If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
College of Medical, Veterinary and Life Sciences, Medical Faculty, Glasgow University School of Dentistry, BACS research group, 378 Sauchiehall Street, Glasgow G2 3JZ, UK
College of Medical, Veterinary and Life Sciences, Medical Faculty, Glasgow University School of Dentistry, BACS research group, 378 Sauchiehall Street, Glasgow G2 3JZ, UK
Centre for Ultrasonic Engineering, Department of Electrical and Electronic Engineering, University of Strathclyde, 204 George Street, Glasgow G1 1XW, UK
College of Medical, Veterinary and Life Sciences, Medical Faculty, Glasgow University School of Dentistry, BACS research group, 378 Sauchiehall Street, Glasgow G2 3JZ, UK
We present a new method for replicating the skull and occlusal surface with an accurate physical model that could be used for planning orthognathic surgery. The investigation was made on 6 human skulls, and a polyvinyl splint was fabricated on the dental cast of the maxillary dentition in each case. A cone beam computed tomogram (CBCT) was taken of each skull and a three-dimensional replica produced. The distorted dentition (as a result of magnification errors and streak artefacts) was removed from the three-dimensional model and replaced by new plaster dentition that was fabricated using the polyvinyl splint and a transfer jig replication technique. To verify the accuracy of the method the human skulls and the three dimensional replica model, with the new plaster dentition in situ, were scanned using a laser scanner. The three-dimensional images produced were superimposed to identify the errors associated with the replacement of the distorted occlusal surface with the new plaster dentition. The overall mean error was 0.72 and SD was (0.26) mm. The accuracy of the method encouraged us to use it clinically in a case of pronounced facial asymmetry.
Various dental articulators are currently available for planning orthognathic surgery, ranging from simple hinge articulators for single jaw procedures to semiadjustable articulators with facebow systems for more complex procedures.
Semiadjustable articulators have been designed for prosthodontic purposes, for which they have proved effective. However, there have been well-documented inaccuracies when semiadjustable articulators have been used for planning orthognathic surgery and fabrication of occlusal wafers.
The main problem has been with computed tomography (CT), which creates streak artefacts as a result of metallic dental restorations and orthodontic brackets.
In addition, the teeth themselves do not replicate accurately because of their complex structure and resultant beam hardening.
A recent study attempted to deal with the streak artefacts produced during CT in patients with orthodontic appliances in situ. In 3 cases the alveolar arches of the maxilla and mandible of the three-dimensional model were replaced with plaster dental models. The composite model was then used for planning orthognathic surgery.
Despite the fact that life-size, three-dimensional models of the patients were produced, the accuracy of replacing the alveolar arches with dental casts was not assessed and the reliability of the technique was not verified.
Computerised three-dimensional virtual planning for orthognathic surgery has recently been introduced.
The technique depends on accurate replacement of the distorted dentition resulting from streak artefacts with a digitally-produced dental model from the scanned impression. This allowed prediction planning, and fabrication of a digital wafer that was converted to a physical one to guide surgical movements of bony segments. The method has obvious potential benefits, but it requires sophisticated software packages, expensive hardware, and advanced expertise in the manipulation of images that might not readily be available in most oral and maxillofacial surgery units.
The purpose of this paper is to describe a technique for the successful removal of the distorted maxillary dentition from the three-dimensional printed skull model and its replacement with a plaster cast made from direct impressions of the natural teeth. This research study presents evidence of the accuracy and the validity of the new method for planning orthognathic surgery that has been lacking from previously published data.
Material and methods
Direct impressions using alginate (Alginot™ Kerr Corporation, Romulus, USA) were taken of the maxillary dentition of 6 human skulls, and dental casts were produced using a class IV dental stone (Shera Hard Rock, Shera, Germany). On each plaster maxillary dental cast a 1 mm polyvinyl splint was fabricated, covering all surfaces of the teeth using a pressure forming machine (Erkodent Erkopress Es-200E, Germany).
An 0.4 mm voxel resolution CBCT scan was taken of the human skulls using an iCAT machine (Imaging Sciences International, Hatfield). The DICOM (Digital Image Communications in Medicine) file format was converted using Maxilim software (Medicim, Belgium) into Standard Tessellation Language (STL) format. Finally, a Z-Corp 310 Plus three-dimensional printer (Z-corp, Burlington, USA) was used to produce the printed skull models.
