Three-dimensional modeling system for unilateral mandibular bone distraction: A clinical case

Abstract

Facial hemiatrophies are anomalies of the first branchial arch and affect one in 4000–5000 newborns. Bone distraction is the technique of choice for the treatment of these dysmorphoses. Mandibular osteodistraction requires prior determination of the characteristics of the distraction vector whose three components will serve to activate the distractor.

The patient, aged 5 years, presented with a right facial hemiatrophy, Grade IB according to the classification of Pruzansky. Tomodensitometric acquisition was obtained with a CT scanner. Software specifically designed for this application allows segmentation of the anatomical elements by a region-growing algorithm. The 3D representation of each element is added to a 3D scene, in which are placed the built-up landmarks necessary for the surgical simulation after 3D cephalometric analysis. The surgical cleavage plane is oriented according to the surgeon's requirements while preserving the predominant anatomical elements. The software allows performance of rotations and translations of the bone segments rendered independently from the cleavage plane. The distances and angles covered during the virtual movement are measured at its conclusion. The aim of moving the bone segments is to render the mandibular occlusion plane parallel to the reference occlusion plane. The vertical growth of the maxilla is realized by secondary recuperation. The distractor used was of an external multidirectional type allowing elongation of the mandibular ramus and mandibular corpus, closure of the goniac angle, and lateralization or medialization of the ramus. On the 15th day, the mandibular angle was reduced by 10°, which allowed closure of the anterior gap and recentering of the incisive areas by a half-cuspid.

The patient presented with a complex bone deficit in the three spatial directions, which allowed the development of software for modeling the distraction. Other clinical cases will be necessary to validate this 3D imaging-based technique.

• Distraction osteogenesis
• biomodelization
• facial hemiatrophy
• tomodensitometric acquisition
• three-dimensional cephalometry

Introduction

Facial hemiatrophies are anomalies of the first branchial arch and affect one in 4000–5000 newborns. They include mandibular, auricular and tegmental deformations and represent many clinical forms listed in the OMENS classification of Vento et al. [1]. Mandibular bone deficit can be treated by interposition of bone substitutes (hydroxyapatite or coral), by interposition of costal bone grafts, or by iliac graft. The graft-based approach has a number of drawbacks: the graft prognosis is uncertain, it requires a second operative sampling site, sequelae are sometimes complicated, and the resulting cutaneous cicatrization is often unsightly.

Bone distraction is the technique of choice for the treatment of these dysmorphoses. MacCarthy [2] described a distraction method adapted to the hypoplastic mandible in the child. Unlike interposition techniques which give a rigid result, mandibular osteodistraction allows a progressive morphological correction, has a low incidence of complications, is relatively harmless to the temporo-mandibular joints, and results in spontaneous acceleration in the growth of the superior maxilla on the side of the mandibular distraction. The biomechanical quality of the regenerated bone at one year after distraction is similar to that of normal bone tissue [3–5].

Mandibular osteodistraction does, however, require prior determination of the characteristics of the distraction vector whose three components will serve to activate the distractor. Distraction effects must be considered in three dimensions to evaluate the bone morphology after treatment. At present, there are two means of simulating the distraction preoperatively: stereolithography and digital 3D biomodelization. Stereolithography creates a polyurethane model of the patient anatomy on which the surgery can be simulated beforehand [6]. Digital 3D models are obtained after processing data from preoperative scans. Many processing softwares are now available; however, some only show a 3D image but do not allow any modification or movement within the 3D scene [7], while others need to work in two dimensions and only translate the simulation to three dimensions afterwards. The software we use allows each bone segment to be manipulated (via cuts, rotations, translations, etc.) as desired directly in the 3D scene, enabling all the steps to be viewed before obtaining the final result. We also simulate the surgical stage of the process.

The aim of the current study was to obtain a virtual 3D simulation of the effect of a unilateral mandibular bone distraction in a young girl who presented with a facial hemiatrophy.

