- Email: [email protected]
The accuracy of video imaging prediction in soft tissue outcome after bimaxillary orthognathic surgery
J Oral Maxillofac Surg 61:333-342, 2003
The Accuracy of Video Imaging Prediction in Soft Tissue Outcome After Bimaxillary Orthognathic Surgery Chien-Hsun Lu, DDS,* Ellen Wen-Ching Ko, DDS, MS,† and Chiung-Shing Huang, DDS, PhD‡ Purpose:
The purpose of the present study was to evaluate the accuracy of the outcome in soft tissue prediction through use of a computer imaging system after bimaxillary orthognathic surgery. Materials and Methods: The study sample consisted of 30 adults who had undergone orthognathic surgery that included the Wassmund and Ko ¨ le procedures and optional genioplasty to correct bimaxillary protrusion. All the patients had lateral cephalometric radiographs and profile photographs taken within 6 months before surgery and at least 6 months after surgery. The computer-generated soft tissue image and the actual postsurgical profile were compared. The accuracy of this computer-generated profile image was evaluated. Results: The results indicated that the nasal tip, soft tissue A point, and upper lip presented the least predicted errors in sagittal plane. While the nasal tip presented higher reliability. Lower lip prediction was found to be the least accurate region and it tended to be located anterior to the actual position. In the vertical plane, most of the predictions revealed higher accuracy than those in the sagittal plane. There were no statistically significant differences between the predictions of the groups with and those without genioplasty. Conclusions: Computer-generated image prediction was suitable for patient education and communication. However, efforts are still needed to improve the accuracy and reliability of the prediction program and to include the consideration of changes in soft tissue tension and muscle strain. The accuracy of this system in soft tissue prediction should be carefully interpreted. © 2003 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 61:333-342, 2003 Improvement of facial aesthetics is a main reason patients request surgical correction of dentofacial deformities.1 The definition of an ideal result of facial improvement after surgical orthodontic treatment is very subjective. The evaluation is affected by different opinions and treatment philosophies. Therefore, a successful orthognathic surgery includes not only the precise surgical technique and occlusal correction
but also the accomplishment of the aesthetic goals that are satisfying to both patients and professionals.2-6 It is hard for the public to imagine the change in facial appearance after orthognathic surgery without a visual reference.6,7 Lateral cephalometric radiographs are commonly used to predict the surgical treatment outcomes. The cephalometric prediction can be performed manually by repositioning an overlaid tracing or by moving templates.8-12 It involves cutting different parts of the acetate tracing and repositioning them over the original cephalometric tracing to simulate the surgical treatment. The predicted posttreatment soft tissue outline is delineated based on the published ratios of soft tissue to hard tissue changes. With the development of computer science, cephalometric landmarks were able to be digitized into the computer and the repositioning of the surgical parts can now be viewed immediately on the screen.13-16 The predicted outline of the posttreatment facial profile is generated by specialized computer software. The 2 aforementioned methods pre-
Received from Chang Gung Memorial Hospital, Taipei, Taiwan. *Attending Staff, Department of Orthodontics and Craniofacial Dentistry. †Attending Staff, Department of Orthodontics and Craniofacial Dentistry. ‡Chairman and Professor, Faculty of Dentistry. Address correspondence and reprint requests to Dr Ko: Department of Orthodontics and Craniofacial Dentistry, Chang Gung Memorial Hospital, 199 Tung-Hwa North Rd, Taipei, 105, Taiwan; e-mail: [email protected] © 2003 American Association of Oral and Maxillofacial Surgeons
0278-2391/03/6103-0009$30.00/0 doi:10.1053/joms.2003.50058
333
334
THE ACCURACY OF VIDEO IMAGING PREDICTION
Table 1. LITERATURE REVIEW OF THE LEAST ACCURATE REGION IN SOFTWARE PREDICTION OF FACIAL SOFT TISSUE CHANGES AFTER ORTHOGNATHIC SURGERY
Author
Least Accurate Region
Predicted Error (mm)
Sample Size
Kazandjian et al27
Mandibular setback
QC, PoP
32
Maxillary impaction
OTP, PrP
Syliangco et al25
Lower lip
39
Mandibular advancement
OTP, PrP
Konstiantos et al23
Lower lip Pronasal Lower lip Chin
⫹2.25 ⫾ 3.63 (H) ⫺2.23 ⫾ 2.85 (V) ⫹1.71 ⫾ 1.37 (H) ⫹2.88 ⫾ 2.62 (V) ⫹1.61 ⫾ 1.24 (H) ⫹1.77 ⫾ 1.26 (V) ⫹1.57 ⫾ 2.0 (H) ⫺1.64 ⫾ 1.4 (V) ⫹1.9 ⫾ 0.38 (H) ⫺0.5 ⫾ 0.78 (V)
30
Sameshima et al26
Upper lip Lower lip Lower lip
21
Le Fort I osteotomy
DP
16
Mandibular advancement
QC
Hing22
Type of Surgery
Software
Abbreviations: H, horizontal plane; V, vertical plane; ⫹, predicted landmarks anterior (horizontal) or inferior (vertical) to the actual one; ⫺, predicted landmarks posterior (horizontal) or superior (vertical) to the actual one; QC, Quick Ceph; PoP, Portrait Planner; OTP, Orthognathic Treatment Planner; PrP, Prescription Portrait; DP, Dentofacial Planner (DentoFacial Software, Inc, Toronto, Canada).
sented only the “line drawing” profile of the surgical simulation. The linear presentation of the facial profile (although it is relatively informative to the professionals) cannot provide a realistic image of the treatment results to the patients and the public.12 Manipulation of patients’ photographs to illustrate treatment goals provides the patients with a more vivid image.17 However, it is believed to be more of an art form than a scientific exercise. Recent prediction techniques integrate video images with cephalograms.3,4,18-20 The patient’s lateral photograph is superimposed on the digitized cephalometric tracing. The computer-generated prediction includes both line drawing tracing and the corresponding facial image. With the facial image, the patient can have an idea of the possible changes of their facial appearance. The predicted image also facilitates the communication between orthodontists, surgeons, and patients.3,6,7 Several studies point to the influence of video image on treatment plans. Sarver et al3 evaluated the treatment result 4 months after surgery; it revealed that 89% of the patients thought that the image predictions were realistic and that the desired results were achieved. It also assisted 83% of the patients to make a decision regarding whether to undergo orthognathic surgery. None of the patients believed that the imaging session had no value. When using video imaging as a presentation tool, Phillips et al21 found that in the standard case presentation group (the group without video image presentation), 47% of the patients ranked dental casts as the most helpful physical record in making the treatment decision and 46% of the patients considered that the acetate profile tracing was the most helpful tool for understanding the treatment expectation. However, in the video image presentation group (the group including video image presentation), the results indicated that 42% of
the patients thought the video imaging presentation was the most helpful tool in decision making and that 39% of the subjects considered it to be the most helpful tool for understanding the treatment outcomes.21 The influence of video imaging on patients’ perception was significant. Showing the video imaging simulation to patients should be done conservatively and carefully to avoid an unrealistic expectation of the final results.12 Phillips et al21 compared the final treatment outcomes in 2 groups: the video image case presentation group and the standard case presentation group; the previous one presented a significantly higher self-image expectation. The authors thought that the visual display of treatment simulation was more realistic than the “line drawing” tracing graph and much more accepted by the public for the actual surgical treatment results. Also, it provided the patients with a more appropriate appreciation for the changes in appearance after the orthognathic surgery. When comparing the quality of various predicted images, Sinclair et al12 found that none of the predicted images were more favorable than the actual profile changes after orthognathic surgery. The accuracy of the computer-predicted images after orthognathic surgery is examined in many studies.12,22-27 It seems that the lips was the least accurate region that the computer could predict regardless of the kind of surgical methods that were performed and the kind of prediction software that was used (Table 1). However, most of the studies evaluated only one jaw surgery. The precision of treatment simulation in simultaneous maxillary and mandibular setback procedure has not been investigated. The purpose of this study was to evaluate the accuracy of a video imaging program in predicting soft tissue outcomes after orthognathic surgery in patents with bimaxillary protrusion who underwent simultaneous maxillary and mandibular setback.
335
LU, KO, AND HUANG
FIGURE 1. A, Cephalometric landmarks used in this study: S, sella; N, nasion; Po, porion; G, glabella; Prn, tip of nose; Sn, subnasale; A⬘, soft tissue A point; UL, upper lip; LL, lower lip; B⬘, soft tissue B point; Pg⬘, soft tissue pogonion; Pg, pogonion; LIX, lower incisal apex; LI, lower incisal tip; UI, upper incisal tip; UIX, upper incisal apex. B, Linear and angular measurement: 1, facial convexity; 2, nasolabial angle; 3, upper lip to facial plane; 4, lower lip to facial plane; 5, upper lip to E plane; 6, lower lip to E plane (facial plane: glabella to soft tissue pogonion; E plane: tip of nose to soft tissue pogonion).
