Software Framework for the Creation and Application of ...

Research ArticleSoftware Framework for the Creation and Application ofPersonalized Bone and Plate Implant Geometrical Models

Nikola Vitkovic Srdan Mladenovic Milan Trifunovic Milan ZdravkovicMiodrag Manic Miroslav Trajanovic Dragan Misic and Jelena Mitic

Faculty of Mechanical Engineering University of Nis Nis Serbia

Correspondence should be addressed to Nikola Vitkovic nikolavitkovicmasfakniacrs

Received 25 June 2018 Revised 30 August 2018 Accepted 12 September 2018 Published 10 October 2018

Academic Editor Emiliano Schena

Copyright copy 2018 Nikola Vitkovic et al is is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Computer-Assisted Orthopaedic Surgery (CAOS) defines a set of techniques that use computers and other devices for planningguiding and performing surgical interventions e important components of CAOS are accurate geometrical models of humanbones and plate implants which can be used in preoperational planning or for surgical guiding during an intervention Softwareframework which is introduced in this study is based on the Model-View-Controller (MVC) architectural pattern and it uses 3Dmodels of bones and plate implants developed by the application of the Method of Anatomical Features (MAF) e presentedframework may be used for preoperative planning processes and for the production of personalized plate implants e main ideaof the research was to develop a novel integrated software framework which will provide improved personalized healthcare to thepatient and at the same time provide the surgeon with more control over the patientrsquos treatment and recovery

1 Introduction

Human skeletal system is often affected by some patho-logical processes injuries and fractures In many cases it isnecessary to perform a surgical intervention sometimesquite urgently For the surgical intervention to be successfulit is particularly important to prepare a good preoperativeplan and to ensure that adequate implants are providedComputer-Assisted Orthopaedic Surgery (CAOS) [1] de-fines a set of techniques that use computers and other de-vices for planning guiding and performing surgicalinterventions and it can be developed for various platforms[2] e important components of CAOS are accurategeometrical models of the affected human bone and per-sonalized 3D models of plate implants which can be used inpreoperational planning or for surgical guiding during theintervention [3] A higher geometrical and anatomical ac-curacy of the applied geometrical models can improve pre-intra- and postoperative procedures and as a consequencepatientrsquos recovery In general geometrical models of humanbones can be created by the application of two different

approaches [4ndash8] e first approach implies the applicationof medical imaging technologies like computed tomography(CT) in order to provide 3D geometrical models of humanorgans Such models can be created in three general ways bythe application of specialized software which is a part ofa medical scanner (eg Vitrea) postprocessing of medicalimages in medical-oriented CAD programs (eg MaterialiseMimics) postprocessing in one of the CAD softwarepackages (eg CATIA) One of the main drawbacks of thisapproach is the inability to create a model of a whole boneat is the case when the bone scan is incomplete due toillness (osteoporosis arthritis cancer etc) or trauma(multiple fractures crushed bones etc) or when the qualityof medical images is not good enough e second approachfor creation of 3D geometrical models of bones or bonesegments is based on the predictive geometrical or statisticalmodel of bones and data obtained from medical images[457ndash10] For example these models can be created by theapplication of statistical methods or FFD (free-form de-formation) techniques Geometric entities of predictivemodels are described by mathematical functions whose

HindawiJournal of Healthcare EngineeringVolume 2018 Article ID 6025935 11 pageshttpsdoiorg10115520186025935

arguments are morphometric parameters that can be readfrommedical images By using this approach it is possible tocreate a 3D geometrical model which corresponds to thepatientrsquos bone ie personalized models Possible disad-vantages of this approach are the inability to read all pa-rameters (for the same reasons as in the first approach)insufficient number of parameters involved in predictivefunctions as well as inadequately defined parameters

