Building up a 3D virtual head for orthognathic surgery.
A scientific essay in Medical Sciences
DOCTORAL THESIS defended in public on 13th of April 2011
In 2005 the Nijmegen 3D lab was set up as a close cooperation between the department of Oral and Maxillofacial Surgery and the Department of Orthodontics and Craniofacial Biology. Within this lab three keystones have been gathered together: a close cooperation between technicians and medical researchers, the use of high-tech hardware and up-to-date software, and the availability of a considerable patient population to enable the build-up of a large database. This way, its main objective could be achieved: enhancement of the accuracy of 3D imaging and especially image fusion of the three main structures of the head. These structures, the facial soft tissue surface, the facial skeleton and the dentition, are also known as the (facial) triad. This triad plays a decisive role as it determines a patient’s facial appearance. Patients with a jaw deformity, also known as a dysgnathic deformity, such as hypoplasia or hyperplasia of the maxilla and/or mandible, sometimes benefit from orthognathic surgery. An example of orthognathic surgery is a bilateral sagittal split osteotomy (BSSO). During such a surgical procedure the lower jaw is split in three segments, which are fixated in the correct intermaxillary position and the correct position in relation to the other structures of the triad to enhance functional and esthetical disabilities. As these surgical procedures induce changes of the facial features, it is important to plan such a procedure meticulously. Previously, orthognathic plannings were based on conventional Xray’s, plain photographs and dental plaster casts. With the introduction of 3D imaging in facial surgery, it is now possible to perform these plannings with 3D photographs representing the facial soft tissues; cone-beam CT derived 3D reconstructions of the facial skeleton; and digital dental models representing the dentition. These can be virtually composed as a 3D virtual head. Three different research lines have been setup in the Nijmegen 3D lab concentrating on implantology, facial surgery / orthodontics and reconstructive surgery. As part of the second research line, the aim of this thesis was ‘improvement of the quality of 3D imaging and image fusion processes and building up a 3D virtual head for orthognathic surgery’. For this thesis four prospective clinical studies and a systematic review of current literature were performed and a technical report to setup of a 3D virtual head for orthognathic surgery was written.
The first study (chapter 2) concentrated on the facial soft tissue surfaces, captured with a 3D stereophotogrammetrical camera setup, which was the first hardware available within the 3D lab. The purpose of the study was to determine the reproducibility and the reliability of a 3D soft tissue cephalometric analysis. The evaluation was performed in a 3D environment (a virtual operating-room) and proved to be reliable. Furthermore, it was concluded that for individual patients, a surface based matching method of pre- and postoperative data is preferred.
The next study (chapter 3) concentrated on the facial skeleton to determine the added value of 3D imaging for orthognathic surgery, such as a BSSO. Within one week after surgical advancement of the mandible, the patients were scanned with a cone-beam CT scanner. The DICOM files were reconstructed into a 3D object, representing the facial skeleton. These were reviewed in the 3D virtual operating room. The 3D reconstruction clearly displayed the displacement of the three segments the split pattern between these segments. Surprisingly, four variations of the lingual split pattern were found, while the surgical procedure of the BSSO itself did not change. So, 3D imaging for orthognathic surgery added a new dimension to the discussion of BSSO techniques.
The aim of the third study (chapter 4) was to assess the positional changes of the different segments of the mandible due to a BSSO and to determine the influence of the different split patterns upon these changes. The latter were analysed with a modified 3D cephalometric hard tissue analysis using voxel based matched pre- and postoperative 3D objects. Both desired movements, such as advancement of the distal segment and unwanted movements, such as outward movement of the proximal segment occurred. Only split patterns with a larger bony contact area resulted in more malpositioning of the proximal segment. Since analysis of all three structures of the triad is best performed in one ‘holistic data set’, a shift was established from the use of individual 3D imaging techniques to methods which fuse at least two different structures of the triad into one fusion model.
The fourth clinical study (chapter 5) concentrated on image fusion of the facial soft tissue surface and the facial skeleton. In order to assess the accuracy of the surface based matching procedure, 3D photographs and 3D reconstructions of 15 patients were fused. The accuracy of this fusion model was verified by two observers. It proved to be reliable to so create a photorealistic and accurate 3D virtual representation of a patient’s face, which could be used in daily clinical practice. Documentation, analysis and treatment planning of dysgnathic patients improved with the above mentioned fusion model. Still, for correct planning of orthognathic surgery, the third structure of the triad, the dentition (including the occlusion), needed to be incorporated in the 3D data set as well.
A systematic review of the literature proved, that there were multiple methods to create a 3D virtual head with an integral fusion model (chapter 6). Nevertheless, the final conclusion of the review underlined that an integral fusion model should consist of CBCT data augmented with a textured facial soft tissue surface and a digital dental cast to elucidate dental artefacts. Moreover, none of the available integral fusion models delivered a 3D virtual head without additional burden to the patients.
Therefore, a new integral fusion model, including all three structures of the triad, was developed to create a 3D virtual head (chapter 7). The method described in the technical report proved to be a non timeconsuming, patient and user friendly method.
Finally, the results of the studies conducted are discussed in chapter 8. In summary, with these clinical studies, the review and the technical report, an improvement of the quality of 3D imaging and image fusion processes for orthognathic surgery could be accomplished.