+ All Categories
Home > Documents > Computer Aided Biomodeling and Analysis of Patient Specific Porous Titanium Mandibular Implants

Computer Aided Biomodeling and Analysis of Patient Specific Porous Titanium Mandibular Implants

Date post: 31-Jan-2023
Category:
Upload: independent
View: 0 times
Download: 0 times
Share this document with a friend
9
Jayanthi Parthasarathy B.D.S. e-mail: [email protected] Binil Starly Assistant Professor e-mail: [email protected] Shivakumar Raman David Ross Boyd Professor e-mail: [email protected] Center for Shape Engineering and Advanced Manufacturing, School of Industrial Engineering, University of Oklahoma, 202 W. Boyd Street, Room 124, Norman, OK 73019-0631 Computer Aided Biomodeling and Analysis of Patient Specific Porous Titanium Mandibular Implants Custom implants for the reconstruction of mandibular defects have recently gained im- portance due to their better performance over their generic counterparts. This is attrib- uted to their precise adaptation to the region of implantation, reduced surgical times, and better cosmesis. Recent introduction of direct digital manufacturing technologies, which enable the fabrication of implants from patient specific data, has opened up a new horizon for the next generation of customized maxillofacial implants. In this article, we discuss a representative volume element based technique in which precisely defined po- rous implants with customized stiffness values are designed to match the stiffness and weight characteristics of surrounding healthy bone tissue. Dental abutment structures have been incorporated into the mandibular implant. Finite element analysis is used to assess the performance of the implant under masticatory loads. This design strategy lends itself very well to rapid manufacturing technologies based on metal sintering processes. DOI: 10.1115/1.3192104 Keywords: CT reconstruction, mandibular implants, porous titanium, layered manufacturing 1 Introduction Mandibular reconstruction after tumor resection presents a sig- nificant challenge to maxillofacial surgeons today 1. Mandibular defects occur due to trauma, orofacial tumors, infection, and other space occupying lesions like cysts. Defects in continuity of the mandible lead to severe facial deformities, major difficulties in verbalization, deglutition, and mastication. The extent of the af- fliction increases with the size and location of the defect. Other than somatic effects, the patient’s quality of life is significantly affected due to the loss of mandibular continuity. Several methods of defect reconstruction, such as autogenous bone grafts, generic alloplastic bone plates, and custom mandibu- lar reconstruction trays are being used presently. Autografts are the “gold standard” from an immune response point of view. However, the use of autografts are limited due to issues such as donor site morbidity, lack of sufficient graft quantity, chances of infection, and patient discomfort 2–5. Generically available mandibular reconstruction plates do not confirm to the exact mor- phology of the resected part of the mandible, making it difficult for post operative dental reconstruction Figs. 1a and 1b. Mar- tola et al. 6 suggested plates that match closely the three- dimensional shape of the mandible to avoid requirements of intra- operative bending. Custom designed and fabricated implants have been found to have advantages of better fit, reduced operating time, and lesser chances for infection, faster recovery, and better cosmesis in cran- iofacial surgery 7–9. Custom mandibular trays have been used for mandibular reconstruction by Samman et al. 10. A mandibu- lar replica was reconstructed from clinical measurements made from the patient’s mandible. The titanium implant is built through a casting or swaging process or CNC milling Fig. 2. Yaxiong et al. 11 and Singare et al. 12,13 fabricated custom mandibular titanium implants using computer aided design CAD software and rapid prototyping RP technologies. Patient specific CT scan data of the defect site was used to produce a virtual reconstruction of the mandibular implant. Data from the contralateral side were mirrored and processed in a virtual CAD environment to arrive at the final shape of the mandibular implant. A stereolithographic SLA physical prototype model of the skull and the final design of the implant were fabricated. The SLA negative mold of the implant was used as a sacrificial part to cast titanium into the mold. Holes were then drilled into the body of the implant to reduce its weight. This process is now extensively used in a clini- cal setting and has greatly improved treatment modalities, but still have inherent deficiencies, enumerated below, making it necessary to explore newer methods for improved treatment outcome. The deficiencies of the above process are as follows. a Titanium implants built using this process are often heavy and can cause discomfort to the patients. The Young’s modulus of titanium is almost five times that of cortical bone resulting in stress shielding effects 14,15. b Secondary processing increases cost and time, resulting in subsequent delay in treatment and increased cost of health care to the patient/insurance provider. c RP models are used only as sacrificial models and sec- ondary manufacturing methodologies, such as casting and swaging, are required to be used for fabrication of the final implants. While virtual reconstruction of the mandible implant has re- sulted in significant design improvements, newer technologies to custom design the implant to match the mechanical properties of the surrounding tissue need to be adopted. New generation of implants would be porous, enabling the in-growth of healthy bone tissue for additional implant fixation and stabilization. Newer im- plants would conform to the external shape of the defect site that is intended to be replaced. More importantly, the effective elastic modulus of the implant should match that of surrounding tissue. Ideally the weight of the implant should also be equal to the weight of the tissue that is being replaced, resulting in increased Manuscript received February 24, 2009; final manuscript received: April 29, 2009; published online September 1, 2009. Review conducted by Vijay Goel. Journal of Medical Devices SEPTEMBER 2009, Vol. 3 / 031007-1 Copyright © 2009 by ASME Downloaded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm
Transcript

1

ndsmvflta

bltHdimpftdo

hciflfaat

2

J

Downlo

Jayanthi ParthasarathyB.D.S.

e-mail: [email protected]

Binil StarlyAssistant Professor

e-mail: [email protected]

Shivakumar RamanDavid Ross Boyd Professor

e-mail: [email protected]