Materials and Methods The sample consisted of 30 patients who met the following criteria: 1) adult older than 19 years (average age, 27.9 years; age range, 19 to 37 years); 2) underwent Wassmund procedure to setback anterior maxilla and Ko ¨ le procedure to setback anterior mandible, with or without genioplasty; and 3) no congenital craniofacial deformities or trauma and head and neck surgical history. The diagnostic records included the lateral cephalometric radiographs and profile photographs within 6 months before surgery and at least 6 months after surgery. The head films and photographs were acquired in the natural head position with teeth in centric occlusion and relaxed lip posture. There were no intraoral fixed orthodontic appliances shown in both the head films and photographs. All of the patients were treated in the Craniofacial Center, Chang Gung Memorial Hospital, Taipei, Taiwan. Both before and after cephalograms were traced on the acetate papers. All the cephalometric tracings and profile photographs were entered in a Pentium-based computer (Intel Pentium III, 500 MHz, Windows 98SE, Microsoft Corporation, Redmond, WA) through a scanner (Astra 2400S; Umax Technologies Inc, Taipei, Taiwan) for analysis. The Dolphin Imaging system software (Version 6; Dolphin Imaging, Canoga Park, CA) was used to store and generate image prediction.
METHOD OF PREDICTION
The sagittal and vertical treatment changes were evaluated by linear measurements within an X-Y coordinate system (Fig 1A). The SN plane was defined as the horizontal reference plane (x-axis), and a line perpendicular to this plane through sella was defined as the vertical reference plane (y-axis). Landmarks of sella (S), nasion (N), and porion (Po) in the presurgical tracing were all transferred to the postsurgical tracing in the same patient. The tracings of presurgical and postsurgical cephalograms were superimposed at the cranial base to ensure the X-Y planes been accurately transferred. The tracings and photographs were then input into the computer system, digitized, and superimposed following instructions of the Dolphin Imaging software. The skeletal and dental landmarks included anterior nasal spine (ANS), upper incisal edge (UI), lower incisal edge (LI), upper incisor root apex (UIX), lower incisal root apex (LIX), and pogonion (Pg) (Fig 1A). The perpendicular distance of each landmark to both reference planes (x- and y-axes) was recorded before and after surgery. The treatment changes of hard tissue in each case were obtained from the differences between presurgical and postsurgical linear measurements. The hard tissue image was moved according to the prescribed distances (the treatment change) by using these calculations and the VTO (visual treatment objectives) function in the software. The predicted soft tissue outline and the corresponding coordinates of the soft tissue
336
THE ACCURACY OF VIDEO IMAGING PREDICTION
Table 2. SOFT TISSUE–TO–HARD TISSUE MOVEMENT RATIO SHOWN IN THE SYSTEM INITIATION FILE OF THE SOFTWARE
Soft Tissue–to–Hard Tissue Movement Ratio for Soft Tissue Landmark (X ratio, Y ratio)
Hard Tissue Movement Maxilla horizontal movement Maxilla rotational movement Anterior maxilla horizontal movement Anterior maxilla rotational movement Mandible horizontal movement Mandible rotational movement Mandible vertical movement Chin horizontal movement Chin rotational movement Chin vertical movement Symphysis horizontal movement Symphysis vertical movement Symphysis rotational movement Mandible rotational movement in orthodontic treatment
Subnasale (1:1,1:1), A⬘ (1:1,1:1), upper lip (1:1,1:1), stomion (1:1,1:1), stomion superior (1:1,1:1) Subnasale (1:1,1:1), A⬘ (1:1,1:1), upper lip (1:1,1:1), stomion (1:6,1:6), stomion superior (1:1,1:1) Subnasale (1:1,1:1), A⬘ (1:1,1:1), upper lip (1:1,1:1), stomion (1:1,1:1), stomion superior (1:1,1:1) Subnasale (1:1, 1:1), A⬘ (1:1,1:1), upper lip (1:1,1:1), stomion (1:1,1:1), stomion superior (1:1,1:1) Stomion inferior (1:1, 1:1), lower lip (1:1,1:1), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1), stomion inferior (1:1,1:1) Stomion inferior (1:1, 1:1), lower lip (1:1,1:1), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1) Lower lip (1:1,1:1), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1), stomion (1:1,1:1), stomion inferior (1:1,1:1) B⬘ (1:2,1:2), Pg⬘ (1:1,1:1) Pg⬘ (1:1,1:1) Pg⬘ (1:1,1:1) Lower lip (1:1,1:1), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1), stomion (1:1,1:1), stomion inferior (1:1,1:1) Lower lip (1:2,1:2), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1), stomion (1:1,1:1), stomion inferior (1:1,1:1) Lower lip (1:2,1:2), B⬘ (1:2,1:2), Pg⬘ (1:1,1:1), stomion inferior (1:1,1:1) Pg⬘ (1:1,1:1), B⬘ (1:1,1:1), upper lip (1:1,1:1)
Abbreviations: X ratio, soft tissue–to– hard tissue movement ratio in horizontal direction; Y ratio, soft tissue–to– hard tissue movement ratio in vertical direction; B⬘, soft tissue B point; Pg⬘, pogonion.
points were automatically generated. The ratios of soft tissue to hard tissue movement are varied according to the specific parts of the hard tissue. The prescription of these ratios is shown in the system initiation file (dolphin.ini) of the Dolphin imaging system (Table 2). The differences in soft tissue outline between predicted tracing and the actual profile were compared (predicted error) to test the accuracy of this system. Seven soft tissue landmarks, including tip of the nose (Prn), subnasale (Sn), soft tissue A point (A⬘), upper lip (UL), lower lip (LL), soft tissue B point (B⬘), and soft tissue pogonion (Pg⬘) were evaluated (Fig 1A). The perpendicular distance of each landmark to the X-Y planes was measured. Also, 4 linear and 2 angular measurements were observed, including upper lip protrusion (perpendicular distance from upper vermilion border to the line connecting soft tissue glabella and Pg⬘), lower lip protrusion (perpendicular distance from lower vermilion border to the line connecting soft tissue glabella and Pg⬘), upper lip to E plane (perpendicular distance form upper vermilion border to the plane connecting the Prn and Pg⬘), lower lip to E plane (perpendicular distance form lower vermilion border to the plane connecting the Prn and Pg⬘), facial convexity (angle formed by glabella-Sn, Sn-Pg⬘), and nasolabial angle (Fig 1B).
METHOD ERROR
Errors associated with digitization and methods of measurement were assessed by retracing and redigitizing on 10 randomly selected cephalometric films. The average differences in these 2 sets of linear measurement were within 0.6 mm. The average differences in these 2 sets of angular measurements were within 1.15o. There were no statistically significant differences between each measurement by comparison with paired t-test (P ⬎ .5).
Results SAGITTAL PLANE
When comparing the landmarks located in the computer-generated prediction with the actual profile change at sagittal plane (Table 3), the mean differences smaller than 1 mm between 2 groups were seen in 3 of the 7 soft tissue measurements, including tip of nose, soft tissue A point, and upper lip. The most accurate region was located at soft tissue A point. The largest differences were shown in the region of the lower lip. In general, the predictions tended to underestimate the amount of soft tissue retraction except for the subnasale and soft tissue pogonion.
337
LU, KO, AND HUANG
Table 3. DIFFERENCES BETWEEN THE ACTUAL AND PREDICTED LINE DRAWINGS
Tip of nose Subnasale Soft tissue A point Upper lip Lower lip Soft tissue B point Soft tissue pogonion
Sagittal Plane (mm) X ⫾ SD
Vertical Plane (mm) X ⫾ SD
⫹0.5 ⫾ 1.2 ⫺1.7 ⫾ 2.1 ⫹0.1 ⫾ 2.0 ⫹0.8 ⫾ 2.7 ⫹4.0 ⫾ 2.3 ⫹3.2 ⫾ 3.0 ⫺1.3 ⫾ 3.2
⫺0.5 ⫾ 1.5 ⫺0.8 ⫾ 1.5 ⫹1.0 ⫾ 2.6 ⫹1.7 ⫾ 2.4 ⫹0.3 ⫾ 3.6 ⫺0.1 ⫾ 3.4 ⫹0.8 ⫾ 3.7
Abbreviations: ⫹, predicted landmarks were anterior (sagittal) or inferior (vertical) to the actual one; ⫺, predicted landmarks were posterior (sagittal) or superior (vertical) to the actual one. Values are given as mean (average of differences between the prediction and actual final result) ⫾ SD (standard deviation of differences between prediction and actual final result).
