Computer Aided Surgery 3:108 ?114 (1998) Clinical Paper Accuracy of Cephalometric and Video Imaging Program Dentofacial Planner Plus? in Orthognathic Surgical Planning Gu?nter Schultes, M.D., D.D.S., Alexander Gaggl, M.D., D.D.S., and Hans Ka?rcher, M.D., D.D.S., Ph.D. Clinical Department of Oral and Maxillofacial Surgery, University Hospital, Graz, Austria ABSTRACT The prediction of profile changes after surgery poses a problem due to the variability of the soft tissue and the differences in soft-tissue translations relative to osseous changes. This study examined the accuracy of computer predictions of such soft-tissue changes. Twenty-five patients with mandibular retrognathia were examined before and after orthognathic surgery. Changes in soft-tissue reference points were correlated to translations of hard-tissue references in the sagittal and vertical planes, and the measurements from these patients were compared to results predicted in preoperative planning by the cephalometric and video imaging program Dentofacial Planner?. In surgical treatment involving advancement of the mandible, the mean operative advancement of the osseous pogonion was 6.06 mm. The corresponding movements of soft-tissue references, expressed as percentages relative to the movement of the osseous pogonion in the sagittal plane, were 98.4% for the soft-tissue pogonion, 93.6% for the soft-tissue menton, and 49.0% for the soft-tissue labral inferior. The same measurements were carried out in the vertical plane and the changes in soft-tissue references were compared to those predicted in preoperative planning using the Dentofacial Planner?. The predicted images were perceived as agreeing with the actual image most frequently in the lip and nasal area, while the highest degree of error was seen in the submental region. An overall predictability of more than 80% can be attained by planning mandibular advancement operations for correction of mandibular retrognathia using the Dentofacial Planner?. Comp Aid Surg 3:108 ?114 (1998). �98 Wiley-Liss, Inc. Key words: video imaging, orthognathic surgery, accuracy INTRODUCTION The accurate prediction of profile changes following orthognathic surgery is problematic because of the variability of the soft tissue and the differences in soft-tissue translations relative to osseous changes. Numerous studies were published about the predictability of profile changes in relation to osseous translations.2? 4,6 ?13,16,17,22 These studies were based on manually repositioning acetate trac- ings of skeletal segments over the original cephalometric tracing in order to simulate the proposed treatment.13 The posttreatment soft-tissue outline was then added, based on accepted ratios of softtissue to hard-tissue changes. Currently, the prediction technique often involves using a digitized image of a lateral cephalometric tracing in conjunction with a video image of the patient.11,19,20,23,24 Received original December 11, 1997; accepted July 27, 1998. Address correspondence/reprint requests to: Dr. D. Gu?nter Schultes, Clinical Department of Oral and Maxillofacial Surgery, Auenbruggerplatz 7, A-8036 Graz, Austria. �98 Wiley-Liss, Inc. Schultes et al.: Orthognathic Surgical Planning The orthodontic and surgical predictions produced from the digitized cephalometric tracing are combined with the video image so that the prediction includes a line drawing and a corresponding facial image.18,20,21 Images of treatment options can then be displayed on the computer terminal. There are two significant advantages to using video imaging preoperatively. First, the communication between doctor and patient is facilitated by involving the patient in the selection of treatment options; there are fewer surprises, fewer unrealistic expectations, and a greater ?bonding? between the doctor and patient during treatment. The second advantage of video imaging is its ability to aid in treatment planning19,23,24; it provides the orthodontist and surgeon with a manipulable image so that a consensus can be reached on the desired soft-tissue outcome and a decision made as to whether monomaxillary or bimaxillary osteotomies and further corrections should be performed. The main question about the predictability of soft-tissue changes by computer aided planning concerns its accuracy. Because there are so many possible differences in profile, depending on the direction of shift of the mandible and due to postoperative tension with either extension or compression of connective tissue, we asked ourselves whether our computer planning program could give a true prediction of the operative outcome. Therefore, the aim of this study was to compare the actual changes in profile observed following mandibular surgical advancement with the outcome predicted by the Dentofacial Planner? (Dental Facial Software Inc., Toronto, Canada). 