Center for Shape Engineering and AdvancedManufacturing,

School of Industrial Engineering,University of Oklahoma,

202 W. Boyd Street, Room 124,Norman, OK 73019-0631

Computer Aided Biomodeling andAnalysis of Patient SpecificPorous Titanium MandibularImplantsCustom implants for the reconstruction of mandibular defects have recently gained im-portance due to their better performance over their generic counterparts. This is attrib-uted to their precise adaptation to the region of implantation, reduced surgical times, andbetter cosmesis. Recent introduction of direct digital manufacturing technologies, whichenable the fabrication of implants from patient specific data, has opened up a newhorizon for the next generation of customized maxillofacial implants. In this article, wediscuss a representative volume element based technique in which precisely defined po-rous implants with customized stiffness values are designed to match the stiffness andweight characteristics of surrounding healthy bone tissue. Dental abutment structureshave been incorporated into the mandibular implant. Finite element analysis is used toassess the performance of the implant under masticatory loads. This design strategylends itself very well to rapid manufacturing technologies based on metal sinteringprocesses. �DOI: 10.1115/1.3192104�

Keywords: CT reconstruction, mandibular implants, porous titanium, layeredmanufacturing

IntroductionMandibular reconstruction after tumor resection presents a sig-

ificant challenge to maxillofacial surgeons today �1�. Mandibularefects occur due to trauma, orofacial tumors, infection, and otherpace occupying lesions like cysts. Defects in continuity of theandible lead to severe facial deformities, major difficulties in

erbalization, deglutition, and mastication. The extent of the af-iction increases with the size and location of the defect. Other

han somatic effects, the patient’s quality of life is significantlyffected due to the loss of mandibular continuity.

Several methods of defect reconstruction, such as autogenousone grafts, generic alloplastic bone plates, and custom mandibu-ar reconstruction trays are being used presently. Autografts arehe “gold standard” from an immune response point of view.owever, the use of autografts are limited due to issues such asonor site morbidity, lack of sufficient graft quantity, chances ofnfection, and patient discomfort �2–5�. Generically available

andibular reconstruction plates do not confirm to the exact mor-hology of the resected part of the mandible, making it difficultor post operative dental reconstruction �Figs. 1�a� and 1�b��. Mar-ola et al. �6� suggested plates that match closely the three-imensional shape of the mandible to avoid requirements of intra-perative bending.

Custom designed and fabricated implants have been found toave advantages of better fit, reduced operating time, and lesserhances for infection, faster recovery, and better cosmesis in cran-ofacial surgery �7–9�. Custom mandibular trays have been usedor mandibular reconstruction by Samman et al. �10�. A mandibu-ar replica was reconstructed from clinical measurements maderom the patient’s mandible. The titanium implant is built throughcasting or swaging process or CNC milling �Fig. 2�. Yaxiong et

l. �11� and Singare et al. �12,13� fabricated custom mandibularitanium implants using computer aided design �CAD� software

Manuscript received February 24, 2009; final manuscript received: April 29,

009; published online September 1, 2009. Review conducted by Vijay Goel.

ournal of Medical Devices Copyright © 20

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

and rapid prototyping �RP� technologies. Patient specific CT scandata of the defect site was used to produce a virtual reconstructionof the mandibular implant. Data from the contralateral side weremirrored and processed in a virtual CAD environment to arrive atthe final shape of the mandibular implant. A stereolithographic�SLA� physical prototype model of the skull and the final designof the implant were fabricated. The SLA negative mold of theimplant was used as a sacrificial part to cast titanium into themold. Holes were then drilled into the body of the implant toreduce its weight. This process is now extensively used in a clini-cal setting and has greatly improved treatment modalities, but stillhave inherent deficiencies, enumerated below, making it necessaryto explore newer methods for improved treatment outcome. Thedeficiencies of the above process are as follows.

�a� Titanium implants built using this process are oftenheavy and can cause discomfort to the patients. TheYoung’s modulus of titanium is almost five times that ofcortical bone resulting in stress shielding effects �14,15�.

�b� Secondary processing increases cost and time, resultingin subsequent delay in treatment and increased cost ofhealth care to the patient/insurance provider.

�c� RP models are used only as sacrificial models and sec-ondary manufacturing methodologies, such as castingand swaging, are required to be used for fabrication ofthe final implants.

While virtual reconstruction of the mandible implant has re-sulted in significant design improvements, newer technologies tocustom design the implant to match the mechanical properties ofthe surrounding tissue need to be adopted. New generation ofimplants would be porous, enabling the in-growth of healthy bonetissue for additional implant fixation and stabilization. Newer im-plants would conform to the external shape of the defect site thatis intended to be replaced. More importantly, the effective elasticmodulus of the implant should match that of surrounding tissue.Ideally the weight of the implant should also be equal to the

weight of the tissue that is being replaced, resulting in increased

SEPTEMBER 2009, Vol. 3 / 031007-109 by ASME

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

pmtbpfpTau

aabmpccm

2

Fscwrspmioa

Fc

0

Downlo

atient comfort. All of the above requirements would have to beade at an affordable cost that is at least equal to if not less than

he present healthcare costs. The recent introduction of electroneam melting �EBM� and direct metal laser sintering �DMLS� forrocessing of titanium has led to the possibility of a one stepabrication of porous custom titanium implants with controlledorosity and desired external and internal characteristics �16�.ypically these rapid manufacturing technologies are utilized inerospace applications but the systems can be easily extended forse in the fabrication of medical implants.