One of the possible applications of geometrical models ofhuman bones is for the internal fixation Internal fixation isa surgical technique in which implants (plates rods pins etc)are placed inside of the human body For bone fixation andreduction plate implants are widely used ere are severaltypes of these implants but most common are dynamiccompression plates (DCP) and locking compression plates(LCP) DCP [11 12] improved healing processes by eliminatingthe need for external immobilization and providing morestability to the fixation assembly In order to fulfil theirfunction DCPs have to be mounted onto the periosteum (thetissue that surrounds bone surface) and should be pressed ontothe bone to achieve stability [12ndash15] As a consequence bloodsupply to the bone can be interrupted and cortical bone porosiscan appear at the site of implant placement A new plate designthe limited contact-dynamic compression plate (LC-DCP) [12]was introduced in order to reduce cortical necrosis In todayrsquospractice locking compression plates (LCP) are most common[14] ese plates provide locking (locking screws) and non-locking function which means that they can combine prop-erties of both the mentioned plates (DCP and LC-DCP) eprecontouring of plate shape ie the personalization of plateshape is a common operation which can be done before orduring the surgical interventione need for precontouring isgenerally determined by a type of plate and bone disease Forexample reconstructive plates [16] require precontouring inorder to achieve mechanical stability but for LCP with lockingscrews it is not required because locking screws provide sta-bility to the fixation system Regardless of the precontouringrequirement the possibility to adjust a plate shape is a goodsoftware feature because it gives the surgeonmore flexibility inplanning the orthopaedic intervention and the research pre-sented in this paper provides that functionality

Software platforms used for preoperative planning andguiding in orthopaedics have been developed for manyyears and there are many different solutions ere arecommercial solutions like Materialise Mimics and Otho-View (httpswwwmaterialisecomenmedical) or Osirix(httpswwwosirix-viewercom) [17] which uses medicalimages in appropriate format (DICOM) to plan a surgicalintervention or to analyse the patientrsquos health state ere isa possibility for surgeons to use software which is distributedwith the scanning systems like Vitrea (httpwwwvitalimagescomvitrea-visionvitrea-advanced-visualization) or Vesalius3D (httpswwwvesalius-3dcom) which enable direct processing of the 2D scannedimages Open-source or free solutions are also used likeRadiAnt (httpswwwradiantviewercom) or MicroDicom(httpwwwmicrodicomcom) which can be used for thereconstruction of DICOM images and creation of a 3Dmodel of the scanned tissue All of this software is based on

the direct implementation of medical images obtained frommedical scanning devices [18 19] In these cases whenscanned data of a human organ are incomplete the properreconstruction of their missing part or the creation ofa personalized plate implant model is very complex andtherefore the accuracy is questionable [4 6 20]

In order to improve pre- intra- and postoperativeprocedures a new software framework based on Model-View-Controller (MVC) architectural pattern [21] was de-veloped e software framework is named PersonalizedOrthopaedic Model-View-Controller (POMVC) frame-work e POMVC framework represents integrated soft-ware environment which uses and improves Method ofAnatomical Features (MAF) [5 7 8] procedures and al-gorithms for the creation of personalized 3D models ofhuman bones and plate implants Integrated means that theframework includes individual software components usedfor the creation of 3D models and connects them in onefunctional platform By the authorsrsquo knowledge there is nosimilar kind of framework currently applied for the creationof bone and plate models or for preoperative planningPossible benefits of POMVC framework application can becreation of complete bone models andor plate implantseven in the cases when the input bone data is incompleteshorter time of surgical intervention due to better pre-operative planning manufacturing of personalized platemodel by the application of additive manufacturing or high-speed machining generation of bone and plate model da-tabases which can be used for further research and others

e paper is structured as follows the first section of thepaper describes bone samples used for the research In thenext section of the paper descriptions of the MAF and plate3D models are presented ereafter the POMVC frame-work and its components are described in detail In the lastsection of the paper the application of the POMVCframework for a real clinical case is presented

2 Materials and Methods

In order to demonstrate the methods and the POMVCframework human tibia and mandible samples were usede samples were scanned by 64-slice CT (MSCT)(Aquillion 64 Toshiba Japan) with a resolution of 512 times

512 px and a slice thickness of 05mm e same sampleswere used as the ones presented in [7 8]

21 Research Concept In this section of the paper firsta short introduction to the methods developed by the au-thors of the research which represent the foundation for theresearch presented in this study is provided Next a de-scription of the software framework its components andapplication in real medical case is shown

211 Method of Anatomical Features (MAF) MAF isa method developed by the authors of this research and itintroduces a new approach for the creation of geometricalmodels of human bones e main objective of MAF ap-plication is to provide complete 3D geometrical models of

2 Journal of Healthcare Engineering

human bones and bone fragments even in the cases wherethe input data of a patientrsquos bone are not complete due to thebone illness fracture or some other trauma In general twodifferent types of output models can be created by MAF

