ORIGINAL ARTICLE

Comparison of soft-tissue proles in Le Fort I osteotomy patients with Dolphin and Maxilim softwares
Nasser Nadjmi,a Azita Tehranchi,b Niloufar Azami,c Bahram Saedi,d and Wouter Mollemanse Antwerp and Leuven, Belgium, and Tehran and Qazvin, Iran

Introduction: To correct dentofacial deformities, a combination of orthodontic treatment and orthognathic surgery is needed. Prediction software packages are benecial in treatment planning and achieving improved outcomes, but before using any software, its reliability and reproducibility must be assessed. The aim of this study was to evaluate the accuracy of 2-dimensional Dolphin (version 10; Dolphin Imaging & Management Solutions, Chatsworth, Calif) and 3-dimensional Maxilim (Medicim, Sint-Niklaas, Belgium) softwares in predicting the softtissue proles of patients who had Le Fort I osteotomies. Methods: The presurgical and postsurgical cone-beam computed tomography synthesized lateral cephalograms of 13 patients were collected. Using the Dolphin and Maxilim softwares, the postsurgical proles were predicted. The positions of the soft-tissue landmarks in prole views were compared with landmarks in the postsurgical photographs. The data were analyzed with the coefcient of reliability and paired-sample t tests. Results: The alpha values of the interclass correlations for each landmark in the x and y planes were between 0.96 and 0.99, except for stomion superior in Maxilim (0.83). The 95% condence interval and the absolute mean of the error showed that errors in the Dolphin software were greater than those in the Maxilim software, but the differences were not signicant (P .0.05), except for soft-tissue A-point. The greatest errors were seen in the chin region. The prediction errors of the nasolabial and mentolabial angles were greater; the prediction error in the Dolphin software was 9 , which has clinical signicance. Conclusions: The Dolphin and Maxilim softwares are both appropriate for clinical use. Their inaccuracies in the prediction of the chin region should be considered in complicated surgical planning. (Am J Orthod Dentofacial Orthop 2013;144:654-62)

orrecting dentofacial deformities needs a combination of orthodontic and surgical approaches.1 Assessment of the results is based not only on the correction of malocclusion, but also on the esthetic improvement of the face.2 Facial appearance is a key factor in a person's social life, and people are sensitive to facial changes after orthognathic surgery.3 Lateral cephalometric analysis and computer predictions are helpful

a Professor and coordinating program director, and oral and maxillofacial surgeon, University of Antwerp; director, team for cleft and craniofacial anomalies, Antwerp, Belgium. b Associate professor, Medical School, Shahid Beheshti University, Tehran, Iran. c Assistant professor, Department of Orthodontics, School of Dentistry, Qazvin University, Qazvin, Iran. d Assistant professor, Medical School, Shahid Beheshti University, Tehran, Iran. e Research engineer, Medical Image Computing, Katholieke Universiteit, Leuven, Belgium. All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conicts of Interest, and none were reported. Address correspondence to: Niloufar Azami, Department of Orthodontics, School of Dentistry, Qazvin University (Medical Sciences), Bahonar Blvd, Qazvin, Iran; e-mail, dr.azami@yahoo.com. Submitted, December 2012; revised and accepted, June 2013. 0889-5406/$36.00 Copyright 2013 by the American Association of Orthodontists. http://dx.doi.org/10.1016/j.ajodo.2013.06.019

for quantication of treatment plans, presurgical orthodontic preparations, comparison of different treatment plans, improved surgical outcomes, communications with other members of the interdisciplinary team, and database management. It is crucial to assess the feasibility of the treatment plan to optimize case management.1,4,5 Several software packages are designed for prediction in orthognathic surgery. The 2-dimensional (2D) softwares are based on lateral cephalometrics and prole photographs, including Dentofacial Planner (Dentofacial Software, Toronto, Ontario, Canada), OPAL, Quick Ceph (Quick Ceph Systems, San Diego, Calif), TIOPS, and Dolphin (Dolphin Imaging & Mangement Solutions, Chatsworth, Calif) that simulate changes in the anteroposterior and vertical facial planes.1 The main shortcomings of these programs are the limitations of the algorithms in the prediction of soft-tissue changes and their inability to predict changes in the transverse plane.6 Three-dimensional (3D) prediction planning is based on cone-beam computed tomography (CBCT) or conventional computed tomography scans for 3D reconstruction of the skull and prediction

