| Objective: The aim of this study wasto evaluate the three-dimensional cephalometric measurement and image fusion techniques using three dimensional digitized models. Partial least squares(PLS) method are used for predictions of hard and soft tissue changes after mandibular advancement, setback, distraction osteogenesis surgeries, and its accuracy was comparedwith the ordinary least square(OLS) method.Part One: Establishme nt of digital cephalometric model using image fusion technology.Materials: Eligible patients diagnosed of jaw deformity and underwent single or double jaw operation in our hospital were enrolled. The following image acquisition devices were used:(1) GALILEOS CBCT system(Sirona Dental Systems, Inc. Charlotte, USA)(2) 3d MD scanning device(3d MD, Gorgia, Atlanta, USA) Methods:1. Collection of 3D facial images; 2.Digital calibration of anatomical landmarks 3.Fusion of C BCT and 3d MD images. Results: the computer simulation of soft tissue image consists of a large number of damped mass spring model and at least 500000 nonlinear spring model polygon connections. Real physical properties of soft tissues and deformation of 3D nodes can becalculated and simulated by the spring damper model, including face, edge tetrahedral, etc.However, due to the nature of soft tissue, the elasticity parameters modulus can only be used as reference for prediction of soft tissue deformation. Conclusion: Model of CBCT and 3d MD image fusion proved to be successful and accurate.Part Two: Evaluation of precision of soft tissue alteration predictions with computerized cephalometry and image fusion techniqueMaterials:Five adult patients were enrolled in the study(2 women and 3 men). Enroll eligibility and other materials were same as part one. Methods: A total of 12 soft tissue cephalometric landmarks and 10 hard tissue landmarks were selected, soft tissue were marked one week before operation using the Simplant O&O software to generate a CBCT and 3d MDreference image. All facial region were marked with selected anatomical landmarks. The average differences of anatomical landmarks were calculated automatically, and difference of anatomic landmarks in the horizontal direction(X axis) and the vertical direction(Y axis) vector were analyzed by statistical calc ulation of Student-t test, and 95% was chosen as the confidence interval. Results: the soft tissue image and 3d MD image was superposed on the CBC T images, the anatomical landmarks were selected between two images in the vertical and horizontal vector difference, reflecting the fusion image precision. The results showed that error range of soft tissue measurement in vertical direction was between 0.01 mm ~ 2.32 mm, the average absolute error was 1.27 mm. Among the 12 soft tissue sites in the vertical direction, the maximum soft tissue error was stomion superius(1.33mm), minimum error was nasal point(0.02mm); in the horizontal direction, the maximum soft tissue error was stomion superius(1.02mm), the minimum error was nasal point(0.01mm). Conclusion: the stud y results showed that the precision of 3d MD optical scanning and CT fusion was satisfied. Region of other anatomic landmark errors wereless than 1mm. Average error between soft tissue landmarks the was 0.27 mm.Part three: Models of prediction of soft tissue change after orthognathic surgery using partial least squares analysisMaterials: 22 patients(9 women, age: 17-27 years old, mean age 21 years; 13 males, age:19-24 years old, mean age 22 years) were enrolled in the study, the patients were diagnosed as severe malocclusion and orthognathic surgery was required for complete correction. Methods: a total of 24 prediction variable sets(X matrix) was setup. The predictive variables for x1...Xp...XP was setrespectively, the type and amount of the patient’s age, gender, orthognathic operation(single, double jaw, genioplasty) and postoperative time was integrated in the agorithum. Valuesof preoperative bone and soft tissue anatomy landmarks including 16 bone landmarks and 28 soft tissue landmarks were measured.Values of 48 soft tissue and bone landmark values in vertical and horizontal directions were calculated. The 64 dependent variablesand 32 soft tissue landmarks(y... Yr... YR) were included in the X and Y axis. Results: Two-stage partial square analysis was used, the results of ANOVA showed that the soft tissue landmarkswere interconnected(P= 0.039). PLS and OLS had significant statistical difference in the horizontal and vertical vector(horizontal vector Sig.: nasal point 0.03, nasal septum <0.02, upper and lower lip <0.0001, B point 0.03, POG 0.04, Menton 0.0259, R point <0.0001.Landmarks in the vertical vector: nasal point <0.0001, nasal