(i) 3D geometrical models these models are standardpolygonal surface and volume models which havebeen used in CAD for many years ey are createdby the application of standard CAD technical fea-tures in CAD software packages on the input modelscreated by medical imaging methods e procedurefor creating such models consists of several pro-cesses CT scanning of the patient segmentation ofmedical data in medical software (eg Mimics)transformation of geometrical data in medicalsoftware to an appropriate format (eg STL) forCAD importing 3D model into CAD softwaremodel manipulation and transformation in CADsoftware and creation of an adequate 3D model(polygonal surface solid etc) e procedure isdescribed in more detail in [5 7]

(ii) Predictive (parametric) models of human bonesthese models are statistical models formed over theinput set of bone samples ey are defined as pointcloud models which represent an approximation ofthe human bone boundary surface Each point co-ordinate (X Y and Z) in the point cloud is defined byan individual parametric function Currently mul-tilinear regression is applied for the creation ofparametric functions and the used algorithm ispresented in Equation (1) and defined in [24]

X ones size d1( 1113857( 1113857d1d2d3d41113858 1113859

A XprimelowastX

K inv(A)

B A

XprimelowastXcoord

M XlowastB

(1)

where Xcoord is the vector of X coordinates defined for theinput set cloud of point di are the morphometric parameters(four) B is coefficient vector and M is the vector of thecalculated values (multilinear regression)

Parametric function is a function whose arguments aremorphometric parameters Morphometric parameters aredimensions which can be acquired (measured) frommedicalimages by using adequate software like Materialise Mimics3D doctor (httpwwwableswcom3d-doctor) or GIMP(httpswwwgimporg) By the application of individuallydefined values of morphometric parameters these modelscan be customized to the geometry and morphology ofa human bone of a specific patient Currently the process ofacquiring parameter data is done externally by using medicalor other image processing software but the intention is to

create in-house software solution which will be integratedinto the POMVC After the point cloud personalization themodel can be further processed in any CAD application andother types of 3D geometrical models can be created [5 7 8]

It is important to note that parametric models are used inclinical cases when 3D data of human bone are missing(eg due to tumour osteoporosis and complex fractures)and so a complete 3D model cannot be created By usingmedical software (eg Mimics and 3D Doctor) or X-rayimages a surgeon can acquire values of measurablemorphometric parameters by using standard techniques[1ndash3] and apply them in parametric functions by using thePOMVC framework As the result of the process a completepoint cloud model of the bone is created despite the lack ofinput data Further description of the process for the cre-ation and application of parametric models can be found in[5 7 8]

212 Personalized Plate Implants Personalized plates arebased on three developed 3D models of the plates presentedin [22 23] ese models are as follows

(i) Solid model of the personalized fixator is modelis a classical CAD model which can be used for thefixator manufacturing by the application of classicalmanufacturing or by the application of additivemanufacturing processes

(ii) Optimal parametric model (OPM) which is basedon the median bone geometry is is a solid modelwhose dimensions can be changed according to thegeometry of a specific bone e median bone ge-ometry is defined for the input set of bone sampleswhich defines a group of people for whom thefixator is applicable is means that the fixator canbe used for the patient whose bone geometriccharacteristics fall within the group

(iii) Parametric points model (PPM) which is based onthe parametric surface model of the human bonedeveloped by the application of MAF e contactsurface between the fixator and the bone is definedby the points whose coordinatesrsquo values are definedas parametric functions [5 7 8]

It is important to note that OPM and PPM models aredeveloped by the authors of this research and described in[22] e geometry of this fixator models can be adjusted insuch a way that they can be used as DCP or LCP fixatorseshape is already precontoured and the holes for screws canbe adjusted to achieve screw locking

213 Description of the Software Framework e developedsoftware framework is based on MVC architectural patternand it can be applied for the creation of personalized 3Dbone and plate models and for preoperative planning insurgery e POMVC framework is composed of threemain components Model View and Controller which arestandard entities of any MVC software architecture [21]

Journal of Healthcare Engineering 3

e role separation is not strict yet each component canhave multiple roles which is influenced by the requestedoutcomes from the software and direct implementatione POMVC framework represents integrated softwareenvironment which uses and improves Method of Ana-tomical Features (MAF) [5 7 8] e integrated frame-work consists of several software components which areconnected into one functional platform In the currentversion of the POMVC framework included softwarecomponents are MS Excel MATLAB and CAD software(CATIA andor Solid Works) All components are in-cluded as commercial or trial versions with adequatelicenses ere are several reasons why this framework wasdeveloped in this way and the most important reasons areas follows