654

Nadjmi et al

655

in 3 planes of space.7 In addition, it is feasible to generate 3D facial models by color texture mapping.8 Three-dimensional systems are now becoming more popular because of their potential in presenting more realistic soft-tissue predictions. However, before applying prediction software in clinical practice, qualitative and quantitative validations are required; furthermore, it is important to dene the magnitude, direction, and location of errors to evaluate the clinical impact of these errors.9 Several studies have assessed the accuracy and reliability of 2D prediction softwares, but the results were often conicting, and most of these inaccuracies and controversies are related to soft-tissue predictions.1,2,10-20 There are limited studies evaluating Dolphin prediction accuracy; in most of them, only hard tissues were evaluated.17,21-23 For 3D prediction softwares, there is even less evidence regarding their validity and reliability. Maxilim software (Medicim, Sint-Niklaas, Belgium) is a 3D image-based environment for assessing the head's anatomy and for preoperative planning for maxillofacial surgery and visualizing the expected soft-tissue outcome. In contrast to the Dolphin software, which uses linear ratios for the soft-tissue response, the Maxilim software uses computational strategies that can model the deformation behavior of the soft tissues to predict the patients postoperative look. Maxilim is based on a biomechanical simulation model that does not use hard-tissue or soft-tissue movement ratios to predict; instead, it mimics the elastic deformation behavior of the soft tisuues.4 The aim of this study was assess whether a 3D orthognathic planning package could predict the changes in the soft-tissue prole more accurately than 2D programs.
MATERIAL AND METHODS

The sample consisted of 13 white adult patients over the age of 20 years (11 women, 2 men) who had orthognathic surgery in Monica Hospital in Antwerp, Belgium, in 2008 and 2009. The inclusion criteria were the following: (1) all patients had a Le Fort I osteotomy alone or combined with a bilateral sagittal split osteotomy; (2) all had completed their growth; (3) all underwent presurgical orthodontic treatment and orthognathic surgery by the same surgeon and with a similar technique with a computer-generated surgical splint; and (4) presurgical records were taken no more than 2 weeks before surgery, and postoperative records were taken 4 months after surgery. Preoperative and postoperative records were taken with the same CBCT machine. To minimize the effect of head position on the facial soft tissues, the following standard protocol was used during preoperative and postoperative scanning and photography procedures:

the preoperative and postoperative scans were taken with the patients in natural head position while gently biting on a thin wafer in centric relation. Natural head position was obtained by asking the patient to look at his or her image in a mirror. Then the distance between the suprasternal notch and soft-tissue pogonion was measured. We called this distance the natural head distance and made sure that it was respected during the preoperative and postoperative scanning procedures, and the photographs were taken at the same time in this dened position. The exclusion criteria were the following: (1) patients with congenital deformities or syndromes, such as cleft lip and palate; and (2) other adjunctive surgical procedures, such as rhinoplasty and liposuction. There were 6 key stages for the predictions and data gathering (all performed by 1 operator [N.A.]): (1) magnication calibration, (2) landmark identication, (3) creation of predictions by Dolphin and Maxilim softwares; (4) superimposition of the 2 predictions on a postoperative prole photograph, (5) measuring the differences in soft-tissue outlines of the superimposed images, and (6) error estimation. For the magnication calibration in the Dolphin software, 2 points 20 mm apart were marked on the lateral cephalograms using Photoshop software (CS3; Adobe, San Jose, Calif), and magnication calibration was done in the Dolphin software. Landmarks were identied using the Dolphin software; 15 soft-tissue and 25 hard-tissue landmarks were digitized. The soft tissue landmarks were glabella, nasal dorsum, pronasale, subnasale, soft-tissue A-point, labrale superior, stomion superior, stomion inferior, soft-tissue B-point, pogonion, gnathian, menton, soft tissue nasion, labrale inferior, and throat point. Predictions by the Dolphin software and the Maxilim software were created. Since validation of the accuracy of soft-tissue predictions was the aim of this study, the amounts of bony movement in 2 planes (anteroposterior and vertical) had to be the same as performed by the surgeon. Before creation of the predictions by the 2 softwares, the preoperative virtually osteotomized bone fragments of each patient were moved to the actual postoperative positions by the iterative closest point algorithm to achieve the best alignment and to align the surfaces of the maxilla and the mandible in 3 dimensions. The amount of bone displacement of each virtually osteotomized bony segment in the anteroposterior and vertical planes was considered to be the actual surgical movement and used by the softwares for prediction of the soft-tissue results. For prediction by the Dolphin software, preoperative photographs were registered on CBCT synthetized lateral