The distribution of the predicted errors (differences between computer-generated prediction and actual postsurgical result) in the sagittal plane were shown in Table 4. The data were divided into 3 categories (error ⬍1 mm, error between 1 and 2 mm, and error ⬎2 mm). The distribution of the data presented wide range of standard deviation with significant bipolar spread, especially in the region of the lower lip. There were 80% with a difference of more than 2 mm in the lower lip region. The most reliable region of the software prediction was located at the tip of the nose, with a difference less than 1 mm in 63% and of less than 2 mm in 90%. On average, the upper lip region, including subnasale, soft tissue A point, and upper lip, presented 53% of predicted errors of less than 2 mm. The lower lip region, including the lower lip and soft tissue B point, showed predicted errors of less than 2 mm in 27%. The overall region presented 49% of predicted errors of less than 2 mm in sagittal plane. VERTICAL PLANE
When comparing the landmarks located in the computer-generated prediction with the actual profile
change at vertical plane (Table 3), the mean differences smaller than 1 mm between 2 groups were seen in 6 of the 7 soft tissue measurements, including the tip of the nose, subnasale, soft tissue A point, lower lip, soft tissue B point, and soft tissue pogonion. In general, the differences noted in the vertical plane were smaller than those found in the sagittal plane. The greatest differences were seen in the region of the upper lip with an average of 1.7 mm and tended to underestimate the amount of soft tissue impaction compared with the actual ones. The most accurate software prediction was located at soft tissue B point. The distributions of the predicted errors in the vertical plane were presented in Table 4. The distributions of data were more concentrated in errors smaller than 2 mm compared with the sagittal plane. In the vertical plane, the most reliable regions of the software prediction were located at the tip of the nose (Prn) and subnasale (Sn). For the tip of the nose, there were 60% with a difference of less than 1 mm and 80% with a difference of less than 2 mm. The least reliable region was located at the soft tissue pogonion. Only 10% of the predicted error was less than 1 mm in difference. In average, the upper lip region showed 62% of predicted errors less than 2 mm. The lower lip region, including lower lip and soft tissue B point, showed 53% of predicted errors less than 2 mm. The overall region presented 58% of predicted errors less than 2 mm in vertical plane. When averaging the errors of sagittal and vertical planes in all landmarks, it showed 54% of predicted errors less than 2 mm. The upper lip and lower lip regions presented 58% and 40% of predicted errors less than 2 mm, respectively. LINEAR AND ANGULAR
The predicted errors of other linear and angular measurements are listed in Table 5. The computergenerated lips are located in a more protrusive position than the actual ones. The distribution of the predicted error related to the upper lip was more
Table 4. FREQUENCY OF PREDICTED ERRORS
Sagittal (%)
Tip of nose Subnasale Soft tissue A point Upper lip Lower lip Soft tissue B point Soft tissue pogonion Overall
Vertical (%)
⬍1 mm
1-2 mm
⬎2 mm
⬍1 mm
1-2 mm
⬎2 mm
63 17 33 33 7 13 23 27
27 27 37 13 13 20 20 22
10 57 30 53 80 67 57 51
60 63 27 30 17 23 10 33
20 13 23 30 37 30 23 25
20 23 50 40 47 47 67 42
NOTE. Predicted errors were divided into 3 categories: error ⬍1 mm, error between 1 and 2 mm, and error ⬎2 mm. Overall value is the average of all predicted errors.
338
THE ACCURACY OF VIDEO IMAGING PREDICTION
Table 5. LINEAR AND ANGULAR DIFFERENCES AND FREQUENCY OF PREDICTED ERRORS
Upper Lower Upper Lower
Linear
Mean ⫾ SD (mm)
⬍2 mm (%)
2-4 mm (%)
⬎4 mm (%)
lip lip lip lip
⫹1.7 ⫾ 1.6 ⫹4.2 ⫾ 2.6 ⫹2.9 ⫾ 1.7 ⫹5.4 ⫾ 2.7
40 27 40 13
43 27 37 20
17 47 23 67
to to to to
E plane E plane GPg⬘ GPg⬘
Angular
Mean ⫾ SD (°)
⬍5° (%)
5°-10° (%)
⬎10° (%)
Facial convexity Nasolabial angle
⫺2.2 ⫾ 5 ⫺10.8 ⫾ 9.9
67 13
30 27
3 60
NOTE. Predicted errors were divided into 3 categories: error ⬍2 mm, error between 2 and 4 mm, and error ⬎4 mm in linear measurement, and error ⬍5°, error between 5° and 10°, and error ⬎10° in angular measurement. Abbreviations: G, glabella; Pg⬘, soft tissue pogonion.
concentrated in errors smaller than 4 mm, whereas the lower lip had a greater range of errors. The prediction in angular measurements tended to be underestimated compared with the actual surgical change. The distribution of data in the nasolabial angle and lower lip position presented a great bipolar spread. COMPARISON BETWEEN GROUPS WITH AND WITHOUT GENIOPLASTY
Comparison of the predicted errors between the patients with and without genioplasty procedure re-
vealed no statistically significant differences (Table 6) except the linear distance of the lower lip to the E plane. The genioplasty group showed greater predicted errors than the nongenioplasty group by 2.2 mm. The significance level was borderline (P ⫽ .05). The distributions of the predicted errors were plotted as scattergrams (Fig 2). The individual points in the scattergrams were obtained with the coordinates of the predicted landmarks minus those of the actual final landmarks. It showed that the predicted error distribution of the tip of nose and subnasale was more
Table 6. COMPARISON BETWEEN GROUPS WITH AND WITHOUT GENIOPLASTY
Soft tissue landmarks Sagittal plane Tip of nose Subnasale Soft tissue A point Upper lip Lower lip Soft tissue B point Soft tissue pogonion Vertical plane Tip of nose Subnasale Soft tissue A point Upper lip Lower lip Soft tissue B point Soft tissue pogonion Linear Upper lip to E plane Lower lip to E plane Upper lip to GPg⬘ Lower lip to GPg⬘ Angular Facial convexity Nasolabial angle
No genioplasty (mean [n ⫽ 23] ⫾ SD mm)
Genioplasty (mean [n ⫽ 7] ⫾ SD mm)
P value
⫹0.4 ⫾ 1.2 ⫺1.6 ⫾ 2.3 ⫹0.2 ⫾ 2.2 ⫹0.9 ⫾ 2.9 ⫹3.8 ⫾ 1.9 ⫹2.9 ⫾ 2.5 ⫺1.1 ⫾ 3.1
⫹0.8 ⫾ 1.1 ⫺2.1 ⫾ 1.1 ⫹0.0 ⫾ 1.0 ⫹0.6 ⫾ 1.8 ⫹4.7 ⫾ 3.3 ⫹4.2 ⫾ 4.4 ⫺2.1 ⫾ 3.8
.50 .43 .75 .76 .50 .47 .57
⫺0.5 ⫾ 1.5 ⫺0.7 ⫾ 1.3 ⫹1.4 ⫾ 2.6 ⫹1.7 ⫾ 2.4 ⫺0.2 ⫾ 3.3 ⫺0.3 ⫾ 3.3 ⫹0.6 ⫾ 3.8
⫺0.7 ⫾ 1.5 ⫺1.1 ⫾ 2.3 ⫺0.4 ⫾ 2.3 ⫹1.6 ⫾ 2.7 ⫹1.8 ⫾ 4.3 ⫹0.4 ⫾ 4.4 ⫹1.3 ⫾ 3.7
.76 .71 .12 .93 .29 .71 .69
⫹1.7 ⫾ 2.9 ⫹3.7 ⫾ 2.5 ⫹2.8 ⫾ 1.9 ⫹4.9 ⫾ 2.5
⫹1.6 ⫾ 1.6 ⫹5.9 ⫾ 2.3 ⫹3.0 ⫾ 1.0 ⫹7.2 ⫾ 2.8
.88 .05* .69 .08
⫺2.1 ⫾ 5.3 ⫺10.7 ⫾ 8.5
⫺2.4 ⫾ 4.2 ⫺11.1 ⫾ 14.2
.89 .94
Abbreviations: G, glabella; Pg⬘, soft tissue pogonion. *P ⱕ .05.
339
LU, KO, AND HUANG
FIGURE 2. Scattergrams of predicted errors. The individual points in the scattergrams were obtained by the coordinates of the predicted landmarks minus those of the actual final landmarks. Positive value indicates the predicted landmarks were anterior (x-axis) or inferior (y-axis) to the actual ones. Negative value indicates the predicted landmarks were posterior (x-axis) or superior (y-axis) to the actual ones.
accurate and concentrated. The tip of the nose, the lower lip, and soft tissue B point were estimated to a more anterior position. Subnasale was estimated to a more superior and posterior position. Upper lip was estimated to a more inferior position. Two case examples of actual surgical profile changes and predicted profile image are presented in Figures 3 and 4.