The objective of this study was to develop a design strategy forpatient specific porous titanium based mandibular implant with

n ideal porosity and desired density taking into considerationoth aesthetic and functional requirements. Dental implant abut-ent structures are built into the implant for assessment of the

orous implant performance during mastication. Effective me-hanical properties of the implant are varied and effects of masti-atory forces on the porous implant are studied using finite ele-ent analysis �FEA� methods.

Materials and MethodsThe general methodology followed in this paper is outlined in

ig. 3. The input data consists of helical two-dimensional �2D� CTcan slices of the patient. 1 mm slices are used extending from 1m below and above the region of interest. Medical imaging soft-are MIMICS™ �Materialise Inc., Leuven, Belgium� is used to

econstruct the three-dimensional �3D� model of the mandiblehowing the defect. Following this, a virtual surgical planningrocedure is performed. 3D data from the contralateral side isirrored to the defect site to arrive at the external geometry of the

mplant. Using point cloud reconstruction tools, the external ge-metry is converted to a CAD based digital model. Dental implantbutments and retention structures for osteosynthesis screws are

Fig. 1 „a… Mandible reconstruction with gemor mandible

ig. 2 Titanium mandible implant fabricated using CNC ma-

hining process

31007-2 / Vol. 3, SEPTEMBER 2009

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

modeled into the implant. A bounding volume is then createdaround the CAD model to enclose the complete model excludingthe implant abutments and the retention structures. Pores are de-signed and propagated throughout the bounding volume. The po-rous bounding cube is imported into MIMICS™, wherein a Booleanoperation is performed to create the porous implant structure. Thisporous model can be further used to drive downstream rapidmanufacturing systems to fabricate the implant.

2.1 External Geometry of the Implant. A case with man-dibular tumor extending only on the left side is chosen for recon-struction as it is a common clinical finding requiring reconstruc-tion. MIMICS™ medical imaging software is used to create the 3Dmodels by performing thresholding and region growing operationsFigs. 4�a� and 4�b�. A final rendered virtual model of the patient’sreconstructed mandible is shown in Fig. 4�c�.

Virtual reality has played an important part in surgical simula-tion as the digital model can be cut, repositioned and tried untilsatisfactory results of the postoperative outcome is arrived at.Surgical simulation for three common clinical scenarios is con-sidered here: Model 1—anterior canine to canine resection,Model 2—premolar to subcondylar resection, and Model3—hemimandibulectomy were conceived. Image data from thecontralateral side are mirrored for deriving the external shape ofthe corrected site, as shown in Fig. 4. The corrected model willthen form the external shape of the metallic implant. All the mod-els are converted to a usable digital format �.iges� file type usingthe method described by Starly et al. �17�.

2.2 Dental Abutment and Retention Structures. Retentionstructures in the form of plates are required for fixing the implantto the normal mandible. Dental implant abutments are also de-signed to enable the implants to be used in a clinical scenario overwhich fixed prosthodontic restorations can be fixed. The retentionstructures are 15 mm in length, 5 mm in width, and 1 mm inthickness and have two holes of 2 mm diameter placed at 3 mmfrom the edge to fix osteosynthesis screws. Three implants abut-ments are designed as per specifications. All implants are 6 mm inheight and have a taper of 15 deg between the top and bottom

ic reconstruction plate and „b… resected tu-

Fig. 3 Roadmap for the design of mandible implant from CT

ner

images

Transactions of the ASME

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

s6r5

�v6sbTwsewbsbi

spprmfuraowat

J

Downlo

urfaces. The implants with diameters of 3.5 mm, 5.00 mm, and.00 mm are placed in the anterior, premolar, and molar regions,espectively. The three implants with the fixtures are seen in Figs.�a�–5�c�.

Density and corresponding weight of the corrected site modelFig. 5�c�� with two different materials are shown in Table 1. Theolume of the model as estimated by the software is given to be0,626 mm3. If the mandible corrected model were to be built ofolid titanium alloy, it would be 2.2 times heavier than the originalone it would replace. More importantly, the elastic modulus ofitanium alloy is about 114 GPa, while that of cortical boneould be 20 GPa. This relative difference in stiffness results in

ignificant stress shielding effects. Therefore, it is imperative tostablish two design criteria for the new implant: �1� reduce theeight of the implant such that it is close to that of the naturalone and �2� reduce the effective stiffness of the material whiletill having titanium alloy as the primary choice of the implantiomaterial. Both of the above two criteria can be met by buildingn porous structures within the mandible structure.

2.3 Internal Porous Architecture of the Implant. Some es-ential considerations while designing porous implants are �1�ore sizes to allow for the in-growth of new bone; �2� final im-lant weight and mechanical properties to be close to that of sur-ounding bone tissue to prevent stress shielding; �3� implantanufacturability and repeatability. The porous implant, apart

rom providing structural support, would also have to function innison with the host tissue as an integral part of the mandible. Toeduce the weight of the titanium alloy implant, we have patternedseries of square holes propagated throughout the interior volumef the implant. The patterning of the holes would reduce theeight, reduce the effective Young’s modulus �effective stiffness�,

nd provide a network of pores for the in-growth of healthy boneissue. To achieve this internal porous architecture, a square

Fig. 4 „a… Thresholding operation; „b… reconstructed mandition of external geometry of the implant

Fig. 5 „a… Canine to canine reconstruction—Model 1; „b… pr

dible reconstruction—Model 3

ournal of Medical Devices

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

bounding volume cube is created containing square pores eachmeasuring 1.5�1.5 mm2 propagated throughout the boundingvolume in the XY, XZ, and YZ planes. This porous block is im-ported into MIMICS™ and intersected with the original solid im-plant, as shown in Fig. 6. The resultant porous model is seen inFig. 7. The final porous model is matched with the healthy rightside mandible of the patient �Fig. 8�.