Materials and methods

Patient

The 5-year-old patient presented with a right facial hemiatrophy, Grade IB according to the classification of Pruzansky [8]. Clinical examination revealed a diminution (15 mm) in facial height on the atrophic side, a tipping of the occlusion plane and bicommissural line on the right side, a deviation (2 mm) of the cutaneous chin point towards the right relative to the median sagittal plane, and a malformation of the ear pavilion corresponding to type II microconcha Figures 1 and 2. Institutional review board approval and parental informed consent were obtained.

Figure 1. Five-year-old patient with a right facial hemiatrophy, Grade IB (classification of Pruzansky): (a) anterior facial view; (b) left profile; (c) right profile. [Color version available online.]

Figure 2. Intraoral views: (a) anterior view; (b) left lateral view; (c) right lateral view. [Color version available online.]

Scan acquisition

The preoperative scan was performed with the mouth closed, using a Somaton Sensation 4 CT scanner (Siemens, Erlangen, Germany) which allows spiral acquisition. Acquisition parameters were as follows: 4 × 1 mm, with a displacement of 2.7 mm per rotation and an incremental reconstruction of 0.8 mm at a tension of 120 kV and an intensity of 70 mA/s, giving an effective dose of 0.3 mSv. Voxel size was 0.36 mm (X) × 0.36 mm (Y) × 0.63 mm (Z). The digital records of the sections were transferred to the image processing laboratory.

Image processing software

Software specifically designed for this application allows segmentation of the anatomical elements by a region-growing algorithm [9]. The following elements were segmented: the apophysis crista galli, the spina nasalis anterior maxillae, the right and left sub- and supra-orbital foramina, the right and left semi-circular canals, the mandibular condyles, the mandible, the maxilla, the mandibular teeth, and the maxillary teeth. The 3D representation of each element is added to a 3D scene, in which are placed the built-up landmarks necessary for the surgical simulation (Figure 3).

Figure 3. Three-dimensional cephalometric analysis as proposed for determining the distraction vector. [Color version available online.]

Landmarks

Concerning the anatomical elements, we defined points allowing the construction of planes from which the measurements and movements of bone segments are performed. These anatomical elements were not affected by the pathology. The five basic planes are as follows:

• A median sagittal plane passing through the middle of a line joining the right and left semi-circular canals, the summit of the apophysis crista galli and the summit of the spina nasalis anterior maxillae.

• A frontal plane passing through the right and left sub-orbital foramina and the healthy supra-orbital foramen.

• A horizontal plane passing through the intersection of the right and left semi-circular canals and forward via the healthy sub-orbital foramen (Figure 3).

• An occlusion plane parallel to Camper's plane (modified) passing through the right and left semi-circular canals (equivalents of the external acoustic meatus) and the spina nasalis anterior maxillae.

• A surgical cleavage plane.

Surgical simulation

The cleavage plane is oriented according to the surgeon's requirements while preserving the predominant anatomical elements (dental nerve, dental germs). The software allows performance of rotations and translations of the bone components rendered independently from the cleavage plane. The distances and angles covered during the virtual movement are measured at its conclusion.

Results

Construction of planes

Plane construction was performed by the software from the points placed by the operator on the previously segmented anatomical elements.

The surgical cleavage plane was defined by three points: the apex of the coronoid process, the apex point of the pre-angular incisure of the mandible, and the projection of the posterior aspect of the germ of 47 on the internal cortex of the mandible. This cleavage plane allowed individualization of a posterior mandibular segment (posterior indent) which corresponded to the rising branch on the atrophic side, and an anterior mandibular segment (anterior indent) which corresponded to the mandibular corpus and the rising branch on the healthy side (Figure 4).

Figure 4. The surgical cleavage plane, positioned behind the germ of 47, and passing by the pre-angular incisure of the mandible. [Color version available online.]