American Journal of Orthodontics and Dentofacial Orthopedics

November 2013  Vol 144  Issue 5

656

Nadjmi et al

Table I. Detailed amounts of bony movement and surgical procedures

Maxillary impaction Maxillary (mm) Bilateral sagittal advancement Anterior/ split osteotomy Genioplasty Patient (mm) posterior (mm) (mm) 1 4 3 3 2 3.5 3.5 4 2 3 5 5 3.5 2 4 4 4 6 5 3 2.5 2.5 4 6 6 7 7 5 5 4.5 8 3 4 4 9 4 4.5 5 10 4 3 1.5 4.5 4 11 3.5 2.5 2 2.5 12 5 2 5 13 0 5 5 3 5

cephalometry radiographs. This process begins with the registration of subnasale of 2 images on each other, and then the magnication and orientation of the 2 images are matched manually. Using the magnied image of superimposition, it is possible to complete the process of superimposition by nite adjustment of the softtissue outlines of lateral cephalometry on the photograph boundaries to obtain the best t of the images on each other. In the Maxilim software, the 3D photographs are aligned to the CBCT skin surface automatically, and a head model is created. Although the Maxilim software can construct 3D images, just the 2D simulated proles were used. Table I shows details of the bony movements and surgical procedures in each patient. The 2 predictions were superimposed on the postoperative photograph. For superimposition and quantitative validation in our study, Adobe Photoshop was used, and the postoperative and the 2 predicted images were superimposed based on the best-t method, so that the predicted and postoperative soft-tissue images were rigidly registered on each other based on unaltered surfaces. In our study, these included the surfaces above the infraorbital rim and the radix of the nose. In addition to the best t of unaltered surfaces, well-dened facial landmarks were digitized, such as soft-tissue nasion, sella, nasion, and right-eye outer canthus. Then superimposition of the sella-nasion line and outer canthussoft-tissue nasion line were used additionally for more accuracy (Fig 1). The same procedure that was used for the superimposition of the preoperative photographs on the preoperative CBCT image was applied to the superimposition

Fig 1. Superimposition of the postoperative photograph on the predicted photographs.

of the postoperative photographs on the postoperative CBCT synthetized lateral cephalometric radiographs in the Dolphin software. These images were used as the Dolphin prediction photographs in Adobe Photoshop for quantitative validation. We measured the differences in the soft-tissue outlines of the superimposed images. Linear measurements between each predicted image and the postoperative prole photographs (Fig 2) were measured.The differences between the computer predictions and the postoperation results were calculated for each landmark across the x-axis and y-axis. A negative value indicated that the computer prediction was more posterior in the x-axis or more superior in the y-axis compared with the postsurgical result. The nasolabial and mentolabial angles were also measured. Error of the method was assessed by analyzing the reliability and reproducibility of the method with the double determinant method by repeated digitizations of the landmarks.
Statistical analysis

All statistical analyses were carried out using SPSS software (version 11; SPSS, Chicago, Ill). Descriptive statistics, paired-sample t tests, and interclass correlations were used.

November 2013  Vol 144  Issue 5

American Journal of Orthodontics and Dentofacial Orthopedics

Nadjmi et al

657

Fig 2. Measurement of the differences of the superimposed photographs. Table II. Means, standard deviations, frequencies of clinically acceptable error, and 95% CI of the horizontal position of the predicted landmarks by the Dolphin software
Landmark Bridge of nose Pronasale Subnasale A-point Labrale superior Stomion superior Stomion inferior Labrale inferior B-point Pogonion Gnathion Menton Mean (mm) 0.0008 0.0331 0.25 0.12 0.12 0.10 0.05 0.06 0.02 0.04 0.05 0.00 Minimum (mm) 1.14 0.31 0.65 0.30 0.45 0.30 0.60 0.40 0.75 0.95 0.70 1.00 Maximum (mm) 0.20 0.33 0.30 0.50 0.65 1.02 0.80 0.80 0.70 0.90 1.10 1.20 SD 0.823 0.505 0.013 0.003 0.124 0.004 0.305 0.091 0.906 0.972 0.834 0.696 Frequency of clinically acceptable error (%) 6 0.5 92.3 92.3 84.6 84.6 92.3 92.3 84.6 84.6 60.6 76.9 56.6 69.2 95% CI (maximum error) 0.33, 0.13 (0.46) 0.14, 0.07 (0.21) 0.30, 0.09 (0.39) 0.20, 0.17 (0.40) 0.39, 0.09 (0.48) 0.36, 0.13 (0.20) 0.23, 0.12 (0.36) 0.32, 0.03 (0.47) 0.26, 0.21 (0.52) 0.30, 0.22 (0.64) 0.25, 0.37 (0.62) 0.32, 0.32 (0.64)