The resultant porous implant will have an equivalent porositygiven by Eq. �1� below. The weight of the porous implant can alsobe calculated given the density of the titanium alloy. The porosityequation is given as

P =V1 − V2

V1�1�

where V1 is the volume of the solid implant, and V2 is the volumeof the final porous implant. The dimensions of the pore are alsovaried to obtain different weight characteristics. In Sec. 2.4, wewill describe how pore dimensions influence the effective stiffnessof the implant.

2.4 Assessment of the Porous Implant Performance UsingFEA Techniques. The long term successful retention and functionof the porous titanium implant depends not only on the implantproviding structural support but also biofunctionality of the pros-thesis. Biofunctionality is defined as the mechanical and physicalproperties that enable the implant to perform its functioning inunison with the host tissue. In this context, this function would beto behave as an integral part of the mandible, taking part in mas-ticatory functions generating and transferring stresses to the ad-joining bone. This is essential for maintaining the natural balancebetween bone apposition and resorption process. High stressesmay lead to bone resorption and ultimate failure, while lowstresses at the bone implant interface could lead to stress shielding

with defect; „c… virtual surgical simulation and reconstruc-

olar to subcondylar reconstruction—Model 2; „c… hemiman-

ble

em

SEPTEMBER 2009, Vol. 3 / 031007-3

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

l

eamusTvw

mcarl

M

MM

Fw

0

Downlo

eading to aseptic loosening due to failure of bone apposition.Recently FEA techniques have been utilized in the biomedical

ngineering context, for analyzing parts of human body using re-listic biomechanical data for the relevant tissues and alloplasticaterials that are used for prosthesis fabrication. FEA has been

sed by Knoll et al. �18� and Tie et al. �19� effectively to study thetresses and biomechanical effects in mandibular reconstruction.herefore the effective mechanical properties of the representativeolume element and performance of the patient specific implantere estimated using FEA.The geometrical complexity of the porous implant �Fig. 8�akes it computationally intensive and cumbersome for the dis-

retization procedure during the mesh definition stage of the FEAnalysis. To avoid this potential bottleneck, we have utilized theepresentative volume element �RVE� method described by Hol-ister et al. �20�, Starly et al. �21�, and Fang et al. �22,23�. In this

Table 1 Estimated weight of the implant

aterial typeDensity�g /cm3�

Weight�g�

andible �bone� 2 121.25andible �titanium alloy� 4.43 268.57

Fig. 6 Bounding porous cube intersected with the implant

ig. 7 Resultant porous implant after the boolean operation

ith interconnected pores

31007-4 / Vol. 3, SEPTEMBER 2009

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

homogenization method, an RVE is selected to represent the po-rous structure of the implant, and the effective Young’s modulusof the porous RVE is then calculated. Once the effective Young’smodulus is determined for a range of porosity values, this valuecan then be the material input into the FEA analysis for assessingthe performance of the implant.

The selected RVE for our implant is the square pore holes, asshown in Fig. 9�a�. The square pore is constrained at one face,while a strain of 0.1% is applied at the opposite face. Periodicboundary conditions were assumed at all other faces. The RVE ismeshed using four node tetrahedral elements. An FEA programANSYS™ �ANSYS Inc.� is used to predict the effective stiffness ofthe porous structure. The effective stiffness can be calculated byusing Eq. �2� as follows

EXX =�X

�X=

� RX

ASX1

��UX

LX� =

RX

0.001 � Ax�2�

where Ax is the area of cross section of the face Sx1, and Rx is theaverage reaction force on the surface Sx1. Since the unit cell issymmetric in the x, y, and z directions, we obtain Exx=Eyy =Ezz=Eeffective. The prediction of effective stiffness was performed fora range of RVE porosities ranging from 28.18% to 78.4%.

2.5 FEA Based Analysis of Mandibular Implants in Re-sponse to Masticatory Forces. Apart from precise fitting andintegration, the mandibular implants would require functioning inunison with the rest of the mandible during mastication. The threeimplant design models 1, 2, and 3 �Figs. 5�a�–5�c�� describedearlier are imported into PRO-MECHANICA™ �PTC Corp., MA�,and the finite element method is used to predict the stresses andstrains generated in the implant during mastication. A pmesh isgenerated within the structure and the three models—anterior, pre-molar to subcondylar, and hemimandible implants—had 722, 777,and 1475 tetrahedral nodes, a respectively.

The material properties assumed in this study are shown inTable 2. The Young’s modulus and the assumed Poisson’s ratio forthe three different material types are shown. Cortical bone is as-sumed to be at 20 GPa, while dense titanium alloy is assumed tobe at 114 GPa. We have considered different Young’s modulusvalues for the porous titanium structure ranging from 3 GPa to114 GPa corresponding to porosities ranging from 0% to 80%.

All loads are assumed vertical and compressive. Table 3 showsthe loads applied on the dental implant abutments, as used byWang et al. �24�. Loads are applied on the two implant abutmentsfor models 1 and 2 as per the Table 2, as shown in Figs. 10�a� and10�b�. Since model 3 is a hemimandible reconstruction, three loadpatterns are assumed: �1� incisor loading, �2� premolar and molarloading, and �3� incisor, premolar, and molar loading, as seen inFig. 10�c�. Loads are applied on the implant abutments, and con-straints are applied in the X, Y, and Z directions on all holesprovided for fitting of osteosynthesis screws. Additionally, in

Fig. 8 Porous mandibular implant fitted to the normal rightmandible

model 3 constraints were applied on the condylar region of the

Transactions of the ASME

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

mii

3

sspmtotd

M

MTP

I

6

J

Downlo

andible. von Mises stresses, displacement, and strains generatedn the implants are determined to evaluate the performance of themplants under masticatory loads.

