1 Introduction

Temporomandibularjoint(TMJ)ankylosis involves fusion ofthe disc-condyle with the temporal articular surface and restricts mandibular movements which can cause speech impairment,difficulty with mastication,airway compromise,limited mouth opening of the mandible and facial asymmetry or psychological stress,therefore cause speech impairment,difficulty with mastication and airway compromise[1-5].TMJ ankylosis is commonly associated with trauma,localand/orsystemic infections,autoimmune diseases or failed surgery.

According to Sawhney's classification[6],the ankylosis was classified into four different types:type I,extensive fibrous adhesions around the joint and minimal bony fusion;type II,at the outer edge of the joint surface was more bony fusion,but no fusion within the more medial area of the joint;type III,between the mandible and the temporal bone was a bridge;and type IV,a majority of bone replaced the joint. Radiographic examination included panoramic radiographs and computed axial and coronal tomography(CT)to determine the extent of the ankylosis and to rule out any other cause of limited mouth opening.

Predominanttreatmentmethods forTMJ ankylosis include interpositional arthroplasty,gap arthroplasty and reconstruction of the joint usingautogenous or alloplastic materials.However,there are some of the disadvantages as follows:(1)creation of pseudoarthrosis and short ramus and an increased risk of extensive re-ankylosis.(2)fibrosis and scar contracture of the temporalis muscle which may cause trismus,some clinicians suggested to perform coronoidectomy to prevent this complication.(3)depression in temporal region and chronicheadache.(4)graftfracture,infection,overgrowth,donor site morbidity ankylotic bone resection remains the most important aspect of the surgicalprotocolto avoid relapse[7].Resecting massive abnormal formation bone at the skull base is difficult especially with distorted anatomy and the high risks of dura mater exposure,leakage of cerebrospinal fluid,nerve and artery injuries[8].

Rehabilitating the mandibular movement and function,preventing complications and promoting mandibular growth are the main treatment objectives.However,compared to normal anatomic structures,the irregular ankylotic bone has no clear margin in patients with TMJ ankylosis.It is difficult to resect massive bony tissue at the skull base including vital anatomic structures,especially with variation.There is a high risk of hemorrhage resulting from injury of the internal maxillofacial artery,nervesand dura materexposure with cerebrospinal fluid leak[8].

3 D reconstruction for surgery is useful in the areas of complex anatomy where precision is required. This study focuses on the 3D reconstruction of the TMJ from CT scan and mimics to final finite element analysis(FEA)working model.Numerical modeling gives the surgeon an insight into the workings structure of the mandible before and after plate fixation and allows for detailed analysis and planning before surgery.Although computer aided design/computer aided manufacturing(CAD/CAM)has been extensively used in many aspects ofmedicalfield,its application in the orthopedic repair of multiple and complex TMJ ankylosis need to be investigated further.To evaluate the clinical effect of different virtual surgical methods on arthroplasty in the treatment of TMJ ankylosis,we mimic the surgical procedure and predict the surgical outcome on virtual TMJ models with the aid of CAD/CAM,and also finally to investigate their applicable values through the comparison of condylar process stress distribution among TMJ ankylosis mandible models under two different virtual surgical procedures and normal mandible model of TMJ ankylosis or normal condition.

2 Materials and methods

2.1 3D reconstruction

Preoperative planning and simulation

A GE light speed 16-detector spiral CT scanner was used for thin-layer scanning in one patient with TMJ ankylosis from the top of the cranium to maxillary level in a thickness of 0.625 mm;in total,307 layers of transverse images were input to a 3D imaging reconstruction workstation and saved as DICOM file format.Data were converted via a compact disc(CD)to a Windows-based computer workstation with Mimics 13.1 software.The patient's individual anatomy result was assessed from multiplanar(axial,coronal,sagittal)and 3D views(Fig.1).