RESULTS

The error range of the Dolphin software in predicting the horizontal position of the soft-tissue landmarks was between 1.41 and 1.20 mm. The greatest errors were related to pronasale, gnathion, menton, and stomion superior. For the rest of the landmarks, there was a high coefcient of reliability between the predicted and the actual postoperative photographs (Table II). Variability in the vertical position of the predicted landmarks was between 1.85 and 1.55 mm. The maximum vertical errors were seen in subnasale, gnathion,

pogonion, and menton. The reliability coefcients were above 0.97 for each predicted landmark (Table III). The correlations of the predicted nasolabial and mentolabial angular measurement with the actual results were 0.72 and 0.81. The range of error for the Maxilim software in the horizontal positions of the predicted landmarks was between 1.60 and 1.50 mm. The greatest errors in horizontal positions were observed in menton, gnathion, and pogonion (Table IV). In the vertical direction, the variability ranged between 4.25 and 2.42 mm. The

American Journal of Orthodontics and Dentofacial Orthopedics

November 2013  Vol 144  Issue 5

658

Nadjmi et al

Table III. Means, standard deviations, frequencies of clinically acceptable error, and 95% CI of the vertical position of the predicted landmarks by the Dolphin software
Landmark Bridge of nose Pronasale Subnasale A-point Labrale superior Stomion superior Stomion inferior Labrale inferior B-point Pogonion Gnathion Menton Mean (mm) 0.008 0.196 0.079 0.126 0.125 0.102 0.034 0.066 0.042 0.167 0.069 0.000 Minimum (mm) 0.20 0.30 1.85 0.30 0.45 0.32 0.50 0.40 0.40 0.65 0.70 1.00 Maximum (mm) 0.12 0.60 0.50 0.50 0.65 1.02 1.05 0.80 1.15 1.25 1.55 1.20 SD 0.087 0.156 0.603 0.246 0.351 0.388 0.378 0.332 0.40 0.502 0.596 0.594 Frequency of clinically acceptable error (%) 6 0.5 100 92.3 92.3 92.3 92.3 92.3 84.6 92.3 92.3 69.2 76.9 69.2 95% CI (maximum error) 0.02, 0.04 (0.10) 0.05, 0.33 (0.38) 0.46, 1.12 (0.64) 0.01, 0.26 (0.26) 0.06, 0.31 (0.36) 0, 0.40 (0.40) 0.23, 0.16 (0.40) 0.24, 0.11 (0.35) 0.18, 0.26 (0.44) 0.10, 0.42 (0.52) 0.26, 0.30 (0.64) 0.32, 0.32 (0.64)

Table IV. Means, standard deviations, frequencies of clinically acceptable error, and 95% CI of the horizontal position

of the predicted landmarks by the Maxilim software

Landmark Bridge of nose Pronasale Subnasale A-point Labrale superior Stomion superior Stomion inferior Labrale inferior B-point Pogonion Gnathion Menton Mean (mm) 0.007 0.040 0.049 0.129 0.039 0.062 0.052 1.00 0.010 0.160 0.031 0.192 Minimum (mm) 0.50 0.28 0.90 0.85 0.88 0.85 0.60 0.80 0.85 1.40 1.40 1.60 Maximum (mm) 0.70 0.45 0.70 0.30 1.05 0.55 0.40 0.30 0.75 0.45 0.75 0.80 SD 0.317 0.207 0.391 0.314 0.439 0.326 0.292 0.351 0.398 0.512 0.616 0.647 Frequency of clinically acceptable error (%) 6 0.5 92.3 100 84.6 92.3 84.6 92.3 92.3 76.9 84.6 76.9 53.3 46.1 95% CI (maximum error) 0.27, 0.67 (0.40) 0.07, 0.15 (0.22) 0.26, 0.16 (0.40) 0.30, 0.04 (0.34) 0.27, 0.20 (0.47) 0.24, 0.11 (0.35) 0.21, 0.10 (0.31) 0.28, 0.08 (0.36) 0.23, 0.21 (0.44) 0.44, 0.12 (0.56) 0.31, 0.37 (0.68) 0.59, 0.16 (0.75)

Table V. Means, standard deviations, frequencies of clinically acceptable errors, and 95% CI of the vertical position of the predicted landmarks by the Maxilim software
Landmark Bridge of nose Pronasale Subnasale A-point Labrale superior Stomion superior Stomion inferior Labrale inferior B-point Pogonion Gnathion Menton Mean (mm) 0.063 0.123 0.020 0.558 0.330 0.893 0.080 0.046 0.089 0.080 0.085 0.192 Minimum (mm) 0.25 0.50 0.35 4.25 2.60 4.10 0.80 0.65 1.00 1.15 1.36 1.60 Maximum (mm) 0.10 0.55 0.40 2.42 0.30 0.52 0.50 0.90 0.80 0.90 1.05 0.80 SD 0.100 0.319 0.226 1.941 0.753 1.498 0.345 0.437 0.501 0.498 0.590 0.674 Frequency of clinically acceptable error (%) 6 0.5 100% 100% 100% 46.2% 76.9% 76.9% 92.3% 84.6% 76.9% 84.6% 76.9% 46.1% 95% CI (maximum error) 0.20, 0.57 (0.37) 0.05, 0.29 (0.34) 0.09, 0.15 (0.24) 1.6, 0.62 (2.22) 0.73, 0.07 (0.08) 2.03, 0.73 (2.76) 0.26, 0.12 (0.38) 0.20, 0.28 (0.48) 0.15, 0.34 (0.59) 0.40, 0.13 (0.53) 0.67, 0.00 (0.67) 0.41, 0.23 (0.64)