The occlusion plane was constructed parallel to the modified Camper's plane passing through the summit point of the medio-vestibular cuspid of 65, which served as the occlusal reference (healthy side). The summit point of the medio-vestibular cuspid of 55 (atrophic side) at a distance from the occlusion plane thus defined was projected perpendicularly on the occlusion plane.

Movement of bone segments

Bone segment movements were performed around the surgical cleavage plane of the mandible Figures 5 5 and 6.

Figure 5. Initial bone position. [Color version available online.]

Figure 6. Result at the end of distraction. [Color version available online.]

A rotation of the anterior segment was performed around the center of rotation situated at the summit of the condyle (healthy side) and along the axis of rotation extending from the healthy condyle to the summit of the pre-angular incisure (pathological side). The summit point of 55 was firmly attached in the 3D scene to the mandibular teeth and the anterior indent. The whole segment was displaced downwards until the summit point of 55 was on the occlusion plane. This provided the vertical component of the distraction to be performed by horizontalization of the occlusion plane. The lowering obtained by this movement was 22 mm.

This vertical displacement was evaluated on the median sagittal plane by measuring the distance separating the projection of the initial inter-incisive point and the projection of the inter-incisive point after distraction (= 22 mm). The sagittal displacement was evaluated by the distance separating these two points projected on the median sagittal plane (= 2 mm). The transverse displacement was evaluated by the distance separating these two points projected on the frontal plane (= 2 mm). Thus, we obtained the three components of the distraction vector.

Rotation of the posterior segment was performed around a center of rotation placed at the summit of the condyle (atrophic side) and along an axis of rotation joining the summit of the two condyles. The rotation gave rise to a displacement (opening) of 6 mm.

Surgical stage

The distractor used in this procedure was of an external multidirectional type, allowing elongation of the mandibular ramus and mandibular corpus, closure of the goniac angle, and lateralization or medialization of the ramus. We consider MacCarthy's distractor to be best suited for multidirectional correction in young children (4–5 years old) who are unable to benefit from the floating callus technique (Figure 7).

Figure 7. Three-dimensional distractor in place after surgery. [Color version available online.]

The mandibular angle was approached by the external vestibular route after sub-periosteal dissection had been performed. Subsequently, bi-cortical osteotomy was carried out under visual control according to the pre-established plot on the virtual model.

Four transjugal pins were set in place on either side of the osteotomy line. The two anterior pins were fixed on the basilar border forward of the pre-angular incisure, and the two posterior pins were fixed at the level of the posterior border of the ramus.

There was a distance of 15 mm between each pair of pins and of 4 mm between each twin pin.

One week after placing the distractor, collectasis was begun at a rhythm of 0.5 mm morning and evening by activating the postero-superior pins, ensuring the growth of the ramus. On the 15th day, the mandibular angle was reduced by 10°, which allowed closure of the anterior gap and recentering of the incisive areas by a half-cuspid. The distractor was removed after 10 weeks’ consolidation.

Discussion

The patient presented with a right mandibular atrophy which led to a homolateral growth defect of the maxilla with an anterior gap and a tipping of the occlusion plane upwards on the atrophic side in the frontal plane (Figure 8). Therefore, the occlusion plane could not be defined by dental landmarks but by a plane parallel to Camper's plane (reference from the total prosthesis) defined by the bone landmarks. The occlusion plane thus defined was the reference position of the occlusion plane (Figure 9). However, on the pathological side, the maxillary teeth were above this plane (vertical growth defect of the maxilla) which cut through the mandibular teeth.

Figure 8. Pathologic occlusal plane. [Color version available online.]

Figure 9. Modified occlusal plane chosen in reference to Camper's plane. [Color version available online.]

The choice of orientation of the cleavage plane had to satisfy several requirements:

• It had to pass behind the germ of 47 (a tooth necessary for the placing of a stable and functional occlusion, thereby giving it a very posterior position).