With the mirroring tool,and using the median sagittal plane as a reference plane,normal anatomic structures of the target area were created by superimposing and comparing the unaffected sides with the affected ones.The normal contour of the affected skull base and glenoid fossa was ascertained,virtual surgery and final evaluation were performed on a virtual 3D mandible model on the basis of osteotomies positioning through calculating with Mimics 13.1 and the extent of reduction required(Fig.2).When the simulation was completed,the analysis of mandible model was exported to Mimics and Abaqus software sequentially.

The model generated in FEA was a linear elastic model with 288382 tetrahedral elements and 59815 nodes using 3-D 4-Node tetrahedral structural solid as element types.Mechanical properties werethen assigned to themodelin Abaqusaccording to those reported by Futterling et al.[9,10].for their cortical bone models(Table 1).There were statistical differences in properties among inner cancellous and outer cortical bone and titanium plate.

Fig.1 The 3D reconstruction of skull models CT scan data were input into the 3D imaging reconstruction workstation and saved as DICOM format and then virtual surgery was underwent after the definition of 3D coordinate sequences with 3D reconstruction software Mimics 13.1

Fig.2 The 3D models of reconstructed mandibles with various grafts

2.2 Finite element analysis

Several configurations of muscle forces and boundary conditions were considered in an attempt to accurately simulate situations of some mandible experiences during mastication and speech(Table 2).The muscle groups involved in the mastication process were positioned on the model in their realistic locations(Fig.3).A condition of nomastication or neutral position was created,i.e.nobiting occurred and the mandible was at rest.The sum of the moments about the TMJ was thus set to zero.The jaw-opening muscles composed of lateral pterygoid,digastric,geniohyoid and mylohyoid while the jaw-closing musles consisted of masseter,temporalis and medial pterygoid.Optimum length of the contractile element was defined by[(optimum sarcomere length—fiber length)/sarcomere length].The series elastic element was modeled as an inextensible wire(muscle length—fiberlength)(Table2).Combined with finite element and rigid-body analysis of human jaw joint dynamics,adjustments to the maximum muscle forces were made until a condition of neutral moments was induced or until almost zero motion was achieved(Table 2).

Table 1 Mechanical properties assigned inthe finite element models[9,10]

Table 2 Masticatory muscle forces[11,12]

Suitable boundary conditions had to be applied to prevent rigid body motion of the model when the muscle forces were activated.The upper area of the condyle,representing the TMJ,was fully constrained[13,14].This constraint was evident in the resulting FEA plots as a stress concentration at both condyles(Fig.3).

2.3 DataAnalysis

The study wasdesigned to evaluate the condyle process stress distribution values under 3 different conditions(normal and TMJ ankylosis conditionsundertwo differentvirtualsurgical operations)and ten points possessing comparable concentrated stressforcesin theanteriorand posterior positions of the left and right side of the dame condyle were chosen to evaluate condylar process stress distribution respectively.The quantitative data were described as±s.

3 Results

FEA simulations were carried out on mandible models of healthy,TMJ ankylosis,virtual surgery and virtual surgery with Titanium plate fixation for a neutral(no mastication)condition.Tensile stress was showed along the superior body of the healthy mandible.It should be noted that the high tensile stresses observed at each of the condyles are direct results of boundary conditions applied to the model as described previously.Areas of compressive stress are evident along the inferior border of the body and rame.The sections of high compressive stress in the condyles are again as results of the boundary conditions put in place.The symphysis region was displaced most during the simulation (Fig.4,healthy mandible).

One case of abnormal TMJ ankylosis was introduced to the model,a large area of tensile stress was observed on the left body and ramus.Increased tensile stresses were also observed at the condyles and coronoid process ofthe TMJ ankylosis.

Once Titanium plate was applied along the superior border of the angle region,a significant reduction in tensile stress was visible in a neutral position.Very little or no compressive stresses were observed,except for a small area on the inferiorright angle just above the fracture.The region of greatestdisplacementwasobserved along the symphysis(Table 3).