positions of the predicted landmarks in the Maxilim software were highly correlated to the actual results in both the horizontal and vertical dimensions. The weakest correlation was observed between the vertical positions of stomion superior, at 0.83 (Table V).

For the nasolabial and mentolabial angles, the correlations between the predicted and actual results were 0.95 and 0.92, respectively. There were no statistical differences between the absolute prediction errors of the Dolphin and Maxilim

November 2013  Vol 144  Issue 5

American Journal of Orthodontics and Dentofacial Orthopedics

Nadjmi et al

659

softwares and the actual results in relation to the horizontal and vertical positions of the landmarks (P .0.05), except for the vertical position of softtissue A-point in the Maxilim simulated prole. The 95% condence interval (CI) showed that for the nose and lower lip the errors of the 2 prediction programs had limited clinical signicance (less than 1 mm). The vertical positions of soft-tissue A-point and stomion superior had the greatest errors associated with the Maxilim prediction program. Gnathion and menton showed the greatest errors in the predictions, but there were no signicant differences regarding the mean errors of each prediction program. For the nasolabial and mentolabial angles, the differences of absolute errors in Dolphin and Maxilim were not signicant (P .0.05). For the 95% CI, the error in the nasolabial angle was larger with the Dolphin software (9 ) relative to the Maxilim software (3 ). For the mentolabial angle, the errors were 9 in Dolphin and 6.8 in Maxilim.
DISCUSSION

The aim of this study was to compare the postsurgical soft-tissue proles of the patients who had Le Fort I osteotomy alone or combined with bilateral sagittal split osteotomy with the proles predicted by the Dophin and Maxilim softwares. Our study showed that the postoperative positions of the soft-tissue points were highly correlated with the predicted ones in both softwares, and that there were no signicant differences in errors between the predictions and the real surgical changes. The vertical and horizontal positions of the nose, upper lip, and lower lip landmarks could be produced with Dolphin and Maxilim at an accuracy of 70%; in the chin region, the accuracy was 50%. The largest systematic difference between the predicted and real changes was observed in the chin region with both softwares. These ndings were consistent with the results of Loh et al,10 and it was assumed that this occurred because of the different amounts of mandibular autorotation between the predicted and actual results after maxillary impaction, caused by inaccuracies in locating the center of the mandibular rotational axis. The complexity of predicting in the chin region has been reported by Xia et al6 because of multiple directional movements of the soft tissues in the chin region. It seems that the chin is the most difcult region to predict; however, as Magro-Filho et al22 mentioned, it seems that Dolphin is more accurate in predictions of the chin area and the base of the mandible than is Dentofacial Planner. Some previous studies reported the greatest inaccuracies in the predictions of the lips,2,20,24-27

although in our study the position of the lips was predicted accurately, and the largest errors in the 95% CI in both softwares were less than 0.5 mm, with the exception of 2 points (vertical positions of labrale superior and stomion superior in Maxilim). In our subjects, the Dolphin software did not predict the changes in the shape of the lips after surgery. For example, increased show of vermilion and some other aspects such as lip lengthening and competency were not predicted and should be adjusted by the operator. In 1 patient who had an increased show of vermilion after maxillary advancement, the Dolphin prediction of the lip shape was distorted because it could not predict the changes in volume and thickness of the lips. Konstiantos et al2 also found that although there were high correlations between the predicted lip landmarks and the postsurgical results, the shape of the lips was not predicted accurately. This limitation was observed not only in Class II subjects, but also in Class III subjects because the upper lip was trapped by the mandibular incisors. Differences in lip tonicity, length, posture, and mass between patients result in more inaccuracies in software that use preprogrammed xed hard-tissue to soft-tissue ratios for prediction of the soft-tissue results.28-30 These limitations did not exist with the predictions of Maxilim, since it is based on tensor model analysis and can predict rolling the lips as free points on the teeth during simulation; it also predicts the changes in lip thickness and show of vermilion more accurately. Although the mean errors of the vertical predictions of A-point and stomion superior in Maxilim were relatively small (0.55 and 0.89 mm, respectively), the ranges were huge: from 4.25 to 2.42 mm, and 4.10 to 0.52 mm, respectively (Table V). This discrepancy also is expressed in the 95% CI values of these points: 2.22 mm for A-point and 2.67 mm for stomion superior; these are much larger than the values for the other points. The 95% CI was less than 0.5 mm (clinically acceptable error) for all other points in the horizontal and vertical positions. It seems that for these 2 points the range and the 95% CI reect the discrepancy better than the mean error. Both softwares were accurate in the prediction of nose landmarks. The systematic errors were insignicant, and inaccuracies in the 95% CI were less than 0.45 mm, with the exception of the vertical position of subnasale in Dolphin (0.64 mm), and clinically acceptable; the prediction accuracy was higher than in other softwares in similar studies.2,10,20 The analysis of angular measurements showed that the amount of systematic errors and the prevalence of signicant clinical errors were increased in both softwares, especially in Dolphin. It seems that this occurred