• It then had to pass by the pre-angular incisure. As this was very pronounced on the pathological side, it was desirable during the movement to open this angle and confer a more harmonious form on the basilar border of the mandible.

• However, the coronoid process is surgically difficult to reach and is the support of muscular insertions. This is why, in practice, the cleavage plane is angled forwards and joins up with the anterior edge of the rising branch beneath the coronoid process.

Therefore, the aim of the movement of the bone segments is to render the mandibular occlusion plane parallel to the reference occlusion plane. The vertical growth of the maxilla is realized by secondary recuperation. This “lowering” of the teeth on the pathological side gives us the vertical component of the distraction vector.

The choice of the vestibular orientation is not an empirical one. Indeed, the external semi-circular canals present a physiological individuality in relation to the rest of the head: their orientation responds only to the action of gravity. Because of this, their spatial location is practically the same for all gnathostomas. Furthermore, it has been shown that growth has little impact on these canals: the labyrinth has almost reached its definitive size as of the 20th week of gestation and only minor modifications occur during development. It undergoes only a very slight rotation during cranio-cephalic expansion, but this takes place around a local fixed point, i.e., the external semi-circular canal. Thus, the sagittal radiological aspect of the external canals is identical at all stages of growth of the cephalic extremity [10], [11]. Fenart [12] reported that, in spite of the large cranio-facial deformations, the external semi-circular canals preserve the same orientation (except in very rare cases of major hydrocephalus).

Evaluation of the symmetry of the hemi-mandibles of the healthy and pathological sides relative to the median sagittal plane can be performed by the software. This geometrical operation allows one to assess whether the bone gain obtained by distraction results in a mandibular morphology closely resembling that of the healthy side. In the present clinical case, the gonion created by distraction remained at a distance from its ideal position as indicated by the symmetry evaluation. This nevertheless enabled recreation of a sufficient jugal volume and improvement in the aesthetic appearance.

Other groups [13–19] have also worked on preoperative 3D visualization but have not taken into account a 3D cephalometric landmark. Segmentation of the whole skull volume is realized in a single step by thresholding without separation of the different anatomical components. Only a surface representation of the skull is obtained, which does not enable the placement of anatomical points on precise bone landmarks necessary for the construction of a cephalometric landmark. Thus, everything is based on a median sagittal plane which serves as a symmetry plane from which an ideal mandible is visualized. The distraction vector should enable one, at the end of distraction, to attain this symmetrical mandibular morphology. The problem is therefore treated without any consideration of occlusion, which we consider to be a major issue in such young children. Indeed, dental occlusion stabilizes the result of the distraction while preserving the bone capital acquired during movements of mastication and bruxism.

Treil et al. [20] have worked on 3D cephalometry, but the landmarks they create do not have enough mandibular components to be useful in the treatment of oto-mandibular syndromes. In particular, no posterior mandibular landmark can be used in the zone where the distraction is performed.

Most groups working in this area only obtain representations of the bone surfaces, but have been able to model the distractor. We model only the movements of the bone segments and do not yet represent the distractor in the distraction zone during movements. However, as the maxilla and mandible are segmented separately, they can be displaced by rotation and translation by reproducing the movements created by distraction. The results obtained are therefore quite coherent. Furthermore, surface representations do not allow performance of precise measurements between two points in the 3D scene. This step is essential for simulating quantitatively the distraction to be realized and for transferring the data into the operative field.

Practical application

The patient in this study presented with a complex bone deficit in the three spatial directions, which allowed the development of a software program for modeling distraction. Figure 10 shows the patient's facial appearance before and after the distraction procedure was performed. Other clinical cases will be necessary to validate this 3D technique.

Figure 10. Patient's face before (a) and after (b) distraction. [Color version available online.]

The standard MacCarthy mandibular distractor is not adapted to treat every case of complex mandibular deformities because of the many multidirectional movements required. Individually designed patient-specific distractors would be ideal for application of this imaging-based technique.