(i) e presented MVC architecture is very scalableeasily maintainable and upgradable [21] e sys-temrsquos components can be replaced by using softwarewhich is appropriate for the selected institutionepossibility to replace one of the components withsome other open-source or in-house-developedsolution adds great opportunity for further re-search and customization

(ii) e systemrsquos open architecture enables applicationof different algorithms for 3D model creation ismeans that meshing of personalized cloud pointmodels can be done by using a selected algorithmlike Delaunay triangulations in custom-made ap-plications or by using specialized CAD softwareCurrently meshing is done in commercial CADcomponent of the system (CATIA) but in the nextversions of the POMVC framework open-sourcesolutions with enabled scripting will be used likeBlender (httpswwwblenderorg) or FreeCAD(httpswwwfreecadweborg)

(iii) Algorithms for the creation of parametric 3D bonemodels originally created in MAF are improvedUntil the development of the POMVC frameworkpersonalized bone 3D models were created by theapplication of parametric functions with strictlydefined number of morphometric parameters[5 7 8] POMVC enables application of parametricfunctions with variable number of morphometricparameters is means that if a surgeon cannotacquire values of all defined morphometric pa-rameters which is the case with the missing bonedata then he can provide only morphometric valueswhich can be acquired from medical images or hecan provide a minimal number of parameters whichare adequate for the specific case Based on the inputdata POMVC will apply an adequate algorithm forthe creation of personalized point cloud e resultmay be a less accurate 3D model yet it will bea complete 3D personalized model of human bone

(iv) Until the development of POMVC plate and boneparametric models were created and used directly inCAD software with strictly defined procedures

[5 7 8] is means that a surgeon or a di-agnostician needs to know specialized softwarefeatures in order to create or manipulate 3D modelsor to engage a designer who will provide support Byintroducing POMVC application of parametricmodels is much simpler because interface for en-tering morphometric data and parametric modelmanipulation can be adjusted to the surgeon needsOf course the presence of a designer is still rec-ommended but not mandatory It must be notedthat the POMVC framework still enables the ap-plication of individual components for specificcases as it is in the clinical case that will be pre-sented in the later section of the paper

(v) Bidirectional connection between componentsincluded in the POMVC framework is establishedis means that bone and plate models created inthe framework can be used for improving para-metric models by enlarging the input set (CAD toExcel) forming the database of bone and platemodels (CAD to database) which can be used forfurther statistical or other AI processing etc

It is important to define howMAF is integrated with thePOMVC framework MAF is a method with many de-veloped procedures for the creation of personalized 3Dmodels of human bones and plate implants It containsalgorithms (1) procedures rules and other elements whichmust be implemented in order to fulfil specific goals emain question is how these elements are integrated orrepresented in the POMVC framework e constructionof personalized 3D geometrical models based on completebone data requires application of just one component of theframework and that is CAD software e input for thisprocedure is a 3D model created in medical imagingsoftware eg Mimics Application of parametric bonemodel requires a different approach First the algorithmfor the creation of a personalized point cloud model ofa specific bone or bone part must be stored in some kind ofa database system e POMVC uses textual files and re-lational database Each algorithm is defined by Model(eg Module in VBA or m file in MATLAB) and based onthe user input the controller component decides whatmodel will call ie perform business logic As alreadystated algorithms cannot work without data (point datamorphometric parameters and plate model parameters) soit is important to enable input of data which will be for-warded to Model component is input is done throughview Component and its UI (eg Excel cells or text fieldsCAD UDF (user-defined feature) or plain parameters)Based on the requirements and input data the frameworkcreates adequate bone and plate models or enables thesurgeon andor designers to do additional modellingprocesses

In order to provide visual representation of the softwarecomponents and their mutual influence UML componentdiagram is created and presented in Figure 1 e de-scription of individual components included in the currentversion of the POMVC framework is as follows


(1) MS Excel (Controller and Model) is component is anentry point for the POMVCis component contains inputfields for entering point data values of morphometric pa-rameters acquired from medical images and for insertingplate parameter values for the plate personalization Itcontains VBA macros which do calculations call othercomponents and forward required data to them esemacros are as follows