septum <0.0001, the upper lip 0.0002, lower lip 0.0004, B pint 0.03, POG 0.04, Menton 0.0259, R <0.0001. Considering the multiple soft tissue between the actual and predicted values of changes, the 95% confidence ellipse scatterplot was used for analysis. The soft tissue anatomy landmarks are compared using both PLS and OLS methods. The 95% confidence interval showed that PLS was significantly higher than that of OLS by the measn of elliptic boundary. Conclusion: the prediction performance of PLS was better than OLS.Part four: Partial Least Squares(PLS) pre diction of the soft tissue changes after Mandibular Setback Surgery(MSS)Materials: subjects consisted of 23 skeletal class III malocclusion patients(8 men and 15 women, mean age 24.5 years, age range 19.5-33 years old, some subjects were the same as experiment two). Inclusion criterion: subjects were of no facial a symmetry; all MSSswere done by the same surgeon; all underwent bilateral sagittal split ramus osteotomy(B point preoperative average value: 6.2 + 2.5mm). Methods: C BCT and 3d MD images of patients underwent MSS were collected6 months before and after operation. PLS algorithm wasused to calculate the predictions ofthe landmarks alterations. Results: the average value of the mandible setback was 3.3 + 1.6mm(gnathion). The change of preoperative and postoperative valuesof the nasal base, upper lip, lower lipand soft tissue pogonion were listed. The maximum alterations were observedat lip point(3.51 + 0.7mm), and foremost point was the subnasal(1.62 + 0.3mm). Result: our study showed that after MSS, there was significant different between the actual value and the PLS value at lower lip andpoint B.(P<0.001). The rest of the anatomical landmarks predictive values and the actual valuespresented no statistical differences. Conclusion: PLS for mandibular setback surgery proved to be valid, but its accuracy was poor in the peristomatous regions.Part Five: Accuracy of partial least squares(PLS) on the predictions ofsoft tissue changes after mandibular advancement(MAS)Materials: the cohort consisted subjects with skeletal class II malocclusion. Inclusion criterion: subjects without facial asymmetry; 2.Subjects underwent MASby the same surgeon; 3.Bilateral sagittal split ramus osteotomy and advancement was adpoted. Methods: Same as experiment four. Results: anatomic landmark alterations were observed after mandibular advancement(three-dimensional cephalometric values). Mandibular advancement mean value was 4.1 + 1.5mm(gnathion). The change between value under the nose, upper lip, lower lip points midpoint midpoint, lips and soft tissue pogonion preoperative and postoperative values are listed. The maximum movement was observed in the nasalspinale(3.51 + 0.7mm), while the minimum movement was observed in the midpoint in the lower lip(1.62 + 0.3mm).Part six: Prediction of soft tissue accuracy using partial least squares(PLS) on mandibular alveolar ridge distraction osteogenesis followed by implant borne prosthesis.Materials: the patient cohort: subjects were diagnosed of trauma or tumor; subjects underwent alveolar bone distraction mandibular distraction.After 5 days of delay period, distraction was applied at 1mm / day, untilthe desired mandible height reached. C BCT and 3d MD images were taken and soft tissue alterations were measured at following time points:DO reached the ideal height(T1), distractor dismantled(6 months after theoperation, T2) and implant restoration is completed(1 years, T3). Image collection method was same as forementioned. Labrale superius, labrale inferius, subnasale, sublabiale and soft tissue pogonion and lower 1/3 face height changes of after the operation were calculated and compared with PLS prediction. Paired sample t test(p =0.05)was used to evaluate the average absolute error, standard deviation, and 95% confidence interval. Results; study results showed that during the DO retention period(T1), PLS efficiency on the change of soft tissue prediction is poor. The subnasal point-the labrale superius(P = 0.05), lower lip mouth point(p=0.003), point B(p=0.002), soft tissue pogonion(p=0.02), point B(p=0.01), the nasal soft tissue pogonion point(p=0.04) PLS prediction values are statistically different from the actual value. The rest of the anatomical landmarks predictive value and the actual values have no obvious statistical differences. Six months after DO, the lower lip, mouth point- the midpoint(p=0.001), point B(p=0.001 PLS) of the predicted value and the actual value of the showed significant differences. The rest of the anatomical landmarks predictive value and the actual value showed no statistical significant difference. |
PDF Full Text Request