Dentofacial Orthop 2006;129:834-7.
13. Noroozi H. Introduction of a new orthodontic treatment planning software;
a fuzzy logic expert system. Int J Orthod 2006;17:25-9.
14. Faber J, Berto PM, Quaresma M. Rapid prototyping as a tool for diagnosis
and treatment planning for maxillary canine impaction. Am J Orthod Dentofacial
15. Vlaar ST, van der Zel JM. Accuracy of dental digitizers. Int Dent J
16. Kusnoto B, Evans CA. Reliability of a 3D surface laser scanner for
orthodontic applications. Am J Orthod Dentofacial Orthop 2002;122:342-8.
aDepartment of Craniomaxillofacial Surgery,
Albert-Ludwig University Freiburg.
bComputer Science and Computer Graphics Department,
Albert-Ludwig University Freiburg.
cHirslanden Medical Center.
Received for publication Apr 29, 2007; returned for revision Jul 16, 2007;
accepted for publication Jul 23, 2007.
© 2008 Mosby, Inc. All rights