(i) Macro which contains the algorithm for the cre-ation of parametric model of a specific bone basedon the input bone samples and measured pointdata As it has already been stated a parametricbone model is created based on the input bonesamples and the applied multilinear regressione multilinear regression algorithm (1) is storedin Excel Module and in external ldquomrdquo textual fileis macro can create parametric functions basedon a specific number of input morphometric pa-rameters ie it can create ldquovariable parametricmodelsrdquo In order to create a specific parametricmodel the textual file is transferred to MATLABby using VBA subroutines and OLE (ObjectLinking and Embedding) technology and calcu-lations are performed externally After the cal-culations new point data (X Y and Z) arereturned to Excel and placed in an adequate tablecell Created parametric functions are stored inexternal textual files for possible later applicationslike further research or statistical processing After

the completion of macro the user has an insightinto each point coordinates for every point ina personalized point cloud because they arepresented in tabular format in Excel

(ii) e second macro is used for transferring point datato CAD software is is done by calling CADsoftware through OLE technology and applyingadequate functions Transferred points are used asthe basis for the calculations of interpolated splinecurves in CAD Software and for the creation ofpersonalized bone surface and solid models If thereis a requirement further model processing can bedone directly in CAD software

ere is another specific function which Excel providesand that is the ability to store parametersrsquo values defined forCAD parametric plate model ie it represents design tablefor the specific plate model is is an excellent functionwhich can be used for a quick creation of personalized platemodel for the specific patient when a 3D model of the boneis available

(2) MatLab (Model) is component is used for all cal-culations In general it can be considered as a Modelandor helper component As already stated multiplelinear regression is currently used as a statistical methodfor tibia and mandible but other statistical and AImethods can also be applied is component uses pointand morphometric parameter data supplied by Excel toperform calculations on the basis of a defined algorithm

MATLAB MATLAB is used forthe calculations Itreceives data fromExcel and returns

results (parametricfunctions) to the

Excel

CAD soware

CAD soware (SolidWorks or CATIA)

is used for the drawingof bone and plate 3D

models Point data areacquired from Excel

MS ExcelExcel VBA Macros are

created and used togetherwith the bone data to createparametric functions and

to transfer point data to theCAD soware

3D personalized bone model

Data Model

3D personalized plate model+formed on

Parametric functions

+created on

Parametric plate model

For the bonesinput set

Model

View

Controllerlaquodeployraquo

laquodeployraquo

laquodeployraquo

laquodeployraquo

Figure 1 UML component diagram of the POMVC architecture


(1) After the calculations all results are returned toExcel

Currently for the mandible bone neural network is alsoapplied for the creation of a predictive model An algorithmthat covers this AI method will be included in the nextrelease of the framework

(3) CAD Software (Model and View) CAD software is mostlyused as View component of the POMVC Point data fora specific bone are forwarded fromMS Excel to CAD softwareand a geometrical model is formed by the use of specifictechnical features (eg Loft surface technical feature of SolidWorks or multisection surface of CATIA) ere is one ad-ditional function of this component which is important isfunction enables the creation of a personalized plate geo-metrical model based on the previously defined parametricsolid model for a specific plate In this stage of the POMVCthe creation of plate models is performed in CAD softwarebased on the values which are manually acquired from thecreated bone model [22] ese values are implemented asparameter values in the UDF (user-defined feature) manuallyor by using a design table from Excel If both the plate andbone models are created then the CAD model can be used toform an assembly and to define possible manufacturingprocedures through conventional manufacturing or by usingadditive technologies (3D printing)

214 Process Description A surgeon is the main user of thesystem His role is to make important decisions and toperform surgical planning e designer is a supporting userof the system His role is to provide technical support for thesurgeon and to improve and correct software systemaccording to the surgeonrsquos recommendations In order tobetter describe main processes of the proposed softwareframework a flowchart diagram of the complete process forthe creation of personalized plate for the specific bone andfracture is presented in Figure 2 e initial processes whichare not included in the diagram are so-called preplanningprocesses and they include patient admittance to the hos-pital and diagnostics of bone disease Medical imaging is animportant diagnostic process and its output is the input forthe first process in the system which is ldquoreading ofmorphometric parametersrdquo In that process the surgeonacquires values of morphometric parameters from medicalimages such are X-rays or CT in adequate software or withother tools (etalons measuring equipment etc) e ac-quired data are inserted in Excel tables as already describedand calculations are performed e surgeon only needs toenter the data and all other tasks are done automaticallyCalculated values are presented to the surgeon (Figure 3)and he only needs to press another button and start thesecond macro which will trigger ldquoExporting data to CADsoftwarerdquo and ldquoCreation of personalized Bone 3D Modelrdquoprocesses