109 Fig. 1. Preoperative cephalogram of a patient with mandibular retrognathia. diographs were digitized using the Scriptel RDT1212, a transparent, backlit model that can be used with IBM Software (International Business Machines Corp., Armonk, NY). The manufacturer specifies an accuracy of ?0.025 in. over the entire surface and an accuracy of ?0.01 in. and a resolution of 0.001 in. (1000 lines/in.) over the anterior 90% of the digitized area. A Steiner analysis was produced for each computerized tracing. While taking each lateral skull radiograph, a metal device of 10-mm height and 10-mm breadth was fixed in the glabella abutment. Comparison of its measurements in reality and on the radiograph showed that there was no enlargement or reduction of the real dimensions within a range of error of no more than 0.1 mm. PATIENTS AND METHODS Registration of Operative Changes in Hard- and Soft-Tissue Structures Twenty-five patients with skeletal mandibular retrognathia were examined pre- and postoperatively. The orthodontic treatment of all patients entailed using a fixed appliance technique. The average preoperative therapy lasted 26 months, and treatment following surgery lasted 9 months. All patients were ?18 years of age and received a sagittal splitting of the mandible as described by Obwegeser and Trauner15 and Dal Pont.5 Fifty lateral cephalograms, 25 preoperative and 25 postoperative, were taken in centric occlusion with the patients? heads fixed in a natural position by a Siemens Cephalostat (Siemens Medical Systems Inc., Iselin, NJ) and with the lips relaxed. The cathode to object distance was 5?, and the object to film distance was 13 cm. Each radiograph was duplicated, and 35 landmarks were permanently marked with pin holes. The marked ra- Lateral cephalograms were performed preoperatively and 12 weeks postoperatively. Both cephalograms were digitized with the Prescription Dentofacial Planner? software program25 (Fig. 1) using the Scriptel RDT-1212. To show the real operative translations and the relationship of softtissue changes to osseous changes pre- and postoperative cephalograms were superimposed and changes in hard- and soft-tissue references in the sagittal and vertical planes were recorded. The sagittal translations were registered parallel to the NSe line (nasion-sella-entrance line), and the vertical translations were registered perpendicular to this. Initially, the sagittal and vertical translations were registered in the region of hard-tissue structures such as the pogonion (oss. po), menton (oss. men), B point (B pt.), incisal point in the maxilla and mandible (inc. max. and inc. mand.), and the spina 110 Schultes et al.: Orthognathic Surgical Planning Fig. 2. Preoperative lateral video image of the same patient with mandibular retrognathia. nasalis anterior. Similarly, the translations in relation to soft-tissue references, such as the pronasal (pronas.), subnasal (subnas.), stomion (sto), superior labral (lab. sup.), inferior labral (lab. inf.), soft-tissue pogonion (po. soft), and soft-tissue menton (men. soft) were recorded. Changes of softtissue reference points were correlated to translations of three hard-tissue references; the oss. po, oss. men., and inc. mand. in the sagittal and vertical plane. The percentage relationships of relative position alterations were calculated and standard deviations determined. Thus, the basic translations of hard- and soft-tissue structures could be demonstrated in all patients. Registration of Differences in Computer Aided Planning and Real Postoperative Profile On the basis of preoperative cephalograms and profile videos, preoperative video imaging (Figs. 2, 3) was performed in order to plan the surgical treatment by simulating the posttreatment profile situation. This simulated cephalogram was superimposed on the postsurgical one, and the hardtissue cephalometric prediction was superimposed on the digitized posttreatment