Each three-dimensional printed skull model was attached to a platform with a 90° vertical column connected to a halo frame. Two locating plates on each side were positioned on to the maxilla, one at the zygomatic buttress and one at the pyriform aperture. They were subsequently removed before the maxilla was sectioned at the Le Fort I level.
A denture relining jig (Dentsply, UK), which consisted of upper and lower sections that were joined by three parallel columns of fixed vertical height, was used to transfer the dentition. The dental alveolar process of the sectioned three-dimensional printed skull model was located upside down in the lower section of the jig using silicone putty (Coltène/Whaledent AG, Altstätten, Switzerland) (Fig. 1). Before the jig was closed the top section of the device was filled with silicone impression material, and an impression was taken of the maxillary dentition of the skull model. Once the material had set the jig was opened, and the dentition of the three-dimensional printed skull model removed. The polyvinyl splint, which had been fabricated on the human skull's dentition, was placed into the indentations of the teeth within the silicone impression, which was attached to the upper section of the jig (Fig. 2). Plaster was poured inside the fitting surface of the splint to reproduce the dentition. Cold cure acrylic resin (Ortho-Care, UK) was placed between the plaster base and the maxilla to join them together. The jig was then reassembled and bolted down to ensure that the dentition was orientated in an accurate position and to maintain the vertical dimension. The jig was reopened and the splint removed, leaving the accurate plaster dentition attached to the dental alveolar process of the three-dimensional printed skull model. The dental alveolar process of the skull bearing the plaster teeth was then removed from the transfer jig, and returned to the three-dimensional printed skull model using the locating plates and screws (Fig. 3).
Fig. 1Transfer jig showing the dental alveolar process located before removal of dentition.
A FARO laser scanner (Scantec, Coventry, U.K.) was used to scan both the natural skull and dentition and the three-dimensional printed skull model with the plaster dentition in place. The scanner allowed three-dimensional surface capture with an accuracy of 0.025 mm according to the manufacturer's specifications. The laser-scanned three-dimensional images were then imported into software that was capable of generating the x, y, and z co-ordinates of a specific point or operator-defined landmark (VRmesh software, Seattle City, USA). Thirteen landmarks were digitised on each skull, 6 landmarks on the incisal edges and cusp tips of the teeth, and the other 7 on the vault of the skull. For each case the three-dimensional printed skull model and the human skull were digitised twice, one week apart; this produced 4 sets of three-dimensional co-ordinates of the 13 landmarks: human 1, printed 1, human 2, and printed 2.
The three-dimensional configurations of the 13 landmarks were superimposed using Procrustes superimpostion.
The following groups of superimpositions were carried out: dentition only (6 landmarks) – human skull dentition and plaster dentition; the vault of the skull (7 landmarks) – human skull vault and vault of three-dimensional printed skull model; and the dentition and the vault of the skull (13 landmarks) – both entire models.
Clinical application
The technique was used for a patient with serious facial asymmetry as a result of right condylar hyperplasia (Fig. 4). A life-sized three-dimensional model of the skull and dentition was printed using rapid protyping. The dentition was replaced with dental plate models using the new technique (Fig. 5). Model surgery was planned to correct the maxillary occlusal canting and to estimate the magnitude of condylectomy required to correct the mandibular asymmetry. Simultaneous sagittal split osteomy of the opposite side was planned to achieve the best possible occlusion and aesthetic improvement (Fig. 6).
Fig. 4A case of facial asymmetry as a result of right condylar hyperplasia.
Fig. 6Planning of model surgery, which consisted of Le Fort I osteotomy to adjust occlusal canting, 18 mm right condylar shave, and sagittal split osteotomy of the opposite side.
After superimposition of the images in each group, the mean distance between the landmarks was recorded (Fig. 7). The blue circles on the horizontal line marked “dentition” indicate the mean error among the 6 landmarks of the human skull dentition and the plaster dentition following superimposition. The overall mean (SD) error was 0.55 (0.37) mm.
Fig. 7Blue circles indicate the mean error between the landmarks of human skull dentition (6 landmarks) and vault (7 landmarks), and landmarks of the plaster dentition and three dimensional printed skull vault after superimposition. Red circles indicate the mean error between the landmarks of the dentition and vault when all 13 landmarks are used. Vertical columns indicate the number of skulls; horizontal rows indicate the superimposition sequence of the skulls.
The blue circles on the horizontal line marked “vault” indicate the mean error among the 7 landmarks of the human skull vault and the printed skull vault after superimposition. The overall mean (SD) error was 0.72 (0.26) mm.