Results

3.1 Effective Young’s Modulus of Porous RVE. A relation-hip between porosity and effective elastic modulus of the con-tructs is first arrived at and the values are used to predict theroperties of the patient the specific implants. For the FEAethod the stress-strain values at the constrained nodes are plot-

ed and the effective elastic modulus is estimated from the slopef the stress-strain curve of the porous constructs. As expected,he effective stiffness of the porous RVE �Ti alloy� gradually re-uced with increasing porosity values. However, for all useful

Fig. 9 „a… RVE with applied boundafor stress distribution within the RVE

Table 2 Material properties

aterialElastic modulus

�GPa�Poisson’s

ratio

andible-cortical bone 20 0.33itanium �dense� 114 0.32orous titanium alloy implants 3 �83.6 %�, 5 �80.6 %�,

10 �73.2%�, 15 �66.4%�,20 �60.2%�, 30 �49.3�, 114 �0�

0.32

Table 3 Loads applied on dental implant abutments

ncisor Premolar Molar

0 150 300

Fig. 10 „a…, „b…, and „c… Loads and constra

ournal of Medical Devices

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

applications, typically porosity values ranging from 55% to 85%are selected, which gives a corresponding effective stiffness torange from 24.7 GPa to 3.2GPa.

Using Fig. 11, it is seen that a 60% porous Ti alloy should havean effective Young’s modulus of about 19.51 GPa. If a 60% po-rous structure is incorporated into Model 3 �Fig. 9�c��, the finalproperties, such as weight and effective volume, can be estimated.A comparison of the properties of the solid titanium, porous tita-nium, and a bone model of the mandibular implant is displayed inFig. 12. Notice that a porosity value of 60% is selected to matchthe weight characteristics of natural healthy bone. Also, the effec-tive stiffness of the implant would match that of the cortical bone.

3.2 Performance of Patient Specific Mandibular Implants1, 2, and 3 in Response to Masticatory Stresses. von Misesstress is commonly used to study the stress measurements in man-dible reconstruction �18,25–27�. In this study, von Mises stressesgenerated during masticatory vertical compressive stresses arestudied to arrive at an acceptable porosity of the implant to func-tion well in a clinical scenario.

Model 1 canine to canine reconstruction. Maximum von Misesstresses of 56.94 MPa is seen to be in the region of the first set ofscrews on the retention structures bilaterally, as seen in Fig. 13.Maximum principal strain was seen in the region of the first set ofscrews. Displacements are seen in the X, Y, and Z directions ofwhich the X and Z displacements were negligible. The maximumdisplacement is seen in the Y direction, at 0.01020 mm for aneffective elastic modulus of 3 GPa, which reduces to 0.00027 mmwhen the elastic modulus is increased to that of dense titanium.The maximum Y displacement is seen in the lower part of theretention structures close to the distal ends of the implant.

onditions and „b… FEA contour plot

ry c

ints on mandibular models 1, 2, and 3

SEPTEMBER 2009, Vol. 3 / 031007-5

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

viamt

rla

ovsrms

Fm

0

Downlo

Model 2 premolar to subcondylar reconstruction. Maximumon Mises stresses for model 2 is 269.9 MPa, as seen in Fig. 14,s also seen at the screws on the first set of retention structuresttached to the proximal end of the implant. A maximum displace-ent of 0.06772 mm in the X direction, 0.045 mm in the Y direc-

ion, and 0.05284 mm in the Z direction are seen.Model 3 hemimandible reconstruction. When a hemimandible

eplacement is done, the implant could be subjected to varyingoading patterns. Simulations of varying loading patterns are dones follows:

1. incisor2. premolar3. molar4. premolar and molar5. incisor, premolar, and molar

Maximum von Mises stresses for model 3 is seen at the screwsn first set of retention structures, as can be observed in Fig. 15.on Mises stresses and shear stresses for the five load patterns arehown in Table 4. The maximum principal stresses are found toange from a minimum of 18.66 MPa for load pattern 1 to aaximum of 90 MPa for load pattern 5. Similarly the maximum

hear stresses ranged from a minimum of 22.44 MPa for load

Fig. 11 Effective elastic modulus predicted by FEA method

ig. 12 Comparison of properties of solid and porous titaniumandibular implant and its bone equivalent

Fig. 13 Model 1 von Mises stresses under masticatory load

31007-6 / Vol. 3, SEPTEMBER 2009

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

pattern 1 to a maximum of 149.2 MPa for load pattern 5. Maxi-mum principal strains were observed in the proximal retentionstructures on the first set of screws and were very low. The maxi-mum principal strains reduce from 0.03 for load pattern 5 for aneffective elastic modulus of 3 GPa �maximum load-minimumstrength� to 0.0006827 when the elastic modulus equaled that ofdense titanium. Maximum displacements are seen for load pattern5 for an effective elastic modulus of 3 GPa and displacements in

Fig. 14 Model 2 von Mises stresses—maximum stresses seenat the screw joints

Fig. 15 Model 3 von Mises stresses—maximum stresses seenat the screw joints

Table 4 Model 3 von Mises stresses

Loadpattern Load locations

Totalload�N�

von Misesstress�MPa�

Shearstress�MPa�

1 Incisor 60 43.01 22.442 Premolar 150 86.64 50.463 Molar 300 164.3 77.994 Premolar and molar 450 245.8 127.45 Incisor, premolar, and molar 510 287.9 149.2

Transactions of the ASME

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

t0p

4

faasonawdficmrb

csiaifinctttspn

iaat0l

ottshi

J

Downlo

he X, Y, and Z directions were 0.04 mm, 0.0619 mm, and.003909 mm, respectively. Maximum displacements for all loadatterns are seen in the Y direction.