e surgeon and the designer can analyse the createdmodel presented in Figure 4 and if there is a requirementfor the bone model improvement it can be done in CADsoftware When the surgeon is satisfied with the created

personalized 3D model of a specific bone the next step isplate model selection e surgeon can choose a plate modelfrom the database of created plate parametric 3D models orit can create a customized plate model based on theboundary surface of the bone [22] In the current stage of theresearch four parametric models of standard plates arecreated three for long bones and one reconstructive plate forthe mandible bone but more will be developed After theselection of a plate parametric values are acquired from thebone model by the use of CAD technical features and theplate model is modified accordingly Parameters are defineddifferently for specific plates and for the tibia bone and plateimplant developed by Mitkovic process of parameter ac-quisition is described in [22] If there is a requirement for themodel correction (eg parameter values are not measuredright the surgeon needs additional bending of the platemodel etc) it is done in CAD software by the designerWhen both the designer and the surgeon are satisfied withthe 3D bone and plate models final assembly is created aspresented in Figures 5(a) and 5(b) In the last step of theoverall process (Finishing processes) 3D models are storedin a database for further research and application

Considering the time required for the creation of 3Dmodels we can state that they can be created in a relativelyshort time considering the described processes When thereare unacceptable deviations in models or the surgeon needsadditional modifications it can take a longer period of timeto create the models depending on the complexity of theproblem Also if there is a requirement for modelmanufacturing then the time for production depends onmodelsrsquo characteristics (geometry shape material etc) andon the applied manufacturing technology

215 POMVC Application in Real Surgical Case

(1) Clinical Case Description e patient was a young (18years) male with progenia e main surgeonrsquos requirementwas to create personalized reconstructive plates which willprovide fixation of mandible parts after surgery e secondimportant requirement was to avoid CT scanning for ac-quiring geometrical and morphometric data For the pur-pose of fulfilling these requirements the POMVC wasapplied As already stated the POMVC uses or createsgeometric models of human bones and plates and the onlycondition is to provide adequate input data (morphometricparameters bone models etc) to the framework

It is important to mention that the POMVC was used fora 3D mandible model creation in individual CAD compo-nent of the framework and for the personalized plate 3Dmodel creation (Excel and CAD components were used) Ashort description of the mandible and a personalized plate3Dmodel creation process is presented in the next section ofthe paper

(2) Acquiring Input Data e patient was scanned by usingan X-ray scanner (Figure 6(a)) and with Sirona SL Ortho-phos 3D device (Figure 6(b)) e X-ray scanning wasperformed with etalon included in order to properly scale


measured values X-ray and Sirona SL scanners were usedtogether because Sirona SL field of view (FOV) size is lessthan required for the creation of a complete 3D model ofmandible bonee scanning FOV is 11 times 10 cm and becauseof that in the 3D medical image of the patient both condylarprocesses were missing as is presented in Figure 6(b) elack of condylar process made it difficult to determine theposition of the rotation axis of the temporomandibular jointon the 3Dmodele rotation axis of the temporomandibularjoint is significant because the mandible moves around it androtates until the good occlusion is achieved is is importantbecause a proper plate shape can be created only when themandible is properly positioned Only then a designer anda surgeon can choose valid points on the surface of themandible model To enable proper geometrical definition ofthe rotation axis an X-ray image was used to acquire themissing geometrical and morphometric data