The entire human skull and three-dimensional printed skull model with the plaster dentition based on the 13 landmarks were superimposed and the mean distances of the 7 landmarks of the vault (red circles/vault) and 6 landmarks on the dentition (red circles/dentition) were measured. This overall superimposition produced larger errors. The overall mean (SD) error for the dentition was 0.74 (0.37) mm and 0.83 (0.27) mm for the vaults.
As well as the landmarks, it was also possible to superimpose the laser-scanned images of each of the human skulls over the three-dimensional printed skull model. The models were aligned initially using rigid registration, followed by further alignment using the iterative closest point algorithm. Only the 7 landmarks of the vaults were used for superimposition to eliminate the effect of the dentition. Fig. 8 shows an example of a superimposed skull with the range of values set between 1 mm and −1 mm. The green parts of the images, such as the infraorbital rims and superior parts of the zygoma, are aligned to within 0.1 mm. However, the red shown on the laterally and buccal surfaces of some of the teeth indicates a larger error, between 0.8 and 1.0 mm.
Fig. 8Human skull mesh superimposed only on the vault of the three-dimensional skull printed model with plaster teeth mesh using VRmesh. The distance between the meshes is indicated by the colour, with red being the least accurate and blue being the most accurate.
We have described a method for the accurate transfer of a plaster dentition to a three-dimensional printed skull model. The technique was used to replace the maxillary teeth, but the same technique could be used to replace the mandibular dentition. The sample size used was relatively small and it was done on human skulls. Further investigations involving patients, including clinical cases of varying complexity, are required to assess the applicability of the technique.
The mean placement error of the dentition was 0.19 mm with a mean placement error for the skull vault of 0.11 mm. However, the mean superimposition and digitisation errors were 0.55 mm for the dentition and 0.72 mm for the vault. These results suggest that errors in superimposition and the digitisation of the landmarks are much greater than the dentition placement technique error itself. If digitisation and superimposition of the landmarks could be improved, therefore, then this should have an appreciable effect on the precision of the technique. With the present levels of accuracy preliminary studies have shown that this replacement technique could be clinically acceptable.
At present most surgeons use lateral cephalographs and dental casts for planning orthognathic surgery. This provides little information about any associated deformities of the jawbones, including condylar abnormalities, mediolateral facial asymmetries, dysmorphology of the chin, or abnormalities of the inferior border and ramus of the mandible. The inclusion of selected portions of the skull and jawbones would improve the accuracy of surgical planning, and would also negate the difficulties associated with the use of the facebow. The credibility and application of the new method have been illustrated by the case presented. It would have been impossible to predict the magnitude of the required condylectomy in three dimensions using standard articulators.
The method has the potential to save operating time, as the bony plates could all be preformed on the three-dimensional skull model preoperatively. This is more applicable in patients with complex craniofacial problems and pronounced asymmetry. To obtain an accurate replica of the patient's skull and dentition would be beneficial in explaining the procedure, and would be of educational value.
We support the concept that the use of orthognathic anatomical replicas of the patients’ jaws should allow surgeons and technocrats to plan orthognathic operations more accurately in the future.
Recent studies have reported “virtual orthognathic planning” and the use of digital wafers. These have been milled and used as a template for repositioning the jaws. However, there are still benefits from a hand-held three-dimensional skull model that allow surgeons to have feedback during surgical planning.
Three-dimensional skull models are expensive compared with the techniques currently available, and rely on experienced personnel to operate the necessary equipment for their production. The cost of this technology is, however, becoming less, which makes it more affordable for larger specialist units. Smaller units and departments could liaise with larger ones to make use of this technology and make it more cost-effective, and this would add another dimension to improving the accuracy of surgery.
Funding
This study was partially funded by the Biotechnology And Craniofacial Sciences (BACS) research group, Glasgow University Dental School.
Acknowledgement
We would like to thank Prof. A. Bowman, Dr. J Whitters, and Ms. A Holms for their help in this study.
References
Ellis III, E.
Tharanon W.
Gambrell K.
Accuracy of face-bow transfer: effect on surgical prediction and postsurgical result.
00365-7&id=gr1.jpg){kind=link}
00365-7&id=gr2.jpg){kind=link}
00365-7&id=gr3.jpg){kind=link}
00365-7&id=gr4.jpg){kind=link}
00365-7&id=gr5.jpg){kind=link}
00365-7&id=gr6.jpg){kind=link}
00365-7&id=gr7.jpg){kind=link}
00365-7&id=gr8.jpg){kind=link}