hard-tissue cephalogram (Fig. 4). Thus, with the computerized hardtissue prediction and the actual hard-tissue final result superimposed (Figs. 5, 6), it was possible to compare and analyze the line drawings of cephalometric soft-tissue outlines in order to determine the accuracy of the soft-tissue line-drawing prediction (Fig. 4). Furthermore, it was possible to compare the preoperative, planned, and postoperative profile by video imaging (Fig. 7). Fourteen linear soft-tissue measurements (Fig. 4) were obtained from the digitized posttreat- Fig. 3. Superimposition of the cephalogram to the lateral video imaging preoperatively. Hard structures are demonstrated in correlation to soft-tissue structures. ment cephalogram and from the prediction line drawing. The differences between the actual posttreatment soft tissue and the predicted posttreatment soft-tissue line drawing were analyzed in the sagittal and vertical planes in order to show the differences between the computer predictions and the actual outcomes. Soft-Tissue Prediction Algorithms of Dentofacial Planner Plus Dentofacial Planner Plus (DFP Plus) is a software application that allows clinicians to link digitized lateral cephalograms to lateral facial images and visualize facial soft-tissue responses to orthodontic treatment and orthognathic surgery. DFP Plus simulates soft-tissue changes using a proprietary series of algorithms developed by Dentofacial Software Inc. and based on retrospective studies of predictive reliability. Earlier versions of DFP Plus used a Fig. 4. Preoperative computer aided planning of mandibular advancement. The picture shows the predicted changes of soft-tissue and hard-tissue structures. Schultes et al.: Orthognathic Surgical Planning 111 Fig. 5. Superimposition of the cephalogram to the lateral video imaging postoperatively. New positions of hard-tissue structures in correlation to soft-tissue structures are demonstrated. predictive model wherein ratios or linear regression equations were used to predict the location of softtissue landmarks relative to hard-tissue change. The current version of DFP Plus employs a nonlinear predictive approach, using a pattern-recognition scheme to categorize the class of facial deformity and then establish movement thresholds prior to which little or no soft-tissue change may be elicited. This nonlinear approach is used for predicting soft-tissue response to maxillary surgery, mandibular rotational changes, mandibular advancement and setbacks, and genioplasties. Accuracy of Cephalometric Radiographs and Measurements While taking each lateral skull radiograph, a metal device of 10-mm height and 10-mm breadth was fixed in the glabella abutment. Comparison of the measurements in reality and on the radiograph Fig. 7. The hard- and soft-tissue cephalometric primary landmarks and the horizontal and vertical reference lines. N, soft tissue nasion; Ns, pronasal; Snu, subnasal upper; Sn, subnasal; Ls, labral superior; Stos, stomion superior; Stoi, stomion inferior; Li, labral inferior; sBu, soft-tissue B upper; sB, soft-tissue B; sBl, soft-tissue B lower; sPg, softtissue pogonion; sGn, soft-tissue gnathion; sMe, soft-tissue menton. showed that there was no enlargement or reduction of the real dimensions within a range of error of no more than 0.1 mm. The main error was that resulting from the reproduction of cephalometric landmarks. After marking the same cephalogram 25 times a mean error of 0.4 mm was recorded in the menton, but in all other points the error was low. The precision of the digitizer itself is high with a resolution of 1000 lines/in. RESULTS Registration of Operative Changes in Hard- and Soft-Tissue Structures Fig. 6. Video images of the preoperative, planned, and postoperative profile. The mean ventral displacement for all patients was 6.06 mm for the osseous pogonion, 6.19 mm for the osseous menton, and 5.05 mm for the lower incisal point. In the vertical plane, the caudal dislocation of the oss. po. was 3.48 mm, that of the oss. men. was 2.96 mm, and that of the inc. mand. was 2.90 mm (Table 1). For confirmation purposes, the accuracy of the superimposition translation of the anterior nasal spine and the inc. max. was determined and found to be congruent up to two decimal places. Following translation of the osseous pogo- 112 Schultes et al.: Orthognathic Surgical Planning Table 1. Direction and Extent