Fig.3 FEA model Muscle forces and boundary conditions by using the model

Fig.4 The Von Mises stress distribution

4 Discussion

The main causes of TMJ ankylosis are trauma and infection[15,16].Estimates of a traumatic origin range from 26%to 75%and that of infection from 44%to 68%.Roychoudhury et al.[17]retrospectively studied 50 cases of TMJ ankylosis and showed that trauma was documented as a major etiologic factor in 86%of all cases.In the present study the main causes were infection and trauma.This may be because most of our patients were from a less well-informed population without easy access to the specialty.The ensuing ankylosis could therefore be a reflection of misdiagnosis,or nondiagnosis of condylar fractures or infections,which had not be treated appropriately or even not be treated at all.

At present,clinicians always perform TMJ ankylosis resection surgery according to their own opinions on the TMJ ankylosis site including size and position of the TMJ ankylosis.There is a lack of validation for the techniques.As a result,a reconstruction of the mandible proves beneficiary as an clinical evidence that the plate will be operated correctly in a specific position according to the numerical analysis.Measurements from the 3D model are precise and can be used further for plate positioning during surgery.The scans are necessary for the 3D model reconstruction and routinely carried outwhen a facialTMJankylosisis detected.The main disadvantage of this protocol is the simplification of the model because of the lack of information available regarding the values of density for such regions like the gums,teeth and other soft tissues.

A healthy mandible is continuously subjected to stress,even in a state of‘rest’while stationary.In an effort to straighten the natural curve of the mandible,muscle movementscause the concentrated areas from tension and compression.These zones observed in the model while in a neutral position were consistent with ones described in previous literature by Gear et al.[18].Their work included the study of Champy et al.[19],who firstly demonstrated these zones of tension along the superior border and zones of compression along the inferior border of the mandible,using a simple cantilever beam model.Displacement at the angle region was further evidence of bending of the mandible due to the unopposed muscle forces.

Development of the large area of tensile stress in the fractured model was a result of the inability of the right body and ramus to support any stresses induced on it,due to the presence of a fracture.With no constraining surface,each side of the fracture was free to move independently.The tensile stressesat the condyle and coronoid process of the fractured segment were due to the increased force of the masseter and medial pterygoid muscles.The lower fractured segment moved downward in response to the direct unopposed pull of the suprahyoid muscle group.

Table 3 Value of condylar process stress distribution(MPa,±s)

Table 3 Value of condylar process stress distribution(MPa,±s)

Condylar process anterior posterior Position left right left Right Normal value 65.3±3.263.4±3.664.3±4.863.6±4.9 Virtual surgery(non-plate fixation)85.8±2.981.7±2.1113.7±25.1111.7±20.9 Virtual surgery(plate fixation)64.3±3.162.9±3.862.7±6.161.2±4.7

After plate fixation,tensile stresses are significantly reduced in the mandible due to the presence of the plate.When compared to the healthy model,tensile stress along the body and angle region of the TMJ ankylosis was absorbed by the titanium plate.The small area of compressive stress along the TMJ ankylosis was attributed to contact between both ankylosis surfaces.The displacement observed previously in the mandible model was stabilised by the plate.Compared with normal model, plate fixation model showed greater displacement while significant decrease was observed when compared to that of TMJ ankylosis model.The ankylosis was therefore stabilised by the plate because of the decrease of both tensile and compressive stresses. Displacement was also reduced,and muscle movements and forces were restored.

5 Conclusions

In summary,3D reco nstruction of the mandible using CT scan data is a valuable tool for clinicians during pre-surgical planning.It is proved to be of great benefit for TMJ gap arthroplasty by using a computer-based and FEA technique.The technique helps to prevent major complications such as injury of surrounding vital structures.Moreover,it has the advantages of reducing surgery time,promoting surgical security and plate placement accuracy,and allowing minimally invasive access.Although FEA method is a valuable tool for clinicians to carry out for pre-surgical planning and virtual operation.It is still needed to be verified during clinical surgery and further observe by postoperative patient CT data.

1 Chidzonga MM.Temporomandibular joint ankylosis:review of thirty-two cases[J].Br J Oral Maxillofac Surg,1999,37(2):123-126.