American Journal of Orthodontics and Dentofacial Orthopedics

November 2013  Vol 144  Issue 5

660

Nadjmi et al

because the errors in different landmarks of an angle were not in the same direction. If landmark prediction in 1 angle arm was overpredicted and in the other arm was underpredicted, then the sum of errors in 2 direction leads to increased inaccuracy of angle prediction. Kaipatur and Flores-Mir25 in a systematic review of prediction computer programs stated that although differences larger than 2 mm are important, when small inaccuracies are added in each plane of space, the combined error might have clinical implications. The errors in the Dolphin predictions were 3- and 1.4-fold higher than in Maxilim for the nasolabial and mentolabial angles, respectively, and 9 for each angle in the 95% CI. De Lira Ade et al23 reported errors of 3.7 and 4.8 , respectively, in 95% CI values for the nasolabial and mentolabial angles with Dolphin predictions. In our study, these angles were predicted less accurately, but there were still acceptable correlations with the actual results, and the differences of systematic errors were insignicant between the 2 softwares. Eales et al20 used the Student t test to compare the prediction of angles by COG software (version 3.4) and found signicant differences for nasolabial predictions. Using the mentioned standard error in that study, we calculated the 95% CI for nasolabial angle error; it showed that the error was 11 , comparable with the Dolphin prediction error but 4 times more than the Maxilim error. In a comparison of Dolphin with Dentofacial Planner, Magro-Filho et al22 noted that Dolphin was better for predicting the nasolabial angle. There were some differences between our results compared with other studies that used other 2D softwares. Eales et al,20 Loh et al,10 Konstiantos et al,2 and Donatsky et al26,27 have shown signicant differences in lower lip prediction errors rather than the chin area. One explanation for these differences might be that different statistical analyses were used in our study in comparison with the others. Loh et al,10 Konstiantos et al,2 and Donatsky et al26,27 compared the mean differences and used Wilcoxon and t tests. Jones et al24 compared the medians to overcome the problems of overestimation and underestimations compensations that might occur in mean comparisons. In our study, the differences of landmark positions between the 2 softwares and the actual positions were compared by the intraclass correclation coefcient. The t test was used for comparison of systematic errors of the 2 softwares. As Power and Hillard17 explained, signicant statistical differences in the mean comparisons do not imply clinical signicance because in the assessment of any software some random error is acceptable. In addition, other factors were considered in our study. There was a wide range of disparity between studies in the timing of postsurgical assessments; some were

taken immediately after surgery,10,26,27 and in 1 study during the rst year after surgery.2 In our study, all postsurgical records were taken 4 months after surgery so that soft-tissue edema was minimized. An important point is that soft-tissue changes occurred even after the rst year after surgery as a result of orthodontic tooth movement, relapse, or bone remodeling, so taking the records of all patients at the same time and before orthodontic tooth movement is essential.31 Other variables are impossible to control and make the surgical soft-tissue predictions difcult; eg, weight change from presurgery to postsurgery, parasthesia of the upper lip, and muscle pose can affect the results. A number of studies that had evaluated other softwares did not consider factors such as uniformity of the samples, accuracy of the surgery, an accurate method to measure the surgical movements, or the effects of preoperative or postoperative orthodontic treatment and relapse.1,2,20 Like Jones et al,24 we mentioned that to measure the changes and assess the accuracy of surgical procedures, presurgical and postsurgical lateral cephalograms were superimposed on the sella-nasion line rather than on more variable landmarks such as overjet, overbite, or molar cusps, which were used in other studies.1,2,20 Superimpositions were also done along maxillary and mandibuar bases to consider the probability of dental movements after surgery. Furthermore, in our study, CBCT synthesized lateral cephalograms were used; they had the advantage of better soft-tissue visualization and landmark identication over conventional lateral cephalograms. Since the results showed no major systematic errors in soft-tissue predictions by these softwares, except in the chin region and as mentioned in a recent systematic review of computer program accuracy, the studied softwares gave accurate predictions with errors of less than 2 mm, except for some points in the lower third of the face.25 The most important differences between the 2 software packages were in their soft-tissue algorithms; predictions of lips and proles; qualitative image characteristics; ability to rene images; cephalograms-photographs linking; magnication calibration; image enhancement options; and compatibility with different operating systems.18,21,28,32 Although both softwares were accurate in predictions of proles, Maxilim has the third dimension, which makes the quantication of volumetric changes and predictions in the transverse plane possible. Maxilims dimensional ability is helpful, especially in complicated orthognathic surgery cases such as cleft lip and palate, and asymmetric face. However, prediction planning of the third dimention was beyond the scope of this study.