of Movement of Osseous and Soft Tissue Structures with Mandibular Advancement Ventral Osseous transposition (mm) Spina nasalis anterior Incisal point, maxilla Osseous pogonion Osseous menton Incisal point, mandible Caudal 0 0 6.06 6.19 5.05 Ventral Soft-tissue transposition (mm) Pronasal Subnasal Stomion Labral superior Labral inferior Pogonion (soft) Menton (soft) 0 0 2.73 1.78 2.96 6.15 5.65 Table 3. Mean Differences between Predicted and Actual Line Drawing SoftTissue Values 0 0 3.48 2.96 2.90 Caudal Cranial 0 0 0.9 1.15 1.46 2.71 3.13 N Ns Snu Sn Ls Stos Stoi Li sBu sB sBl sPg sGn sMe Sagittal (mm) SD Vertical (mm) SD 0 0.16 0.38 0.11 0.16 0.16 0.16 0.66 1.00 1.16 1.22 1.44 1.55 1.55 0 0 0 0 0.39 0.49 0.58 0.83 1.06 1.09 1.22 1.10 1.04 1.07 0 0 0.44 0.27 0 0 0.11 0.11 0.27 0.33 0.38 0.166 0.72 1.00 0 0 0 0 0.51 0.70 0.88 0.93 1.43 1.15 1.34 1.24 1.30 1.38 N, soft tissue nasion; Ns, pronasal; Snu, subnasal upper; Sn, subnasal; Ls, labral superior; Stos, stomion superior; Stoi, stomion superior; Li, labral inferior; sBu, soft-tissue B upper; sB, soft-tissue B; sBl, soft-tissue B lower; sPg, soft-tissue pogonion; sGn, soft-tissue gnathion; sMe, soft-tissue Menton. nion, the relative soft-tissue changes in the sagittal plane were 98.4% for the soft-tissue pogonion, 93.6% for the soft-tissue menton, 29.6% for the labral superior, 45.2% for the stomion, and 49.0% for the labral inferior (Table 2). The relationships of these soft-tissue references to the osseous menton (oss. men.) and lower incisal point (inc. mand.) in the sagittal plane are also recorded in Table 2. Upon correction of mandibular retrognathia, the percentage change for the soft-tissue pogonion relative to the osseous pogonion in the vertical plane was 78%. The relative percentage changes for other soft-tissue references were 90% for the soft tissue menton, 51% for the labral superior, 26% for the stomion, and 42% for the labral inferior (Table 2). Further percentage changes for soft- tissue references relative to the osseous menton and lower incisal point are also recorded in Table 2. Registration of Differences in Computer Aided Planning and Real Postoperative Profile The mean difference between actual postsurgical and predicted line drawing soft-tissue values is recorded in Table 3. The predicted images were perceived as agreeing with the actual image most frequently in the upper lip and nasal area, while the labiomental fold and mental areas showed the poorest correlation with a mean error of 1.5 mm. Overall lower differences but higher standard deviation Table 2. Percentage Distribution of Soft-Tissue Versus Osseous Transposition after Mandibular Advancement Changes in sagittal relationship po. (soft)/oss. po. sto./oss. po. lab. sup./oss. po. lab. inf./oss. po. men. (soft)/oss. po. Changes in vertical Relationship lab. sup./oss. po. sto./oss. po. lab. inf./oss. po. po. (soft)/oss. po. men. (soft)/oss. po. Percent s 98.4 45.2 29.6 49.0 93.6 7.3 6.4 6.5 6.9 7.2 11.3 12.0 11.1 8.9 12.3 51 26 42 78 90 Percent s Percent s men. (soft)/oss. men. po. (soft)/oss. men. lab. inf./oss. men. sto./oss. men. lab. sup./oss. men. 91.3 99.3 47.8 44.0 28.8 6.4 6.7 7.0 5.7 6.3 lab. sup./inc. mand. sto./inc. mand. lab. inf./inc. mand. po. (soft)/inc. mand. men. (soft)/inc. mand. 35 54 58 82 89 8.1 7.3 7.5 7.4 7.7 lab. sup./oss. men. sto./oss. men. lab. inf./oss. men. po. (soft)/oss. men. men. (soft)/oss. men. 39 30 49 92 94 12.4 11.4 12.1 12.4 14.3 lab. sup./inc. mand. sto./inc. mand. lab. inf./inc. mand. po. (soft)/inc. mand. men. (soft)/inc. mand. 40 31 50 93 93 13.3 12.5 11.9 10.8 14.3 Schultes et al.: Orthognathic Surgical Planning were noted in the vertical plane. In contrast to this, there were higher values but lower standard deviation in the sagittal plane. Statistics The paired t test was employed to determine the significance of the stated relationships. The determined values were plotted according to the probability net. For every average value, the standard deviation (s) was determined as shown in Tables 2 and 3. By means of the paired t test, the concordance of the calculated averages and standard deviations was compared to the actual postoperative average. A change in the sagittal relationship was found, revealing a probability of error of p ? 0.15. For changes in the vertical relationship, a slightly lower probability of error of p ? 