2 Chiu WK,Luk WK,Cheung LK.Three-dimensional accuracy ofimplantplacementin a computer-assissted navigation system[J].Int J Oral Maxillofac Implants,2006,21(3):465-470.

3 Lopez EN,Dogliotti PL.Treatment of temporomandibular joint ankylosis in children:Is it necessary to perform mandibular distraction simultaneously?[J].J Craniofac Surg,2004,15(5):879-885.

4 Donkor P,Acheampong AO.Intra-articular ramus ostectomy combined with costochondral grafting for the treatment of recurrent ankylosis of the mandible[J].Br J Oral Maxillofac Surg,2006,44(6):497-500.

5 Su-Gwan K.Treatment of temporomandibular joint ankylosis with temporalismuscle and fascia flap [J].IntJOral Maxillofac Surg,2001,30(3):189-193.

6 Schobel G,Millesi W,Watzke IM,et al.Ankylosis of the temporomandibular joint.Follow-up of thirteen patients[J].Oral Surg Oral Med Oral Pathol,1992,74(1):7-14.

7 Vasconcelos BC,Porto GG,Bessa-Nogueira RV,et al.Surgical treatment of temporomandibular joint ankylosis:Follow-up of 15 cases and literature review[J].Med Oral Patol Oral Cir Bucal,2009,14(1):E34-E8.

8 Yu HB,Shen,GF,Zhang SL,et al.Navigation-guided gap arthroplasty in the treatmentoftemporomandibularjoint ankylosis[J].Int J Oral Maxillofac Surg,2009,38(10):1030-1035.

9 Futterling S,Klein R,Straber W,et al.Automated finite element modelling of a human mandible with dental implants[C].In:Proceedings of the sixth international conference in CentralEurope on computer graphics and visualization.University of West Bohemia,Campus Bory,Plzen,Czech Republic,1998,103-110.

10 Kavanagh EP,Frawley G,Kearns G,et al.Use of finite element analysis in presurgical planning: treatment of mandibular fractures[J].Ir J Med Sci,2008,177(4):325-331.

11 Koolstra JH,Van-Eijden TM.Combined finite-element and rigid-body analysisofhuman jaw jointdynamics[J].J Biomech,2005,38(12):2431-2439.

12 Zee MD,Cattaneo PM,Dalstra M.On the development of a rigid-body modelofthe mandible [C].In:International Symposium on Computer Simulation in Biomechanics.Cleveland,Ohio,USA,2005,28-30.

13 Erkmen E,Simsek B,Yucel E,et al.Three-dimensional finite element analysis used to compare methods of fixation after sagittal split ramus osteotomy: setback surgery-posterior loading[J].Br J Oral Maxillofac Surg,2005,43(2):97-104.

14 Nagasao T,Kobayashi M,Tsuchiya Y,et al.Finite element analysis of the stresses around endosseous implants in various reconstructed mandibular models[J].J Craniomaxillofac Surg,2002,30(3):170-177.

15 Chidzonga MM.Temporomandibular joint ankylosis:review of thirty-two cases[J].Br J Oral Maxillofac Surg,1999,37(2):123-126

16 Kaban LB,Perrott DH,Fisher K.A protocol for management of temporomandibular joint ankylosis[J].J Oral Maxillofac Surg,1990,48(11):1145-1152.

17 Roychoudhury A,Parkash H,Trikha A.Functional restoration by gap arthroplasty in temporomandibular joint ankylosis:a report of 50 cases[J].Oral Surg Oral Med Oral Pathol Oral Radiol Endod,1999,87(2):166-169.

18 Gear AJL,Apasova E,Schmitz JP,et al.Treatment modalities for mandibular angle fractures.J Oral Maxillofac Surg,2005,63(5):655-663.

19 Champy M,Lodde JP,Jaeger JH.Biomechanical basis of mandibularosteosynthesisaccording to the FX Michelet method[J].Rev Stomatol Chir Maxillofac,1976,77(1):248-251.