DiscussionThe study has developed a repeatable design strategy for the

abrication of porous titanium mandibular implants with predict-ble properties. The main advantages over the previous methodsre that implants with controlled shape and porosity can be de-igned for ingrowth of tissues for better integration �28�. The sec-nd advantage of the design strategy is the restoration of the origi-al anatomy for better esthetics that would improve patient’scceptance. Dental rehabilitation, which has been a major problemith generic reconstruction plates and bone grafts, has been ad-ressed by construction of dental implant abutments over whichxed prosthodontic restorations can be made for improved masti-atory functions. Mechanical properties of the implants have beenodified for better integration. Weight of the implant has been

educed and effective elastic modulus of the implant material haseen reduced close to that of cancellous and cortical bone.

4.1 Stresses Generated in Model 1, 2, and 3 Due to Masti-atory Loads. Medical grade titanium has an ultimate tensiletrength of 970 MPa and a yield strength of 930 MPa. The implanttself can be porous but the retention structures and the dentalbutments are fabricated as dense parts. Failure of the titaniummplant from a strength perspective is highly unlikely. Howeverailure of the implant can occur due to high stresses at the screwsn the retention structure, which would be transmitted to theeighboring cortical bone with a tensile strength of 92–185 MPa,ausing microfissures leading to loosening of the implant in shorterm �29�. In the long run, constant bone remodeling can reducehe risk; however, constant high stresses can lead to bone resorp-ion leading to loosening of the implant �18�. Since the retentiontructures for all the three implants resembled fracture fixationlates the results were compared with stresses generated in ge-eric fracture fixation plates.

Model 1. The maximum von Mises stress is 56.94 MPa, whichs much lower compared with the tensile strength of titanium. Themount of stresses transferred to the screws and on to the corticalnd cancellous bone would be very less than that of bone andherefore within acceptable limits. Maximum displacement of.01020 mm seen in the Y direction is well within the toleranceimits.

Model 2. The premolar to subcondylar model represented onef the most common clinical scenarios. Restoration of facial es-hetics and masticatory functions are the main criteria in designinghis implant. Buccal rotation of the implant during mastication,hort thin segment of the residual ramus for distal fixation, andigh masticatory loads are some critical considerations in design-

Table 5 Model 2—stress inscrews of variou

ng the implant. Hence the right design would have the least

ournal of Medical Devices

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

stresses at the osteosynthesis screw holes in the retention struc-tures. von Mises stresses at the screws were 269.9 MPa, only 25%of the ultimate strength of titanium giving a safety factor of 4. Ahigh factor of safety is desired due to the fact that we have per-formed a static study and the assumption of a uniform elasticmodulus for the implant. Increasing the size of the screws isshown to reduce the stresses. Maurer et al. �30� found the maxi-mum tensile strength of 610 MPa for 1.5 mm and 2 mm screws,and the acceptable chewing force would be 89.1 N and 157.5 N,respectively. Knoll et al. �18� increased the diameter of the stan-dard reconstruction plates and their new design of reconstructionplates from 2.7 mm to 4 mm �1.5 times�. They found the stressesat the screws reduced from 1112 N to 199 N, 918 N to 191 N, 109N to 66 N, and 74 N to 66 N, respectively. The authors also variedthe configuration of the screws and the least stresses were foundwith rectangular configuration of the screws.

In this study for Model 2, to reduce the stresses, the followingmodifications are made to the original design of the retentionstructures seen in Fig. 5. Table 5 shows the von Mises stresses inthe various modifications. Increasing the number of screws tothree �model 2a� while keeping the diameter constant did notshow significant reduction in stresses—254.7 MPa. Increasing thediameter by 1.5 times to 3 mm �model 2b� reduced the von Misesstresses from 269.9 MPa by 51.76% to 130.2 MPa. The stressesare further reduced to 117.8 MPa �Fig. 16� by merging the twoplates at the proximal end into one, as shown in Table 5—model2c.

Model 3. From Table 5, it is observed that Model 2c with the4–�3 mm screw plate has the least stress associated with theplate design. Therefore this plate design is incorporated in Model3a, as shown in Fig. 17. A maximum load of 510 N was appliedon the incisor, premolar, and molar abutments �Table 4�. The von

odifications of the retention plate systems

s m

Fig. 16 von Mises stress in the finalized design for model 2c

SEPTEMBER 2009, Vol. 3 / 031007-7

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

MFpnMo

c�tsmw

sdtbatCSbtsFni

0

Downlo

ises stress patterns are reduced, as seen in Figs. 17 and 18. Fromig. 17, it is seen that the stresses are reduced by 44.35% com-ared with Model 3 �Fig. 15�. The maximum von Mises stress forormal load on the incisor, premolar, and molar together is 150.2Pa. The final design of the implant would therefore be capable

f withstanding five times the normal load.

4.2 Micromotion. Micromotion is unavoidable during masti-ation. Bone tolerates micromotion in the range of 50–150 �m31� beyond which bone resorption occurs leading to loosening ofhe implant. The displacements in the X, Y, and Z directions weretudied for the three implant models, and the maximum displace-ent is found to be 0.068 mm, as seen in Fig. 19, which is wellithin the tolerance limits.