November 2013  Vol 144  Issue 5

American Journal of Orthodontics and Dentofacial Orthopedics

Nadjmi et al

661

A possible limitation of this study was the few subjects (6) with Le Fort advancement surgery. In addition, the degree of anteroposterior advancement was relatively small, and future studies will be needed to validate the predictive capacity of these softwares on larger anteroposterior surgical movements.
CONCLUSIONS

This study shows that both Dolphin and Maxilim softwares are accurate in predictions of the soft-tissue prole and can be used for planning and prediction in orthognathic surgery. There are inaccuracies in predictions of the chin region that should be considered in complicated surgical planning.
ACKNOWLEDGMENTS

This paper has been entirely drawn up from a postgraduate Thesis, which was completed by Dr Niloufar Azami under the supervision of Dr Nasser Nadjmi and Dr Azita Tehranchi with the cooperation of the Orthodontic Department, Dental School, Shahid Beheshti Medical Sciences University, Tehran, Iran.
REFERENCES 1. Cousley RJ, Grant E. The accuracy of preoperative orthognathic predictions. Br J Oral Maxillofac Surg 2004;42:96-104. 2. Konstiantos AK, O'Reilly MT, Close J. The validity of the prediction of soft tissue prole changes after LeFort I osteotomy using the dentofacial planner (computer software). Am J Orthod Dentofacial Orthop 1994;105:241-9. 3. Farkas LG. Anthropometry of the head and face. 2nd ed. New York: Lippincott Williams and Wilkins; 1994. 4. Mollemans W, Schutyser F, Nadjmi N, Maes F, Suetens P. Predicting soft tissue deformations for a maxillofacial surgery planning system: from computational strategies to a complete clinical validation. Med Image Anal 2007;11:282-301. 5. Sarver DM. Esthetic orthodontic and orthognathic surgery. St Louis: Mosby; 1998. 6. Xia J, Samman N, Yeung RW, Shen SG, Wang D, Ip HH, et al. Three dimensional virtual reality surgical planning and simulation workbench for orthognathic surgery. Int J Adult Orthod \Orthognath Surg 2000;15:265-82. 7. Burk DL, Mears DC, Cooperstein LA, Herman GT, Udupa JK. Threedimensional computed tomographic imaging and interactive surgical planning. J Comput Tomogr 1987;10:1-10. 8. Chabanas M, Luboz V, Payan Y. Patient specic nite element model of the face soft tissues for computer-assisted maxillofacial surgery. Med Image Anal 2002;7:131-51. 9. Jacobson R, Sarver DM. The predictability of maxillary repositioning in Le Fort I orthognathic surgery. Am J Orthod Dentofacial Orthop 2002;122:142-54. 10. Loh S, Heng JK, Ward-Booth P, Winchester L, McDonald F. A radiographic analysis of computer prediction in conjunction with orthognathic surgery. Int J Oral Maxillofac Surg 2001;30:259-63. 11. Carter AC, Larson BE, Guenthner TA. Accuracy of video imaging in mandibular surgery. Int J Adult Orthod Orthognath Surg 1996;11: 289-300.