Discussion The results of this study showed that the software predictions of surgical profile changes were more accurate in the vertical plane than in the sagittal plane. The results were expected because the surgery mainly involves the sagittal algorithms. The tip of nose was the most reliable region that the software could predict. The treatment simulation of lower lip was shown to be more anterior than the actual postsurgical result. These findings were similar to those of several clinical studies.22-29 There was only one ex-
ception, when Sinclair et al12 used the program Portrait to predict the postsurgical outcome of patient with mandibular advancement. Their results indicated that the prediction of lower lip position was significantly posterior to the actual postsurgical result. For the soft tissue pogonion, our study indicated that the predicted position was more posteriorly located than the actual result. This finding was similar to the results of Kazandjian et al,27 whose surgical method was mandibular setback surgery. The least accurate predicted landmark was the lower lip measured in the sagittal plane. This finding had general agreement with several clinical studies, including Hing,22 Konstiantos et al,23 Syliangco et al,25 Sameshima et al,26 and Kazandjian et al,27 although various surgical methods and different prediction software were used (Table 1). The lower lip showed the least accurate prediction. This could be explained in several ways: the lower lip is pliable and subject to the influence of incisor position and angulation, soft tissue thickness and tonicity, perioral musculature,
340
THE ACCURACY OF VIDEO IMAGING PREDICTION
FIGURE 3. A 28-year-old female patient underwent Wassmund and Ko¨ le procedures without genioplasty. From left to right is shown presurgical image computer-generated prediction, and actual final image. Note the differences of the lips between the predicted and actual images. The lips thickness was not compressed in the predicted image.
and underlying muscle attatchments.25,30 In comparison with to previous studies22,23,25-27 (Table 1), the accuracy of lower lip prediction in the sagittal plane was less in our study. Previous studies mainly evaluated the soft tissue changes after single jaw surgery. However, this study investigated the change after 2-jaw surgery, which made the prediction of lower lip more complicated. The concurrent backward movement of the soft tissue with the underlying hard tissue did not fully express the change in lip thickness in this software
FIGURE 4. A 34-year-old female patient underwent Wassmund and Ko¨ le procedures without genioplasty. From left to right is shown presurgical image, computer-generated prediction, and actual final image. Note the mentalis strain was relieved after the surgery; it is persisted in the predicted image.
(Fig 3). The soft tissue change in the profile photograph seemed to move in a single graphic piece without morphing within it. Therefore, the relief of muscle strain at lip region after backward jaw movement was not accurately expressed in this computer program (Fig 4). Kazandjian et al27 also indicated this phenomenon in their study. The average soft tissue– to– hard tissue ratio was generally used in imaging software programs to generate the predicted images. Such methods only portray the change of the facial outline and do not include the change in individual
341
LU, KO, AND HUANG
soft tissue posture. The soft tissue thickness and muscle tonicity may not be shown precisely in the predicted images. Errors in linear measurement showed that the soft tissue prediction tended to overestimate the distance from lips to the facial plane (glabella to soft tissue pogonion) and those to the E plane (tip of nose to soft tissue pogonion). This would make the predicted lips look more protrusive than the actual ones. The results were similar to those of Upton,24 who presented that the predicted distance from lips to E plane were larger than actual. However, Sinclair et al12 indicated an underestimation of the prediction value from lips to E plane. The measurements of angular prediction tended to be underestimated in facial convexity and nasolabial angle. The result was also similar to those of Upton et al,24 who showed that the actual postsurgical nasolabial angle was larger than the software prediction. In the surgical treatment of bimaxillary protrusion, the frequency with that the software was able to make accurate postsurgical predictions varied. In average, the upper lip region presented 58% of predicted errors less than 2 mm. The lower lip region showed 40% of predicted errors less than 2 mm. These results were similar to the study of Kazandjian et al,27 who used Quick Ceph (Orthodontic Processing, San Diego, CA) and Portrait Planner (Rx Data Inc, Ooltewah, TN) to evaluate the accuracy of video imaging for predicting the soft tissue profile after mandibular setback surgery, with 56% and 56% of predicted errors less than 2 mm in upper lip region and 57% and 48% of predicted errors less than 2 mm in the lower lip region, respectively. On the other hand, Syliangco et al25 used Orthognathic Treatment Planner (Pacific Coast Software, Pacific Palisades, CA) and Prescription Portrait (Rx Data Inc) to predict soft tissue changes in mandibular advancement, showed more accurate predictions in their study, with 98% and 96% of predicted errors less than 2 mm in the upper lip region and 69% and 78% of predicted errors less than 2 mm in the lower lip region, respectively. The various results could be explained that our study and the study of Kazandjian et al27 mainly involve compression of the image and the study of Syliangco et al25 mainly related to expansion of the image. This indicated that different direction of hard tissue movement may influence the accuracy of the prediction programs. In this study, there were no statistically significant differences in the predicted errors between the group with and without genioplasty. This result was similar to the result of Upton et al,24 who used Quick Ceph to predict the postsurgical soft tissue profile; they also did not find any statistically significant differences between the patients with and without genioplasty.
However, in our study, the sample size of the genioplasty group was relatively smaller. The clinical significance needs to be further investigated.
References 1. Kiyak HA, Bell R: Psychosocial considerations in surgery and orthodontics, in Proffit WR, White RP Jr (eds): Surgical Orthodontic Treatment. St Louis, MO, Mosby–Year Book, 1991, pp 71-95 2. Proffit WR, White RP: The need for surgical orthodontic treatment, in Proffit WR, White RP Jr (eds): Surgical Orthodontic Treatment. St Louis, MO, Mosby–Year Book, 1991, pp 2-24 3. Sarver DM, Johnston MW, Maturas VJ: Video-imaging for planning and counseling in ortho-surgery. J Oral Maxillofac Surg 46:939, 1988 4. Sarver DM, Johnston MW: Video-imaging: Techniques for superimposition of cephalometric radiography and profile images. Int J adult Ortho Orthognath Surg 5:241, 1990 5. Sarver DM, Matukas VJ, Weissman SM: Incorporation of plastic surgery in the planning and treatment of orthognathic surgical cases. Int J Adult Orthod Orthognath Surg 6:227, 1991 6. Ackerman JL, Proffit WR: Communication in orthodontic treatment planning: Bioethical and informed consent issues. Angle Orthod 65:253, 1995 7. Turpin D: The need for video imaging. Angle Orthod 65:243, 1995 8. Proffit WR: Treatment planning: The search for wisdom, in Proffit WR, White RP Jr (eds): Surgical Orthodontic Treatment. St Louis, MO, Mosby–Year Book, 1991, pp 142-191 9. Cohen M: Mandibular prognathism. Am J Orthod 51:368, 1965 10. McNeil RW, Proffit WR, White RP: Cephalomertic prediction for orthodontic surgery. Angle Orthod 42:154, 1972 11. Henderson D: The assessment and management of bony deformities of the middle and lower face. Br. J Plast Surg 24:287, 1974 12. Sinclair PM, Kilpelainen P, Phillips C, et al: The accuracy of video imaging in orthognathic surgery. Am J Orthod Dentofac Orthop 107:177, 1995 13. Bhatia SN, Sowray JH: A computer-aided design for orthognathic surgery. Br J Oral Maxillofac Surg 22:237, 1984 14. Harradine NWT, Birnie DJ: Computerized prediction of the results of orthognathic surgery. J Maxillofac Surg 13:245, 1985 15. Walters H, Walters DH: Computerised planning of maxillofacial osteotomies: The program and its clinical application. Br J Oral Maxillofac Surg 24:178, 1986 16. Lew KK: The reliability of computerized soft tissue prediction following bimaxillary anterior subapical osteotomy. Int J Adult Orthod Orthognath Surg 7:97, 1992 17. Kinnebrew MC, Hoffnab DR, Carlton DM: Projecting the softtissue outcome of surgical and orthodontic manipulation of the maxillofacial skeleton. Am J Orthod 84:508, 1983 18. Laney TJ, Kuhn BS: Computer imaging in orthognathic and facial cosmetic surgery. Oral Maxillofac Clin North Am 2:659, 1990 19. Takahashi I, Takahashi T, Mitsuhiko H, et al: Application of video surgery to orthodontic diagnosis. Int J Adult Orthod Orthognath Surg 4:219, 1989 20. Turpin DL: Computers coming on line for diagnosis and treatment planning. Angle Orthod 60:163, 1990 21. Phillips C, Hill BJ, Cannac C: The influence of video imaging on patients’ perceptions and expectations. Angle Orthod 65:263, 1995 22. Hing NR: The accuracy of computer generated prediction tracings. Int J Oral Maxillofac Surg 18:148, 1989 23. Konstiantos KA, O’Reilly MT, Close J: The validity of the prediction of soft tissue profile changes after LeFort I osteotomy using the Dentofacial Planner (computer software). Am J Orthod Dentofac Orthop 105:241, 1994 24. Upton PM: Evaluation of video imaging prediction in combined maxillary and mandibular orthognathic surgery. Am J Orthod Dentofac Orthop 106:451, 1994 25. Syliangco ST, Sameshima GT, Kaminishi RM, et al: Predicting soft tissue changes in mandibular advancement surgery: A
342 comparison of two video imaging systems. Angle Orthod 67: 337, 1997 26. Sameshima GT, Kawakami RK, Kaminishi RM, et al: Predicting soft tissue changes in maxillary impaction surgery: A comparison of two video imaging systems. Angle Orthod 67:347, 1997 27. Kazandjian S, Sameshima GT, Champlin T, et al: Accuracy of video imaging for predicting the soft tissue profile after mandibular set-back surgery. Am J Orthod Dentofacial Orthop 115: 382, 1999