0.1 was determined. DISCUSSION It is still problematic to give a prognosis of the postoperative profile of patients undergoing orthognathic surgery that is due to the differences in soft-tissue translations relative to osseous translocations. The assumption of a 1:1 relationship of osseous to soft-tissue references is actually only valid near the pogonion. Several investigators2? 4,8,12,14,16 reported this when examining the chin regions of patients following mandibular advancement. In our study a relationship of 0.98:1.0 was determined upon comparing soft- and hard-tissue pogonion in patients after mandibular advancement. The main error of the computer prediction was about 20% in the sagittal plane. Thus, the Dentofacial Planner? showed a larger and more protrusive prediction than the actual value in the chin area. The more caudal the reference point, the larger and more protrusive the value became. This was contrary to the results of Sinclair et al.21 who demonstrated the greatest probability of errors in the upper lip region using the video imaging program Prescription Planner/Portrait software (Rx Data Inc., Ooltewah, TN). In our study the cranial reference points showed a low margin of error when using the DFP Plus. The midface and upper lip region showed an error of less than 0.2 mm and the lower lip region an error of 0.6 mm, demonstrating a high significance of prediction for the lip area. Lines and Steinha?user12 found a ratio of 0.67: 1.0 and Mommerts and Marxer14 reported a ratio of 0.55:1.0 for the lower lip after operative mandibular advancement, but our study showed an even smaller ratio for similar patients. We recognized a 113 lesser degree of sagittal movement in the lip area and therefore a higher predictability of soft-tissue changes compared to the chin region. This difference was confirmed by the relation of the lower incisal point and the labral inferior. The absolute values of the changes in stomion relation were found to lie between the average values of the superior and inferior labral in the sagittal plane. In the vertical plane different percentages were found for the relationship of soft-tissue to hard-tissue references: with a ratio of 0.92:1.0 there was a strong relative dependency of upper lip position on the vertical position of the osseous pogonion, the lower lip following the osseous pogonion in a ratio of 1:1 after operative mandibular advancement. Here again the prediction in the pogonion area was very accurate. Furthermore, there was a lower level of mean error in the vertical plane than in the sagittal plane, comparing all reference points. Only in the submental area could errors greater than 0.5 mm be recorded. This may be explained by the thickness of the soft-tissue structures that were being stretched in this area, making prediction more difficult. Significant changes in lip contour were found postoperatively, as well as a new configuration of the chin. For operative mandibular advancement, the soft tissue structures of the pogonion and menton followed the corresponding osseous references in ratios of less than 1:1 (0.78:1.0 and 0.94:1.0, respectively). In video imaging contour changes in chin area in comparison to preoperative chin contour were not as severe as in postoperative reality. The reason for this may be the relatively high probability of prediction errors in the submental area, which can be explained by the high portion of fat tissue in this region. Therefore, it seems to be difficult to predict movements in fat tissue following mandibular translocation but muscular displacement should be easier to predict; this was demonstrated in the lip region where a closer correlation of video imaging and postoperative situation was recorded in the lip contour. In contrast to our experiences with the DFP Plus, Sinclair et al.21 reported better predictability for the chin region using another planning program produced by Rx Data Inc., although they did not discuss submental changes. It can be concluded that mandibular advancement in correction of mandibular dysgnathia results in significant changes in the lower face, which can be predicted in the vertical and horizontal plane with only a small degree of error, provided that transduction of cephalograms has been performed 114 Schultes et al.: Orthognathic Surgical Planning accurately, as shown by Baskin and Cisneros.1 This statement is valid for the DFP Plus and the Quick Ceph