4.3 Fabrication of Porous Mandible by 3D Printing. Totudy the feasibility of fabrication of patient specific porous man-ible with inbuilt dental abutments, a realistic digital reconstruc-ion of a porous model, as shown in Fig. 20, is conceived. Theody of the mandible is designed with square pores in the X, Y,nd Z direction and dental abutments are designed as dense struc-ures. The STL file of the model is transferred to a 3D printer �Zorp.� and printed. The rapid prototype model is seen in Fig. 21.imilar procedures can be used to fabricate the model in metalased rapid manufacturing systems such as EBM and DMLS sys-ems available from commercial vendors. Sample porous titaniumtructures with varying porosities fabricated with EBM are seen inig. 22. The entire porous model in medical grade titanium wasot made due to cost considerations. Future research activities willnclude the in-vivo evaluation of the porous structures.

Fig. 17 Final model 3b von Mises stresses

Fig. 18 Models 3a and 3b comparison von Mises stress

31007-8 / Vol. 3, SEPTEMBER 2009

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

5 ConclusionA design strategy has been developed for the eventual direct

fabrication of titanium implants for mandibular reconstructionwith mechanical properties close to that of bone. Advantages ofthis design strategy are preparation of custom porous titaniumimplants with controlled mechanical properties. The mechanicalproperties within the implant can also be varied according to theanatomical region of implantation. With the predicted mechanicalproperties, masticatory and other functions of the mandible can betested and implants with better longevity can be fabricated. Thisstrategy can be applied to the design of custom implants for otherparts of the body as well. Patients requiring mandibular recon-struction would have better esthetics and functions with this newinnovative design of implants.

Fig. 19 Maximum displacement for effective elasticmodulus—3 GPa for Model 3

Fig. 20 CAD design of porous mandible for 3D printing

Fig. 21 3DP model of porous mandible

Transactions of the ASME

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

A

sMp

R

pa

J

Downlo

cknowledgmentWe would like to thank the Office of OU Research for financial

upport. In addition, we gratefully acknowledge the assistance ofedical Modeling LLC, Golden, CO for the fabrication of the

orous titanium blocks using the electron beam melting process.

eferences�1� Zhou, L. B., Shang, H. T., Hua, M., Li, D., Sigare, S., Chen, B. L., Liu, Y. P.,

and Zhao, J. L., 2008, “Reconstruction Of Curved Mandibular Angle DefectsUsing a New Internal Transport Distraction Device: An Experiment in Goats,”Br. J. Oral Maxillofac Surg., 46, pp. 445–448.

�2� St John, T. A., Vaccaro, A. R., Sah, A. P., Schaefer, M., Berta, S. C., Albert, T.,and Hilibrand, A., 2003, “Physical and Monetary Costs Associated With Au-togenous Bone Graft Harvesting,” Am. J. Orthop., 32, pp. 18–23.

�3� Silber, J. S., Anderson, D. G., Daffner, S. D., Brislin, B. T., Leland, J. M.,Hilibrand, A. S., Vaccaro, A. R., and Albert, T. J., 2003, “Donor Site Morbid-ity After Anterior Iliac Crest Bone Harvest for Single-Level Anterior CervicalDiscectomy and Fusion,” Spine, 28, pp. 134–139.

�4� Shimko, D. A. and Nauman, E. A., 2007, “Development and Characterizationof a Porous Poly�Methyl Methacrylate� Scaffold With Controllable Modulusand Permeability,” J. Biomed. Mater. Res., Part B: Appl. Biomater., 80�2�, pp.360–369.

�5� Schlickewei, W. and Schlickewei, C., 2007, “The Use of Bone Substitutes inthe Treatment of Bone Defects—the Clinical View and History,” Macromol.Symp., 253, pp. 10–23.

�6� Martola, M., Lindqvist, C., Hänninen, H., and Al-Sukhun, J., 2007, “Fractureof Titanium Plates Used for Mandibular Reconstruction Following AblativeTumor Surgery,” J. Biomed. Mater. Res., Part B: Appl. Biomater., 80B�2�, pp.345–352.

�7� Lee, M.-Y., Chang, C.-C., Lin, C.-C., Lo, L.-J., and Chen, Y.-R., 2002, “Cus-tom Implant Design for Patients With Cranial Defects,” IEEE Eng. Med. Biol.Mag., 21�2�, pp. 38–44.

�8� D’Urso, P. S., Earwaker, W. J., Barker, T. M., Redmond, M. J., Thompson, R.G., Effeney, D. J., and Tomlinson, F. H., 2000, “Custom Cranioplasty UsingStereolithography and Acrylic,” Br. J. Plast. Surg., 53, pp. 200–204.

�9� Connell, H. M., Statham, P. F., Collie, D. A., Walker, F. S., and Moos, K. F.,1999, “Use of a Template for Custom Cranioplasty,” Phidias—EC FundedNetwork Project on Rapid Prototyping in Medicine, 2, pp. 7–8.

�10� Samman, N., Luk, W. K., Chow, T. W., Cheung, L. K., Tideman, H., andClark, R. K. F., 1999, “Custom Made Titanium Mandibular ReconstructionTray,” Dent. J., 44�3�, pp. 195–199.

�11� Yaxiong, L., Dichen, L., Bingheng, H. S., and Li, G., 2003, “The CustomizedMandible Substitute Based on Rapid Prototyping,” Rapid Prototyping J., 9�3�,pp. 167–174.

�12� Singare, S., Dichen, L., Bingheng, L., Yanpu, L., Zhenyu, G., and Yaxiong, L.,2004, “Design and Fabrication of Custom Mandible Titanium Tray Based onRapid Prototyping,” Med. Eng. Phys., 26, pp. 671–676.