THE ACCURACY OF VIDEO IMAGING PREDICTION 28. Lines PA, Steinhauser WW: Soft tissue changes in relationship of movement of hard structures in orthognathic surgery. J Oral Surg 32:891, 1974 29. Quast DC, Biggerstaff RH, Haley JV: The short and long-term soft-tissue profile changes accompanying mandible advancement surgery. Am J Orthod 84:29, 1983 30. Stella JP, Streater MR, Epker BN, et al: Predictability of upper soft tissue changes with maxillary advancement. J Oral Maxillofac Surg 47:697, 1989
The Accuracy of Video Imaging Prediction in Soft Tissue Outcome After Bimaxillary Orthognathic Surgery Chien-Hsun Lu, DDS,* Ellen Wen-Ching Ko, DDS, MS,† and Chiung-Shing Huang, DDS, PhD‡ Purpose:
The purpose of the present study was to evaluate the accuracy of the outcome in soft tissue prediction through use of a computer imaging system after bimaxillary orthognathic surgery. Materials and Methods: The study sample consisted of 30 adults who had undergone orthognathic surgery that included the Wassmund and Ko ¨ le procedures and optional genioplasty to correct bimaxillary protrusion. All the patients had lateral cephalometric radiographs and profile photographs taken within 6 months before surgery and at least 6 months after surgery. The computer-generated soft tissue image and the actual postsurgical profile were compared. The accuracy of this computer-generated profile image was evaluated. Results: The results indicated that the nasal tip, soft tissue A point, and upper lip presented the least predicted errors in sagittal plane. While the nasal tip presented higher reliability. Lower lip prediction was found to be the least accurate region and it tended to be located anterior to the actual position. In the vertical plane, most of the predictions revealed higher accuracy than those in the sagittal plane. There were no statistically significant differences between the predictions of the groups with and those without genioplasty. Conclusions: Computer-generated image prediction was suitable for patient education and communication. However, efforts are still needed to improve the accuracy and reliability of the prediction program and to include the consideration of changes in soft tissue tension and muscle strain. The accuracy of this system in soft tissue prediction should be carefully interpreted. © 2003 American Association of Oral and Maxillofacial Surgeons J Oral Maxillofac Surg 61:333-342, 2003 Improvement of facial aesthetics is a main reason patients request surgical correction of dentofacial deformities.1 The definition of an ideal result of facial improvement after surgical orthodontic treatment is very subjective. The evaluation is affected by different opinions and treatment philosophies. Therefore, a successful orthognathic surgery includes not only the precise surgical technique and occlusal correction
but also the accomplishment of the aesthetic goals that are satisfying to both patients and professionals.2-6 It is hard for the public to imagine the change in facial appearance after orthognathic surgery without a visual reference.6,7 Lateral cephalometric radiographs are commonly used to predict the surgical treatment outcomes. The cephalometric prediction can be performed manually by repositioning an overlaid tracing or by moving templates.8-12 It involves cutting different parts of the acetate tracing and repositioning them over the original cephalometric tracing to simulate the surgical treatment. The predicted posttreatment soft tissue outline is delineated based on the published ratios of soft tissue to hard tissue changes. With the development of computer science, cephalometric landmarks were able to be digitized into the computer and the repositioning of the surgical parts can now be viewed immediately on the screen.13-16 The predicted outline of the posttreatment facial profile is generated by specialized computer software. The 2 aforementioned methods pre-
Received from Chang Gung Memorial Hospital, Taipei, Taiwan. *Attending Staff, Department of Orthodontics and Craniofacial Dentistry. †Attending Staff, Department of Orthodontics and Craniofacial Dentistry. ‡Chairman and Professor, Faculty of Dentistry. Address correspondence and reprint requests to Dr Ko: Department of Orthodontics and Craniofacial Dentistry, Chang Gung Memorial Hospital, 199 Tung-Hwa North Rd, Taipei, 105, Taiwan; e-mail: [email protected] © 2003 American Association of Oral and Maxillofacial Surgeons
0278-2391/03/6103-0009$30.00/0 doi:10.1053/joms.2003.50058
333
334
THE ACCURACY OF VIDEO IMAGING PREDICTION
Table 1. LITERATURE REVIEW OF THE LEAST ACCURATE REGION IN SOFTWARE PREDICTION OF FACIAL SOFT TISSUE CHANGES AFTER ORTHOGNATHIC SURGERY
Author
Least Accurate Region
Predicted Error (mm)
Sample Size
Kazandjian et al27
Mandibular setback
QC, PoP
32
Maxillary impaction
OTP, PrP
Syliangco et al25
Lower lip
39
Mandibular advancement
OTP, PrP
Konstiantos et al23
Lower lip Pronasal Lower lip Chin
⫹2.25 ⫾ 3.63 (H) ⫺2.23 ⫾ 2.85 (V) ⫹1.71 ⫾ 1.37 (H) ⫹2.88 ⫾ 2.62 (V) ⫹1.61 ⫾ 1.24 (H) ⫹1.77 ⫾ 1.26 (V) ⫹1.57 ⫾ 2.0 (H) ⫺1.64 ⫾ 1.4 (V) ⫹1.9 ⫾ 0.38 (H) ⫺0.5 ⫾ 0.78 (V)
30
Sameshima et al26
Upper lip Lower lip Lower lip
21
Le Fort I osteotomy
DP
16
Mandibular advancement
QC
Hing22
Type of Surgery
Software
Abbreviations: H, horizontal plane; V, vertical plane; ⫹, predicted landmarks anterior (horizontal) or inferior (vertical) to the actual one; ⫺, predicted landmarks posterior (horizontal) or superior (vertical) to the actual one; QC, Quick Ceph; PoP, Portrait Planner; OTP, Orthognathic Treatment Planner; PrP, Prescription Portrait; DP, Dentofacial Planner (DentoFacial Software, Inc, Toronto, Canada).
sented only the “line drawing” profile of the surgical simulation. The linear presentation of the facial profile (although it is relatively informative to the professionals) cannot provide a realistic image of the treatment results to the patients and the public.12 Manipulation of patients’ photographs to illustrate treatment goals provides the patients with a more vivid image.17 However, it is believed to be more of an art form than a scientific exercise. Recent prediction techniques integrate video images with cephalograms.3,4,18-20 The patient’s lateral photograph is superimposed on the digitized cephalometric tracing. The computer-generated prediction includes both line drawing tracing and the corresponding facial image. With the facial image, the patient can have an idea of the possible changes of their facial appearance. The predicted image also facilitates the communication between orthodontists, surgeons, and patients.3,6,7 Several studies point to the influence of video image on treatment plans. Sarver et al3 evaluated the treatment result 4 months after surgery; it revealed that 89% of the patients thought that the image predictions were realistic and that the desired results were achieved. It also assisted 83% of the patients to make a decision regarding whether to undergo orthognathic surgery. None of the patients believed that the imaging session had no value. When using video imaging as a presentation tool, Phillips et al21 found that in the standard case presentation group (the group without video image presentation), 47% of the patients ranked dental casts as the most helpful physical record in making the treatment decision and 46% of the patients considered that the acetate profile tracing was the most helpful tool for understanding the treatment expectation. However, in the video image presentation group (the group including video image presentation), the results indicated that 42% of
the patients thought the video imaging presentation was the most helpful tool in decision making and that 39% of the subjects considered it to be the most helpful tool for understanding the treatment outcomes.21 The influence of video imaging on patients’ perception was significant. Showing the video imaging simulation to patients should be done conservatively and carefully to avoid an unrealistic expectation of the final results.12 Phillips et al21 compared the final treatment outcomes in 2 groups: the video image case presentation group and the standard case presentation group; the previous one presented a significantly higher self-image expectation. The authors thought that the visual display of treatment simulation was more realistic than the “line drawing” tracing graph and much more accepted by the public for the actual surgical treatment results. Also, it provided the patients with a more appropriate appreciation for the changes in appearance after the orthognathic surgery. When comparing the quality of various predicted images, Sinclair et al12 found that none of the predicted images were more favorable than the actual profile changes after orthognathic surgery. The accuracy of the computer-predicted images after orthognathic surgery is examined in many studies.12,22-27 It seems that the lips was the least accurate region that the computer could predict regardless of the kind of surgical methods that were performed and the kind of prediction software that was used (Table 1). However, most of the studies evaluated only one jaw surgery. The precision of treatment simulation in simultaneous maxillary and mandibular setback procedure has not been investigated. The purpose of this study was to evaluate the accuracy of a video imaging program in predicting soft tissue outcomes after orthognathic surgery in patents with bimaxillary protrusion who underwent simultaneous maxillary and mandibular setback.
335
LU, KO, AND HUANG
FIGURE 1. A, Cephalometric landmarks used in this study: S, sella; N, nasion; Po, porion; G, glabella; Prn, tip of nose; Sn, subnasale; A⬘, soft tissue A point; UL, upper lip; LL, lower lip; B⬘, soft tissue B point; Pg⬘, soft tissue pogonion; Pg, pogonion; LIX, lower incisal apex; LI, lower incisal tip; UI, upper incisal tip; UIX, upper incisal apex. B, Linear and angular measurement: 1, facial convexity; 2, nasolabial angle; 3, upper lip to facial plane; 4, lower lip to facial plane; 5, upper lip to E plane; 6, lower lip to E plane (facial plane: glabella to soft tissue pogonion; E plane: tip of nose to soft tissue pogonion).
