| Maxillofacial defects or deformities seriously damaged facial appearance, affectedmental status, and reduced quality of life. It was of importance to make appropriate andimmediate repairment on these defects. The current repairing methods mainly includedsurgery and prosthesis. For example, surgical intervention might be a preferred approachfor patients with minor defects, while prostheses repairment might be more beneficial forthose with extensive or complicated defects. With the recent development ofcomputer-aided design and computer-aided manufacturing (CAD/CAM), the technologyhas been applied into maxillofacial prosthetics. The CAD/CAM had advantages over thetraditional method and opened up a new way to better outcome. During the production ofprosthesis with CAD/CAM, it is essential to determine an appropriate facial reference toachieve a natural and harmonious prosthesis. Previous studies showed that natural headposition reflects true appearance of an individual in real life，and is considered as anoptimal reference for facial evaluation.In most of studies on recording and adjusting NHP, investigators used additional headdevices to help achieving NHP, which, to some extent, might compromise the process the other way around. Herein, the present study reported a simple, easy-to-use approach toachieving NHP without attached facial devices. First, we took the full-face photographs ofa head of a relaxed subject who stand and focused on an infinite horizon. Then we usedthose reference photographs to calibrate three-dimensional surface images taken by3dMDface scanner. In the end, we used the adjusted3-dimensional surface images to correct CTmodels.The main experimental procedures were presented as follows:1. A self-designed cartridge was made of acrylic,250mm in height,200mm in widthand300mm in length, and fixed on a tripod. A one-way mirror,250mm in height,200mmin width, was set as the frontal wall of the cartridge. Based on the cartridge, aphotographing system was set up to aid the achievement of NHP in frontal view.2. The frontal facial photographs of42volunteers (male:26, female:16) were obtainedusing the established imaging systems at baseline,1h,1d,3d and7d, respectively. Afteradjustment of the photographs with a plumb line as reference, we measured the anglesbetween the line connecting two outer canthus and horizontal line (positive for clockwisedirection, negative for counter clockwise direction). Then we tested angles of each timepoint using on-way ANOVA and Bonferroni’s multiple tests, and tested angles betweendifferent sex groups using Mann-Whitney U test. P value less than0.05was used forsignificance.3. We established a method to correct3-dimensional surface images according to thefull-face photographs of NHP. Specifically, we calculated the3-dimensional coordinates offour marked points on the surface of facial images using Geomagic Studio software, andthen used the points to correct3-dimensional facial images to NHP. Using the method, wecompared angle difference of the corrected3-dimensional facial images between baselinevs.1h,1d,3d, or7d respectively.4. We used two methods to correct CT-reconstructed,3-dimensional head model basedon the corrected3-dimensional facial surface images; one is the correction based on stablefacial marks and target registration; the other is based on general appearance and globalregistration. We chose the relatively recognizable facial marks (e.g., inner or outer canthus, basis nasi) to help match and correct the3-dimensional CT model, which were not easilydeformed in supine position during the CT acquisition. The difference between thecorrected CT models was3-dimensoinally compared.The main research results were presented as follows：1. We have designed and made a cartridge with a one-way mirror wall, built a novelphotographing system, and established a method to obtain full-face photographs. Usingthis system, we could easily record the NHP of a given subject in frontal view after a fewsimple steps such as correction by a plumb line.2. We have evaluated the full-face photographs of42volunteers at different timepoints obtained by the above imaging system, resulting in that no statistically significantangle difference was observed between groups at different time points or between male vs.female groups. It was indicated that our imaging system is reproducible, and the full-facephotographs may be helpful for correcting3-dimensional images as a useful reference.3. We have established a novel method to correct3-dimensional images of randomhead orientation to NHP using2-dimensional photographs of NHP in frontal view. Toequalize the2-dimensional and3-dimensional facial images, we induced a scalingcoefficient (p), which is calculated as follows:(a2-b2-c2) p2±2(m2-b2p2)(n2-c2p2)+(m2+n2-l2)=0. We also wrote aprogram to automatically calculate all required statistics. Our study showed that the angledifference was significantly increased at longer time points. The increment peaked at3dbut slowed down at longer time point (7d), which was similar to the results from otherstudies. Our method might be a feasible and useful tool to study the reproducibility of3-dimensional NHP imaging and correction.4. We have used two different methods to correct3-dimensional CT model accordingto full-face images of NHP. Theoretically, the method based on stable facial marks wasmore precise than that based on general appearance. The study, however, showed thatthere was no significant difference between the two methods. Because the later methodwas relatively simple and easy to implement, it would be more useful and practical in the correction process of a CT head model. |
PDF Full Text Request