12. Chunmaneechote P, Friede H. Mandibular setback osteotomy: facial soft tissue behavior and possibility to improve the accuracy of the soft tissue prole prediction with the use of a computerized cephalometric program: Quick Ceph Image Pro: v. 2.5. Clin Orthod Res 1999;2:85-98. 13. Csaszar GR, Bruker-Csaszar B, Niederdellmann H. Prediction of soft tissue proles in orthodontic surgery with the Dentofacial Planner. Int J Adult Orthod Orthognath Surg 1999;14: 285-90. 14. Giangreco TA, Forbes DP, Jacobson RS, Kallal RH, Moretti RJ, Marshall SD. Subjective evaluation of prole prediction using video imaging. Int J Adult Orthod Orthognath Surg 1995;10:211-7. 15. Kolokitha OE, Athanasiou AE, Tuncay OC. Validity of computerized predictions of dentoskeletal and soft tissue prole changes after mandibular setback and maxillary impaction osteotomies. Int J Adult Orthod Orthognath Surg 1996;11:137-54. 16. Legan HL, Burstone CJ. Soft tissue cephalometric analysis for orthognathic surgery. J Oral Surg 1980;38:744-51. 17. Power G, Hillard T. Dolphin imaging software: an analysis of the accuracy of cephalometric digitization and orthognathic prediction. Int J Oral Maxillofac Surg 2005;34:619-26. 18. Gosset CB, Preston B. Prediction accuracy of computer assisted surgical visual treatment objectives as compared with conventional visual treatment objectives. J Oral Maxillofac Surg 2005; 63:609-17. 19. Pektas ZO, Kircelli BH, Cilasun U, Uckan S. The accuracy of computer-assisted surgical planning in soft tissue prediction following orthognathic surgery. Int J Med Robot 2007;3:64-71. 20. Eales EA, Newton C, Jones ML, Sugar A. The accuracy of computerized prediction of the soft tissue prole: a study of 25 patients treated by means of the Le Fort I osteotomy. Int J Adult Orthod Orthognath Surg 1994;9:141-52. 21. Smith JD, Thomas PM, Proft WR. A comparison of current prediction imaging programs. Am J Orthod Dentofacial Orthop 2004; 125:527-36. 22. Magro-Filho O, Magro-Ernica N, Queiroz TP, Aranega AM, Garcia IR Jr. Comparative study of 2 software programs for predicting prole changes in Class III patients having double-jaw orthognathic surgery. Am J Orthod Dentofacial Orthop 2010;137: 452.e1-5. 23. de Lira Ade L, de Moura WL, de Barros Vieira JM, Nojima MG, Nojima L. Surgical prediction of skeletal and soft tissue changes in Class III treatment. J Oral Maxillofac Surg 2012; 70:e290-7. 24. Jones RM, Khambay BS, McHugh S, Ayoub AF. The validity of a computer assisted simulation system for orthognathic surgery (CASSOS) for planning the surgical correction of Class III skeletal deformities: single-jaw versus bimaxillary surgery. Int J Oral Maxillofac Surg 2007;36:900-8. 25. Kaipatur NR, Flores-Mir C. Accuracy of computer programs in predicting orthognathic surgery soft tissue response. J Oral Maxillofac Surg 2009;67:751-9. 26. Donatsky O, Bjorn-Jorgensen J, Hermund NU, Nielsen H, Holmqvist-Larsen M, Nerder PH. Accuracy of combined maxillary and mandibular repositioning and of soft tissue prediction in relation to maxillary antero-superior repositioning combined with mandibular set back. A computerized cephalometric evaluation of the immediate postsurgical outcome using the TIOPS planning system. J Craniomaxillofac Surg 2009;37:279-84. 27. Donatsky O, Bjorn-Jorgensen J, Hermund NU, Nielsen H, Holmqvist-Larsen M, Nerder PH. Immediate postoperative outcome of orthognathic surgical planning, and prediction of positional changes

American Journal of Orthodontics and Dentofacial Orthopedics

November 2013  Vol 144  Issue 5

662

Nadjmi et al

in hard and soft tissue, independently of the extent and direction of the surgical corrections required. Br J Oral Maxillofac Surg 2011;49:386-91. 28. Syliangco ST, Sameshima GT, Kaminishi RM, Sinclair PM. Predicting soft tissue changes in mandibular advancement surgery: a comparison of two video imaging systems. Angle Orthod 1997; 67:337-46. 29. Sameshima GT, Kawakami RK, Kaminishi RM, Sinclair PM. Predicting soft tissue changes in maxillary impaction surgery: a comparison of two video imaging systems. Angle Orthod 1997;67:347-54.

30. Kazandjian S, Sameshima GT, Champlin T, Sinclair PM. Accuracy of video imaging for predicting the soft tissue prole after mandibular set-back surgery. Am J Orthod Dentofacial Orthop 1999;115:382-9. 31. Hack B, Otolo D. Long term stability of soft tissue changes after Le Fort I osteotomy. Am J Orthod Dentofacial Orthop 1993;104: 544-55. 32. Lu CH, Ko EW, Huang CS. The accuracy of video imaging prediction in soft tissue outcome after bimaxillary orthognathic surgery. J Oral Maxillofac Surg 2003;61:333-42.

November 2013  Vol 144  Issue 5

American Journal of Orthodontics and Dentofacial Orthopedics