Materials and Methods The sample consisted of 30 patients who met the following criteria: 1) adult older than 19 years (average age, 27.9 years; age range, 19 to 37 years); 2) underwent Wassmund procedure to setback anterior maxilla and Ko ¨ le procedure to setback anterior mandible, with or without genioplasty; and 3) no congenital craniofacial deformities or trauma and head and neck surgical history. The diagnostic records included the lateral cephalometric radiographs and profile photographs within 6 months before surgery and at least 6 months after surgery. The head films and photographs were acquired in the natural head position with teeth in centric occlusion and relaxed lip posture. There were no intraoral fixed orthodontic appliances shown in both the head films and photographs. All of the patients were treated in the Craniofacial Center, Chang Gung Memorial Hospital, Taipei, Taiwan. Both before and after cephalograms were traced on the acetate papers. All the cephalometric tracings and profile photographs were entered in a Pentium-based computer (Intel Pentium III, 500 MHz, Windows 98SE, Microsoft Corporation, Redmond, WA) through a scanner (Astra 2400S; Umax Technologies Inc, Taipei, Taiwan) for analysis. The Dolphin Imaging system software (Version 6; Dolphin Imaging, Canoga Park, CA) was used to store and generate image prediction.
METHOD OF PREDICTION
The sagittal and vertical treatment changes were evaluated by linear measurements within an X-Y coordinate system (Fig 1A). The SN plane was defined as the horizontal reference plane (x-axis), and a line perpendicular to this plane through sella was defined as the vertical reference plane (y-axis). Landmarks of sella (S), nasion (N), and porion (Po) in the presurgical tracing were all transferred to the postsurgical tracing in the same patient. The tracings of presurgical and postsurgical cephalograms were superimposed at the cranial base to ensure the X-Y planes been accurately transferred. The tracings and photographs were then input into the computer system, digitized, and superimposed following instructions of the Dolphin Imaging software. The skeletal and dental landmarks included anterior nasal spine (ANS), upper incisal edge (UI), lower incisal edge (LI), upper incisor root apex (UIX), lower incisal root apex (LIX), and pogonion (Pg) (Fig 1A). The perpendicular distance of each landmark to both reference planes (x- and y-axes) was recorded before and after surgery. The treatment changes of hard tissue in each case were obtained from the differences between presurgical and postsurgical linear measurements. The hard tissue image was moved according to the prescribed distances (the treatment change) by using these calculations and the VTO (visual treatment objectives) function in the software. The predicted soft tissue outline and the corresponding coordinates of the soft tissue
336
THE ACCURACY OF VIDEO IMAGING PREDICTION
Table 2. SOFT TISSUE–TO–HARD TISSUE MOVEMENT RATIO SHOWN IN THE SYSTEM INITIATION FILE OF THE SOFTWARE
Soft Tissue–to–Hard Tissue Movement Ratio for Soft Tissue Landmark (X ratio, Y ratio)
Hard Tissue Movement Maxilla horizontal movement Maxilla rotational movement Anterior maxilla horizontal movement Anterior maxilla rotational movement Mandible horizontal movement Mandible rotational movement Mandible vertical movement Chin horizontal movement Chin rotational movement Chin vertical movement Symphysis horizontal movement Symphysis vertical movement Symphysis rotational movement Mandible rotational movement in orthodontic treatment
Subnasale (1:1,1:1), A⬘ (1:1,1:1), upper lip (1:1,1:1), stomion (1:1,1:1), stomion superior (1:1,1:1) Subnasale (1:1,1:1), A⬘ (1:1,1:1), upper lip (1:1,1:1), stomion (1:6,1:6), stomion superior (1:1,1:1) Subnasale (1:1,1:1), A⬘ (1:1,1:1), upper lip (1:1,1:1), stomion (1:1,1:1), stomion superior (1:1,1:1) Subnasale (1:1, 1:1), A⬘ (1:1,1:1), upper lip (1:1,1:1), stomion (1:1,1:1), stomion superior (1:1,1:1) Stomion inferior (1:1, 1:1), lower lip (1:1,1:1), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1), stomion inferior (1:1,1:1) Stomion inferior (1:1, 1:1), lower lip (1:1,1:1), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1) Lower lip (1:1,1:1), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1), stomion (1:1,1:1), stomion inferior (1:1,1:1) B⬘ (1:2,1:2), Pg⬘ (1:1,1:1) Pg⬘ (1:1,1:1) Pg⬘ (1:1,1:1) Lower lip (1:1,1:1), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1), stomion (1:1,1:1), stomion inferior (1:1,1:1) Lower lip (1:2,1:2), B⬘ (1:1,1:1), Pg⬘ (1:1,1:1), stomion (1:1,1:1), stomion inferior (1:1,1:1) Lower lip (1:2,1:2), B⬘ (1:2,1:2), Pg⬘ (1:1,1:1), stomion inferior (1:1,1:1) Pg⬘ (1:1,1:1), B⬘ (1:1,1:1), upper lip (1:1,1:1)
Abbreviations: X ratio, soft tissue–to– hard tissue movement ratio in horizontal direction; Y ratio, soft tissue–to– hard tissue movement ratio in vertical direction; B⬘, soft tissue B point; Pg⬘, pogonion.
points were automatically generated. The ratios of soft tissue to hard tissue movement are varied according to the specific parts of the hard tissue. The prescription of these ratios is shown in the system initiation file (dolphin.ini) of the Dolphin imaging system (Table 2). The differences in soft tissue outline between predicted tracing and the actual profile were compared (predicted error) to test the accuracy of this system. Seven soft tissue landmarks, including tip of the nose (Prn), subnasale (Sn), soft tissue A point (A⬘), upper lip (UL), lower lip (LL), soft tissue B point (B⬘), and soft tissue pogonion (Pg⬘) were evaluated (Fig 1A). The perpendicular distance of each landmark to the X-Y planes was measured. Also, 4 linear and 2 angular measurements were observed, including upper lip protrusion (perpendicular distance from upper vermilion border to the line connecting soft tissue glabella and Pg⬘), lower lip protrusion (perpendicular distance from lower vermilion border to the line connecting soft tissue glabella and Pg⬘), upper lip to E plane (perpendicular distance form upper vermilion border to the plane connecting the Prn and Pg⬘), lower lip to E plane (perpendicular distance form lower vermilion border to the plane connecting the Prn and Pg⬘), facial convexity (angle formed by glabella-Sn, Sn-Pg⬘), and nasolabial angle (Fig 1B).
METHOD ERROR
Errors associated with digitization and methods of measurement were assessed by retracing and redigitizing on 10 randomly selected cephalometric films. The average differences in these 2 sets of linear measurement were within 0.6 mm. The average differences in these 2 sets of angular measurements were within 1.15o. There were no statistically significant differences between each measurement by comparison with paired t-test (P ⬎ .5).
Results SAGITTAL PLANE
When comparing the landmarks located in the computer-generated prediction with the actual profile change at sagittal plane (Table 3), the mean differences smaller than 1 mm between 2 groups were seen in 3 of the 7 soft tissue measurements, including tip of nose, soft tissue A point, and upper lip. The most accurate region was located at soft tissue A point. The largest differences were shown in the region of the lower lip. In general, the predictions tended to underestimate the amount of soft tissue retraction except for the subnasale and soft tissue pogonion.
337
LU, KO, AND HUANG
Table 3. DIFFERENCES BETWEEN THE ACTUAL AND PREDICTED LINE DRAWINGS
Tip of nose Subnasale Soft tissue A point Upper lip Lower lip Soft tissue B point Soft tissue pogonion
Sagittal Plane (mm) X ⫾ SD
Vertical Plane (mm) X ⫾ SD
⫹0.5 ⫾ 1.2 ⫺1.7 ⫾ 2.1 ⫹0.1 ⫾ 2.0 ⫹0.8 ⫾ 2.7 ⫹4.0 ⫾ 2.3 ⫹3.2 ⫾ 3.0 ⫺1.3 ⫾ 3.2
⫺0.5 ⫾ 1.5 ⫺0.8 ⫾ 1.5 ⫹1.0 ⫾ 2.6 ⫹1.7 ⫾ 2.4 ⫹0.3 ⫾ 3.6 ⫺0.1 ⫾ 3.4 ⫹0.8 ⫾ 3.7
Abbreviations: ⫹, predicted landmarks were anterior (sagittal) or inferior (vertical) to the actual one; ⫺, predicted landmarks were posterior (sagittal) or superior (vertical) to the actual one. Values are given as mean (average of differences between the prediction and actual final result) ⫾ SD (standard deviation of differences between prediction and actual final result).