program (Orthodontic Processing, Coronado), according to Baskin and Cisneros,1 because the reliability and reproducibility of landmarks is similar in both programs. Overall, a predictability of more than 80% can be stated. The highest degree of error was determined to occur in the submental area, and the error grows as the extent of mandibular advancement increases. The less the soft tissue structures follow this advancement, the smaller the error. Exceptions to this statement in bimaxillary and further complex osteotomies should be examined in the future. For an accurate prediction of all changes in the face, a 3-dimensional examination must be performed, so differences between our results and those reported elsewhere in the literature could have arisen because a 2-dimensional examination cannot explore a 3-dimensional movement and its consequences. 11. 12. 13. 14. 15. 16. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Baskin HN, Cisneros GJ (1997) A comparison of two computer cephalometric programs. J Clin Orthod 4:231?233. Bell WH, Dann JJ (1973) Correction of dentofacial deformities by surgery in the anterior part of the jaw: a study of stability and soft tissue changes. Am J Orthod 64:162?166. Bell WH, Jacobs JD (1980) Surgical orthodontic correction of maxillary retrusion by Le Fort I osteotomy. J Maxillofac Surg 8:84 ? 88. Bell WH, Proffit WR, White RP (1980) Surgical Correction of Dentofacial Deformities. Philadelphia, PA: W. B. Saunders & Co. Dal Pont G (1961) Retromolar osteotomy for the correction of prognathism. J Oral Surg 19:42? 46. Dann JJ, Fonseca RJ, Bell WH (1976) Soft tissue changes associated with total maxillary advancement: A preliminary study. J Oral Surg 34:19 ?24. Epker BN, Fish LC (1986) Dentofacial Deformities, Vols. I and II. St. Louis: Mosby. Ewing M, Ross RB (1992) Soft tissue response to mandibular advancement and genioplasty. Am J Orthod 6:550 ?555. Freihofer HB (1977) Changes and nasal profile after maxillary advancement in cleft and non-cleft patients. J Maxillofac Surg 5:20 ?26. Hernandez?Orsini R, Jacobson A, Sarver DM et al. (1989) Short term and long term tissue profile changes 17. 18. 19. 20. 21. 22. 23. after mandibular advancements using rigid fixation techniques. Int J Adult Orthod Orthognath Surg 4:209 ?214. Laney TJ, Kuhn BS (1990) Computer imaging in orthognathic and facial cosmetic surgery. Oral Maxillofac Clinics N Am 2:659 ? 668. Lines PA, Steinha?user EW (1974) Soft tissue changes in relationship to movement of hard structures in orthognathic surgery: A preliminary report. J Oral Surg 32:891? 897. McNeil RW, Proffit WR, White RP (1972) Cephalometric prediction for orthodontic surgery. Angle Orthod 42:154 ?164. Mommaerts MY, Marxer H. (1987) A cephalometric analysis of the long term soft tissue profile changes with accompany the advancement of mandible by sagittal split ramus osteotomies. J Craniomaxillofac Surg 15:127?134. Obwegeser H, Trauner R (1955) Zur operationstechnik bei der progenie und anderen unterkieferanomalien. Dtsch Zahn Mund Kieferheilkd 23:11?17. Quast DC, Biggerstaff RH, Haley JV (1983) Short term and long term soft tissue profile changes accompanying mandibular advancement surgery. Am J Orthod 84:29 ?37. Radney LJ, Jacobs JD (1981) Soft tissue changes associated with surgical total maxillary intrusion. Am J Orthod 80:191?199. Sarver DM (1993) Video imaging?A computer facilitated approach to communication and planning in orthognathic surgery. Br J Orthod 20:187?191. Sarver DM, Johnston MW (1990) Video imaging: Techniques for superimposition of cephalometric radiography and profile images. Int J Adult Orthod Orthognath Surg 5:241?248. Sarver DM, Johnston MW, Matukas VJ (1988) Video imaging for planning and counseling in orthognahtic surgery. J Oral Maxillofac Surg 46:939 ?945. Sinclair PM, Kiplelainen P, Phillips C, White RP, Rogers L, Sarver DM (1995) The accuracy of video imaging in orthognathic surgery. Am J Orthod Dentofac Orthop 107:177?185. Schmuth G. (1990) Kieferorthopa?die I. Munich: Urban & Schwarzenberg. Takahashi I, Takahashi T, Mitsuhiko H et al. (1989) Application of video surgery for orthodontic diagnosis. Int J Adult Orthod Orthognath Surg 4:219 ?222. Turpin DL (1990) Computers coming on line for diagnosis and treatment planning [Editorial]. Angle Orthod 60:163? 164. Walker RP (1990) Dentofacial Planner User Manual. Toronto: Dentofacial Software, Inc.