�13� Singare, S., Yaxiong, L., Dichen, L., Bingheng, L., Sanhu, H., and Gang, L.,2006, “Fabrication of Customized Maxillo-Facial Prosthesis Using Computer-Aided Design and Rapid Prototyping Techniques,” Rapid Prototyping J.,12�4�, pp. 206–213.

�14� Robertson, D. M., Pierre, L., and Chahal, R., 1976, “Preliminary Observationsof Bone Ingrowth Into Porous Materials,” J. Biomed. Mater. Res., 10, pp.335–344.

Fig. 22 Porous titanium

�15� Ryan, G., Pandit, A., and Apatsidis, D. P., 2006, “Fabrication Methods of

ournal of Medical Devices

aded 16 Sep 2009 to 129.15.14.53. Redistribution subject to ASME

Porous Metals for Use in Orthopedic Applications,” Biomaterials, 27, pp.2651–2670.

�16� Harryson, O., and Cormier, D. R., 2006, “Direct Fabrication of Custom Or-thopedic Implants Using Electron Beam Melting Technology,” AdvancedManufacturing Technology for Medical Applications, I. Gibson, ed., pp. 191–206.

�17� Starly, B., Fang, Z., Wei, S., Shokoufandeh, A., and Regli, W., 2005, “ThreeDimensional Reconstruction for Medical CAD Modelling,” Computer AidedDesign and Applications, 2�1–4�, pp. 431–438.

�18� Knoll, W. D., Gaida, A., and Maurer, P., 2006, “Analysis of Mechanical Stressin Reconstruction Plates for Bridging Mandibular Angle Defects,” J. CraniofacSurg., 34, pp. 201–209.

�19� Tie, Y., Wang, D. M., Ji, T., Wang, C. T., and Zhang, C. P., 2006, “ThreeDimensional Finite Element Analysis Investigating the Biomechanical Effectsof Human Mandibular Reconstruction With Autogenous Bone Grafts,” J.Craniomaxillofac Surg., 34, pp. 290–298.

�20� Hollister, S. J., Lin, C. Y., Satio, E., Schek, R. D., Taboas, J. M., Willams, J.M., Partee, B., Flanagan, C. L., Diggs, A., Wilke, E. N., Van Lenthe, G. H.,Miller, R., Wirtz, T., Das, S., Feinberg, S. E., and Krebsbsbch, P. H., 2005,“Engineering Craniofacial Scaffolds,” Orthod. Craniofac. Res., 8, pp. 162–173.

�21� Starly, B., Lau, W., Bradbury, T., and Sun, W., 2006, “Internal ArchitectureDesign Methodology for Tissue Replacement Structures,” CAD, 38�2�, pp.115–124.

�22� Fang, Z., Starly, B., and Sun, W., 2005, “Computer-Aided Characterization forEffective Mechanical Properties of Porous Tissue Scaffolds,” Comput.-AidedDes., 37, pp. 65–72.

�23� Lian, Q., Li, D.-C., Tang, Y.-P., and Zhang, Y.-R., 2006, “Computer AidedCharacterization of Effective Mechanical Properties of Porous Tissue Scaf-folds,” Comput.-Aided Des., 38�5�, pp. 507–514.

�24� Wang, D., Wang, C. T., Zhang, X., and Xu, L., 2005, “Design and Biome-chanical Evaluation of a Custom Lateral Mandible Titanium Prosthesis,” 27thAnnual International Conference of the Engineering in Medicine and BiologySociety, 2005, IEEE-EMBS, Vol. 6, pp. 6188–6191.

�25� Cox, T., Kohn, M. W., and Impelluso, T., 2006, “Computerized Analysis ofResorbable Polymer Plates and Screws for the Rigid Fixation of MandibularAngle Fractures,” Int. J. Oral Maxillofac Surg., 61, pp. 481–487.

�26� Fernandez, J. R., Gallas, M., Burguera, M., and Viano, J. M., 2003, “A Three-Dimensional Numerical Simulation of Mandible Fracture Reduction WithScrewed Miniplates,” J. Biomech., 36, pp. 329–337.

�27� Lovald, S. T., Khraishi, T., Wagner, J., Baack, B., Kelly, J., and Wood, J.,2006, “Comparison of Plate-Screw Systems Used in Mandibular Fracture Re-duction: Finite Element Analysis,” ASME J. Biomech. Eng., 128, pp. 654–656.

�28� Lopez-Heredia, M. A., Goyenvalle, E., Aguado, E., Pilet, P., Leroux, C., Dor-get, M., Weiss, P., and Layrolle, P., 2008, “Bone Growth in Rapid PrototypedPorous Titanium Implants,” J. Biomed. Mater. Res., Part B: Appl. Biomater.,85A�3�, pp. 664–673.

�29� An, Y. H., 2000, Mechanical Properties of Bone, Mechanical Testing of Boneand the Bone-Implant Interface, CRC, Boca Raton, FL, pp. 41–59.

�30� Maurer, P., Holweg, S., and Schubert, J., 1999, “Finite-Element-Analysis ofDifferent Screw-Diameters in the Sagittal Split Osteotomy of the Mandible,” J.Craniomaxillofac Surg., 27, pp. 365–372.

�31� Szmukler-Moncler, S., Salama, H., Reingewirtz, Y., and Dubruille, J. H., 1998,“Timing of Loading and Effect of Micromotion on Bone-Dental Implant In-terface: Review of Experimental Literature,” J. Biomed. Mater. Res., 43�2�,

rts fabricated with EBM

pp. 192–203.

SEPTEMBER 2009, Vol. 3 / 031007-9

license or copyright; see http://www.asme.org/terms/Terms_Use.cfm


Recommended