The distribution of the predicted errors (differences between computer-generated prediction and actual postsurgical result) in the sagittal plane were shown in Table 4. The data were divided into 3 categories (error ⬍1 mm, error between 1 and 2 mm, and error ⬎2 mm). The distribution of the data presented wide range of standard deviation with significant bipolar spread, especially in the region of the lower lip. There were 80% with a difference of more than 2 mm in the lower lip region. The most reliable region of the software prediction was located at the tip of the nose, with a difference less than 1 mm in 63% and of less than 2 mm in 90%. On average, the upper lip region, including subnasale, soft tissue A point, and upper lip, presented 53% of predicted errors of less than 2 mm. The lower lip region, including the lower lip and soft tissue B point, showed predicted errors of less than 2 mm in 27%. The overall region presented 49% of predicted errors of less than 2 mm in sagittal plane. VERTICAL PLANE
When comparing the landmarks located in the computer-generated prediction with the actual profile
change at vertical plane (Table 3), the mean differences smaller than 1 mm between 2 groups were seen in 6 of the 7 soft tissue measurements, including the tip of the nose, subnasale, soft tissue A point, lower lip, soft tissue B point, and soft tissue pogonion. In general, the differences noted in the vertical plane were smaller than those found in the sagittal plane. The greatest differences were seen in the region of the upper lip with an average of 1.7 mm and tended to underestimate the amount of soft tissue impaction compared with the actual ones. The most accurate software prediction was located at soft tissue B point. The distributions of the predicted errors in the vertical plane were presented in Table 4. The distributions of data were more concentrated in errors smaller than 2 mm compared with the sagittal plane. In the vertical plane, the most reliable regions of the software prediction were located at the tip of the nose (Prn) and subnasale (Sn). For the tip of the nose, there were 60% with a difference of less than 1 mm and 80% with a difference of less than 2 mm. The least reliable region was located at the soft tissue pogonion. Only 10% of the predicted error was less than 1 mm in difference. In average, the upper lip region showed 62% of predicted errors less than 2 mm. The lower lip region, including lower lip and soft tissue B point, showed 53% of predicted errors less than 2 mm. The overall region presented 58% of predicted errors less than 2 mm in vertical plane. When averaging the errors of sagittal and vertical planes in all landmarks, it showed 54% of predicted errors less than 2 mm. The upper lip and lower lip regions presented 58% and 40% of predicted errors less than 2 mm, respectively. LINEAR AND ANGULAR
The predicted errors of other linear and angular measurements are listed in Table 5. The computergenerated lips are located in a more protrusive position than the actual ones. The distribution of the predicted error related to the upper lip was more
Table 4. FREQUENCY OF PREDICTED ERRORS
Sagittal (%)
Tip of nose Subnasale Soft tissue A point Upper lip Lower lip Soft tissue B point Soft tissue pogonion Overall
Vertical (%)
⬍1 mm
1-2 mm
⬎2 mm
⬍1 mm
1-2 mm
⬎2 mm
63 17 33 33 7 13 23 27
27 27 37 13 13 20 20 22
10 57 30 53 80 67 57 51
60 63 27 30 17 23 10 33
20 13 23 30 37 30 23 25
20 23 50 40 47 47 67 42
NOTE. Predicted errors were divided into 3 categories: error ⬍1 mm, error between 1 and 2 mm, and error ⬎2 mm. Overall value is the average of all predicted errors.
338
THE ACCURACY OF VIDEO IMAGING PREDICTION
Table 5. LINEAR AND ANGULAR DIFFERENCES AND FREQUENCY OF PREDICTED ERRORS
Upper Lower Upper Lower
Linear
Mean ⫾ SD (mm)
⬍2 mm (%)
2-4 mm (%)
⬎4 mm (%)
lip lip lip lip
⫹1.7 ⫾ 1.6 ⫹4.2 ⫾ 2.6 ⫹2.9 ⫾ 1.7 ⫹5.4 ⫾ 2.7
40 27 40 13
43 27 37 20
17 47 23 67
to to to to
E plane E plane GPg⬘ GPg⬘
Angular
Mean ⫾ SD (°)
⬍5° (%)
5°-10° (%)
⬎10° (%)
Facial convexity Nasolabial angle
⫺2.2 ⫾ 5 ⫺10.8 ⫾ 9.9
67 13
30 27
3 60
NOTE. Predicted errors were divided into 3 categories: error ⬍2 mm, error between 2 and 4 mm, and error ⬎4 mm in linear measurement, and error ⬍5°, error between 5° and 10°, and error ⬎10° in angular measurement. Abbreviations: G, glabella; Pg⬘, soft tissue pogonion.
concentrated in errors smaller than 4 mm, whereas the lower lip had a greater range of errors. The prediction in angular measurements tended to be underestimated compared with the actual surgical change. The distribution of data in the nasolabial angle and lower lip position presented a great bipolar spread. COMPARISON BETWEEN GROUPS WITH AND WITHOUT GENIOPLASTY
Comparison of the predicted errors between the patients with and without genioplasty procedure re-
vealed no statistically significant differences (Table 6) except the linear distance of the lower lip to the E plane. The genioplasty group showed greater predicted errors than the nongenioplasty group by 2.2 mm. The significance level was borderline (P ⫽ .05). The distributions of the predicted errors were plotted as scattergrams (Fig 2). The individual points in the scattergrams were obtained with the coordinates of the predicted landmarks minus those of the actual final landmarks. It showed that the predicted error distribution of the tip of nose and subnasale was more
Table 6. COMPARISON BETWEEN GROUPS WITH AND WITHOUT GENIOPLASTY
Soft tissue landmarks Sagittal plane Tip of nose Subnasale Soft tissue A point Upper lip Lower lip Soft tissue B point Soft tissue pogonion Vertical plane Tip of nose Subnasale Soft tissue A point Upper lip Lower lip Soft tissue B point Soft tissue pogonion Linear Upper lip to E plane Lower lip to E plane Upper lip to GPg⬘ Lower lip to GPg⬘ Angular Facial convexity Nasolabial angle
No genioplasty (mean [n ⫽ 23] ⫾ SD mm)
Genioplasty (mean [n ⫽ 7] ⫾ SD mm)
P value
⫹0.4 ⫾ 1.2 ⫺1.6 ⫾ 2.3 ⫹0.2 ⫾ 2.2 ⫹0.9 ⫾ 2.9 ⫹3.8 ⫾ 1.9 ⫹2.9 ⫾ 2.5 ⫺1.1 ⫾ 3.1
⫹0.8 ⫾ 1.1 ⫺2.1 ⫾ 1.1 ⫹0.0 ⫾ 1.0 ⫹0.6 ⫾ 1.8 ⫹4.7 ⫾ 3.3 ⫹4.2 ⫾ 4.4 ⫺2.1 ⫾ 3.8
.50 .43 .75 .76 .50 .47 .57
⫺0.5 ⫾ 1.5 ⫺0.7 ⫾ 1.3 ⫹1.4 ⫾ 2.6 ⫹1.7 ⫾ 2.4 ⫺0.2 ⫾ 3.3 ⫺0.3 ⫾ 3.3 ⫹0.6 ⫾ 3.8
⫺0.7 ⫾ 1.5 ⫺1.1 ⫾ 2.3 ⫺0.4 ⫾ 2.3 ⫹1.6 ⫾ 2.7 ⫹1.8 ⫾ 4.3 ⫹0.4 ⫾ 4.4 ⫹1.3 ⫾ 3.7
.76 .71 .12 .93 .29 .71 .69
⫹1.7 ⫾ 2.9 ⫹3.7 ⫾ 2.5 ⫹2.8 ⫾ 1.9 ⫹4.9 ⫾ 2.5
⫹1.6 ⫾ 1.6 ⫹5.9 ⫾ 2.3 ⫹3.0 ⫾ 1.0 ⫹7.2 ⫾ 2.8
.88 .05* .69 .08
⫺2.1 ⫾ 5.3 ⫺10.7 ⫾ 8.5
⫺2.4 ⫾ 4.2 ⫺11.1 ⫾ 14.2
.89 .94
Abbreviations: G, glabella; Pg⬘, soft tissue pogonion. *P ⱕ .05.
339
LU, KO, AND HUANG
FIGURE 2. Scattergrams of predicted errors. The individual points in the scattergrams were obtained by the coordinates of the predicted landmarks minus those of the actual final landmarks. Positive value indicates the predicted landmarks were anterior (x-axis) or inferior (y-axis) to the actual ones. Negative value indicates the predicted landmarks were posterior (x-axis) or superior (y-axis) to the actual ones.
accurate and concentrated. The tip of the nose, the lower lip, and soft tissue B point were estimated to a more anterior position. Subnasale was estimated to a more superior and posterior position. Upper lip was estimated to a more inferior position. Two case examples of actual surgical profile changes and predicted profile image are presented in Figures 3 and 4.
