I read with interest the article on 4-dimensional analysis of stomatognathtic function (Terajima M, Endo M, Aoki Y, Yuuda K, Hayasaki H, Goto TK, et al. Four-dimensional analysis of stomatognathtic function. Am J Orthod Dentofacial Orthop 2008;134:276-87). This article dealt with computer-aided simulation of jaw movement and occlusal contact analysis. Although the authors should be applauded for several aspects of their study, it has some limitations to which I want to draw attention.
The mandibular movement pattern is the most complex in the human body and is determined by both posterior and anterior guidance—ie, the temporomandibular joints (TMJs) and occlusal contacts. It is really difficult to reconstruct in a 3-dimensional (3D) way and simulate mandibular movement and occlusal contact for each patient. In general, implementation of 3D animation consists of the following steps: (1) image acquisition and digitalization of the TMJs, jaws, and teeth; (2) pathway recording for mandibular movement; (3) establishing the 3D coordinate system and determining the positions for the various digital images; and (4) movement animation by mechanical kinematic simulation and occlusal analysis by collision detection.
Computed tomography (CT) scanning is most commonly used to acquire TMJ and jaw images. Yet it is still unsettled as to data processing for TMJ images, especially for the disc. During CT scanning, it is recommended to keep the patient in the intercuspal central position (ICP), which will be convenient for further combining the images with those of the teeth in the intercuspal central occlusion (ICO). It is also important to determine the threshold spatial relationship among the various parts of digital models for mandibular movement animation and occlusal analysis.
The mandibular and maxillary teeth (casts) are digitized with a 3D laser scanner or an optical scanner. There are 3 ways to establish the ICO relationship between the maxillary and mandibular digital images. (1) Mounted in ICO relationship by hand, both casts are scanned and digitized. Then the maxillary and mandibular digital tooth images are superimposed to the images of the ICO, respectively. In this way, the spatial relationship between the digital tooth images for ICO would be determined. Nonetheless, there is some discrepancy between the intraoral and extraoral occlusal relationships by mounting casts. The discrepancy might result from error of the impression procedure and mobility of the natural teeth. (2) The second method to determine the ICO relationship is with the computer by collision detection—ie, modify the position of the images of the maxillary digital teeth to find the spatial relationship with the vast amount of occlusal contact areas. This method is time consuming, and the position obtained might not be the same as the oral condition. (3) The third way is to use some references. Attached to the teeth in the oral environment, the references are involved in CT scanning and impression-taking procedures. After digitalization of CT images, the reference images obtained from the cast images with laser scanning are superimposed. Thus, the ICO relationship is established. Also, the ICO images obtained are usually not satisfactory just at a glance, because there is a significant difference between the precision of the images from the CT scan and the laser scan.
The unified coordinate system is important to deal with the various kinds of data, and the method to establish the coordinate system is determined by the patterns of mandibular traces from the patient. The mandibular movement tracker is used to record its movement data. This tracker is based on measuring the tracks of marked points (3 or 4 transducers attached to the mandible) in correspondence to the maxilla. There are 2 ways for the transducers to be attached to the mandible. The first is to attach them to a special frame fixed to the labial and buccal sides of the dentition with adhesion; the other way is to attach them to the dentition directly. For the first method, there must be an occlusal plate (eg, the ARCUS Digma system [Kavo, Biberach, Germany] and a jaw motion analyzer [Zebris Medical Gmbh, Isny, Germany]) with a predetermined dimension in which a coordinate system for recording the jaw-tracking information is involved. During tracking recording, there is usually a procedure to register the position of the maxilla: let the patient bite at the occlusal plate with a bite-record material. Therefore, we can establish the unified coordinate system using the digital image of the occlusal plate and the coordinate system information involved. Mounted on the occlusal plate via the tooth impressions on the bite record material, the maxillary cast and the plate were scanned and digitized. Then the spatial position of the maxillary teeth in the unified coordinate system can be determined. According to the ICO relationship of the maxillary and mandibular teeth, the position of the mandibular teeth in the coordinate system can be determined, and the same for the jaws. In this procedure, scanning, digitizing, and superimposing were done repeatedly with the discrepancy generated each time. For the method characterized by attaching the transducers to the teeth directly, it is easier to establish the coordinate system from the tooth images with the transducer images.
To describe the movement of jaws, the hard tissues are always treated as rigid bodies. However, this is inaccurate to simulate the variability of biologic systems—eg, the distortion and deformation of the condyle, disc, glenoid fossa, and jaws during mandibular movement, and the mobility of the teeth during loading conditions. In computer simulation of mandibular movements, the tracks of the condyles are calculated from the tracks of several transducers. For the mandibular movement tracks record system in which the transducers are attached to the frames extraorally, the distance from the transducer to the condyle is relatively long. All the distortion and deformation in the teeth, jaws, and condyles will accumulate on the condyle tracks calculated. Therefore, there will be a discrepancy of the condyle position calculated by mechanical kinematic simulation from its actual position. Sometimes we can even see that the condyle image invades the skull base. In the record system in which the transducers are attached to the teeth directly, this will not happen. But it is worse that the relative distance between every 2 transducers will keep on changing just because of the distortion and deformation of bones and the mobility of teeth. When calculating the condyle tracks from these data, we must make some adjustments with the data, which may lead to significant discrepancies.
No matter what method is used to integrate the 3D maxillofacial skeleton images from CT scan and tooth-surface images from optical scan, there will be a significant discrepancy because of the distinguishing difference between the resolution of the two kinds of scanning systems. In image processing for integrated images, the surfaces must undergo a smoothing and removing procedure to obtain the continuous and smooth digital surface imagine required for later analysis; this will also decrease the accuracy. The integrated image obtained by this way is absolutely different from the actual spatial relationship of the original structure. The accuracy of spiral CT is about 0.2 to 0.4 mm, whereas, for cone-beam CT, it can reach 0.03 mm. The accuracy for laser scanning and surface remodeling is about 0.02 to 0.05 mm. The discrepancy from image processing might be 0.1 mm at least. The sensitivity for the mandibular movement record system is 0.1 mm. Furthermore, when determining the occlusal contact area by collision detection with the computer, we must set a distance (usually –0.05 to +0.05 mm) beforehand, just to make certain subjectively that it is the occlusal contact area if the distance between the 2 triangles from the maxillary and mandibular tooth surfaces is less than this distance. As we all know, the perceivable thickness of a tooth is about 0.1 to 0.15 mm. The systematic discrepancy of 3D reconstruction and animation of occlusal contact is apparently larger than 0.1 to 0.15 mm. For these problems, the reproduction of dynamic, excursive contacts seems to lower the reliability.
Individual mandibular movements and occlusal contacts are intricate and difficult to describe appropriately; 3D animation is the right tool for better teaching and understanding. Although the accuracy is inefficient for scientific research, there is still much work to do.