Discussion The results of this study showed that the software predictions of surgical profile changes were more accurate in the vertical plane than in the sagittal plane. The results were expected because the surgery mainly involves the sagittal algorithms. The tip of nose was the most reliable region that the software could predict. The treatment simulation of lower lip was shown to be more anterior than the actual postsurgical result. These findings were similar to those of several clinical studies.22-29 There was only one ex-
ception, when Sinclair et al12 used the program Portrait to predict the postsurgical outcome of patient with mandibular advancement. Their results indicated that the prediction of lower lip position was significantly posterior to the actual postsurgical result. For the soft tissue pogonion, our study indicated that the predicted position was more posteriorly located than the actual result. This finding was similar to the results of Kazandjian et al,27 whose surgical method was mandibular setback surgery. The least accurate predicted landmark was the lower lip measured in the sagittal plane. This finding had general agreement with several clinical studies, including Hing,22 Konstiantos et al,23 Syliangco et al,25 Sameshima et al,26 and Kazandjian et al,27 although various surgical methods and different prediction software were used (Table 1). The lower lip showed the least accurate prediction. This could be explained in several ways: the lower lip is pliable and subject to the influence of incisor position and angulation, soft tissue thickness and tonicity, perioral musculature,
340
THE ACCURACY OF VIDEO IMAGING PREDICTION
FIGURE 3. A 28-year-old female patient underwent Wassmund and Ko¨ le procedures without genioplasty. From left to right is shown presurgical image computer-generated prediction, and actual final image. Note the differences of the lips between the predicted and actual images. The lips thickness was not compressed in the predicted image.
and underlying muscle attatchments.25,30 In comparison with to previous studies22,23,25-27 (Table 1), the accuracy of lower lip prediction in the sagittal plane was less in our study. Previous studies mainly evaluated the soft tissue changes after single jaw surgery. However, this study investigated the change after 2-jaw surgery, which made the prediction of lower lip more complicated. The concurrent backward movement of the soft tissue with the underlying hard tissue did not fully express the change in lip thickness in this software
FIGURE 4. A 34-year-old female patient underwent Wassmund and Ko¨ le procedures without genioplasty. From left to right is shown presurgical image, computer-generated prediction, and actual final image. Note the mentalis strain was relieved after the surgery; it is persisted in the predicted image.
(Fig 3). The soft tissue change in the profile photograph seemed to move in a single graphic piece without morphing within it. Therefore, the relief of muscle strain at lip region after backward jaw movement was not accurately expressed in this computer program (Fig 4). Kazandjian et al27 also indicated this phenomenon in their study. The average soft tissue– to– hard tissue ratio was generally used in imaging software programs to generate the predicted images. Such methods only portray the change of the facial outline and do not include the change in individual
341
LU, KO, AND HUANG
soft tissue posture. The soft tissue thickness and muscle tonicity may not be shown precisely in the predicted images. Errors in linear measurement showed that the soft tissue prediction tended to overestimate the distance from lips to the facial plane (glabella to soft tissue pogonion) and those to the E plane (tip of nose to soft tissue pogonion). This would make the predicted lips look more protrusive than the actual ones. The results were similar to those of Upton,24 who presented that the predicted distance from lips to E plane were larger than actual. However, Sinclair et al12 indicated an underestimation of the prediction value from lips to E plane. The measurements of angular prediction tended to be underestimated in facial convexity and nasolabial angle. The result was also similar to those of Upton et al,24 who showed that the actual postsurgical nasolabial angle was larger than the software prediction. In the surgical treatment of bimaxillary protrusion, the frequency with that the software was able to make accurate postsurgical predictions varied. In average, the upper lip region presented 58% of predicted errors less than 2 mm. The lower lip region showed 40% of predicted errors less than 2 mm. These results were similar to the study of Kazandjian et al,27 who used Quick Ceph (Orthodontic Processing, San Diego, CA) and Portrait Planner (Rx Data Inc, Ooltewah, TN) to evaluate the accuracy of video imaging for predicting the soft tissue profile after mandibular setback surgery, with 56% and 56% of predicted errors less than 2 mm in upper lip region and 57% and 48% of predicted errors less than 2 mm in the lower lip region, respectively. On the other hand, Syliangco et al25 used Orthognathic Treatment Planner (Pacific Coast Software, Pacific Palisades, CA) and Prescription Portrait (Rx Data Inc) to predict soft tissue changes in mandibular advancement, showed more accurate predictions in their study, with 98% and 96% of predicted errors less than 2 mm in the upper lip region and 69% and 78% of predicted errors less than 2 mm in the lower lip region, respectively. The various results could be explained that our study and the study of Kazandjian et al27 mainly involve compression of the image and the study of Syliangco et al25 mainly related to expansion of the image. This indicated that different direction of hard tissue movement may influence the accuracy of the prediction programs. In this study, there were no statistically significant differences in the predicted errors between the group with and without genioplasty. This result was similar to the result of Upton et al,24 who used Quick Ceph to predict the postsurgical soft tissue profile; they also did not find any statistically significant differences between the patients with and without genioplasty.
However, in our study, the sample size of the genioplasty group was relatively smaller. The clinical significance needs to be further investigated.
References 1. Kiyak HA, Bell R: Psychosocial considerations in surgery and orthodontics, in Proffit WR, White RP Jr (eds): Surgical Orthodontic Treatment. St Louis, MO, Mosby–Year Book, 1991, pp 71-95 2. Proffit WR, White RP: The need for surgical orthodontic treatment, in Proffit WR, White RP Jr (eds): Surgical Orthodontic Treatment. St Louis, MO, Mosby–Year Book, 1991, pp 2-24 3. Sarver DM, Johnston MW, Maturas VJ: Video-imaging for planning and counseling in ortho-surgery. J Oral Maxillofac Surg 46:939, 1988 4. Sarver DM, Johnston MW: Video-imaging: Techniques for superimposition of cephalometric radiography and profile images. Int J adult Ortho Orthognath Surg 5:241, 1990 5. Sarver DM, Matukas VJ, Weissman SM: Incorporation of plastic surgery in the planning and treatment of orthognathic surgical cases. Int J Adult Orthod Orthognath Surg 6:227, 1991 6. Ackerman JL, Proffit WR: Communication in orthodontic treatment planning: Bioethical and informed consent issues. Angle Orthod 65:253, 1995 7. Turpin D: The need for video imaging. Angle Orthod 65:243, 1995 8. Proffit WR: Treatment planning: The search for wisdom, in Proffit WR, White RP Jr (eds): Surgical Orthodontic Treatment. St Louis, MO, Mosby–Year Book, 1991, pp 142-191 9. Cohen M: Mandibular prognathism. Am J Orthod 51:368, 1965 10. McNeil RW, Proffit WR, White RP: Cephalomertic prediction for orthodontic surgery. Angle Orthod 42:154, 1972 11. Henderson D: The assessment and management of bony deformities of the middle and lower face. Br. J Plast Surg 24:287, 1974 12. Sinclair PM, Kilpelainen P, Phillips C, et al: The accuracy of video imaging in orthognathic surgery. Am J Orthod Dentofac Orthop 107:177, 1995 13. Bhatia SN, Sowray JH: A computer-aided design for orthognathic surgery. Br J Oral Maxillofac Surg 22:237, 1984 14. Harradine NWT, Birnie DJ: Computerized prediction of the results of orthognathic surgery. J Maxillofac Surg 13:245, 1985 15. Walters H, Walters DH: Computerised planning of maxillofacial osteotomies: The program and its clinical application. Br J Oral Maxillofac Surg 24:178, 1986 16. Lew KK: The reliability of computerized soft tissue prediction following bimaxillary anterior subapical osteotomy. Int J Adult Orthod Orthognath Surg 7:97, 1992 17. Kinnebrew MC, Hoffnab DR, Carlton DM: Projecting the softtissue outcome of surgical and orthodontic manipulation of the maxillofacial skeleton. Am J Orthod 84:508, 1983 18. Laney TJ, Kuhn BS: Computer imaging in orthognathic and facial cosmetic surgery. Oral Maxillofac Clin North Am 2:659, 1990 19. Takahashi I, Takahashi T, Mitsuhiko H, et al: Application of video surgery to orthodontic diagnosis. Int J Adult Orthod Orthognath Surg 4:219, 1989 20. Turpin DL: Computers coming on line for diagnosis and treatment planning. Angle Orthod 60:163, 1990 21. Phillips C, Hill BJ, Cannac C: The influence of video imaging on patients’ perceptions and expectations. Angle Orthod 65:263, 1995 22. Hing NR: The accuracy of computer generated prediction tracings. Int J Oral Maxillofac Surg 18:148, 1989 23. Konstiantos KA, O’Reilly MT, Close J: The validity of the prediction of soft tissue profile changes after LeFort I osteotomy using the Dentofacial Planner (computer software). Am J Orthod Dentofac Orthop 105:241, 1994 24. Upton PM: Evaluation of video imaging prediction in combined maxillary and mandibular orthognathic surgery. Am J Orthod Dentofac Orthop 106:451, 1994 25. Syliangco ST, Sameshima GT, Kaminishi RM, et al: Predicting soft tissue changes in mandibular advancement surgery: A
342 comparison of two video imaging systems. Angle Orthod 67: 337, 1997 26. Sameshima GT, Kawakami RK, Kaminishi RM, et al: Predicting soft tissue changes in maxillary impaction surgery: A comparison of two video imaging systems. Angle Orthod 67:347, 1997 27. Kazandjian S, Sameshima GT, Champlin T, et al: Accuracy of video imaging for predicting the soft tissue profile after mandibular set-back surgery. Am J Orthod Dentofacial Orthop 115: 382, 1999
THE ACCURACY OF VIDEO IMAGING PREDICTION 28. Lines PA, Steinhauser WW: Soft tissue changes in relationship of movement of hard structures in orthognathic surgery. J Oral Surg 32:891, 1974 29. Quast DC, Biggerstaff RH, Haley JV: The short and long-term soft-tissue profile changes accompanying mandible advancement surgery. Am J Orthod 84:29, 1983 30. Stella JP, Streater MR, Epker BN, et al: Predictability of upper soft tissue changes with maxillary advancement. J Oral Maxillofac Surg 47:697, 1989