In accordance with the anatomical and morphologicalcharacteristics of the mandible two anatomical referencepoints are defined on the X-ray image in GIMP softwarementon (ME) and gonion (GO) [8] Menton is the lowestpoint on the mandibular symphysis and gonion is the mostinferior point of the mandibular angle e horizontal(mandibular) line was obtained by connecting two ana-tomical points e next step was to determine the positionof the occlusion line e occlusion line is an imaginary linethat theoretically touches the incisal edges of the incisors andthe tips of the occluding Surgeons determined the position

of the required points and elements on the 2D X-ray imagethe point of rotational axis ME GO incisal edges andoccluding tips Next lines which are perpendicular to theocclusion and horizontal line were created ese perpen-dicular lines are going through the point of rotational axise distances between the cross section of perpendicularlines with occlusion and horizontal line and selected pointson the mandible were measured and scaled according to theetalon size ese distances were used to determine theposition of the rotational axis in 3D ereafter the positionof the rotation axis of the temporomandibular joint (point)on the 2D medical image is determined as presented inFigure 6(a) e same procedure was performed in 3D onlyinstead of lines occlusion and horizontal planes were cre-ated To determine the proper position of the rotational axisdistances which were measured in the 2D image weretransferred to 3D ese distances provided positions ofperpendicular planes on both the horizontal and occlusionplane (same as for lines in 2D) By using the intersection ofplanes perpendicular to the occlusion andmandibular planethe rotational axis was determined

(3) 6e Creation of Plate Models In this specific case goodocclusion could be achieved by performing a cut on themandible bone and by repositioning the parts In co-operation with maxillofacial surgeons the line of the cut isdetermined on the surface of the mandible polygonalmodel e cut line is positioned in front of the seventh

Reading ofmorphometric

parameters

Inserting datainto Excel

Performcalculations

in Excel

Export data to CADsoware

Create 3Dpersonalizedbone model

Acquireparameters valuesfor defined plate

Create personalizedplate model

Analysis of platebone assembly

Validatemodel

Modelcorrection

Validateassembly

Finishingprocesses

Modelscorrection

Yes

No

Yes

No

Parametricfunctions

Valuesof

morphometricparameters

Morphometricparameters

valuesMedicalimages

End

Storedpersonalizedbone model

Databaseof created

modelsPersonalizedbone model

Personalizedplate model

Personalizedbone model

Database ofpersonalizedbone models

and plates

Start

Database ofparametric 3Dmodel of thedefined plates

Figure 2 Flowchart diagram of the process for the creation of the personalized bone and plate 3D models


tooth on both the left and the right side After cutting thepolygonal model along the cutting line on the left and theright side rotating and moving of the mandible is done inrelation to the rotation axis of the joint Having been

rotated and moved the mandible is brought into the ap-propriate position in which the aesthetically pleasingappearance of the face is achievedemost important goalwas achieved resulting in good occlusion

Figure 3 Point and morphometric parameter data presented in MS Excel I calculated point data for the specific bone II parametersdefined for the input set III parameters for the specific bone IV first macro and V second macro

Figure 4 Tibia bone surface model created in CAD software component (Solid Works) I interpolated spline curves II human tibiasurface model created on the basis of interpolated spline curves


Next the mandible polygonal model was processed andall anatomical points were transferred to Excel by usingscript in CATIA developed by the authors By using thisapproach a bidirectional connection was established be-tween two components of the POMVC Transferred ana-tomical points were used for enlarging the bone input set forthe parametric model creation and for the following creationof the personalized plate models

Following the maxillofacial surgeonrsquos propositionspecic anatomical points from the set of dened pointswere selected on the mandible parts (Figure 7(a)) e linewhich connects these points determined the position of theplate (Figure 7(a)) e only thing left is to determine theshape of the plate which conforms to the boundary surfaceof the mandible is was done by creating a surface modelof the plate contact surface by using anatomical pointsaround the dened line as presented in Figure 7(b) eplate model on the other side of the mandible was createdin the same way After the surgeons approved the surface

model of the plate the model was transformed to a solidmodel in order to prepare it for 3D printing (Figure 7(c))by adding thickness of 3mm Physical models of bothplates were printed on the CreatBot 3D printer and pre-sented in Figure 7(d) Surgeons used these models forprecontouring of reconstructive plates before surgeryIntervention was performed and surgeons were very sat-ised with the results ie plates were precontoured per-fectly More information about the patientrsquos recovery willbe available in the following period

3 Conclusion

Software framework presented in this research enablescreation of personalized models of bone and plate implantscustomized to the geometry morphology and anatomy ofthe specic patient Its scalable architecture and independentcomponents provide a lot of possibilities for further im-provement and adjustments e main intention of the

I

II

III

IV

V

(a)

I

II

III

IV

(b)

Figure 6 Assembly of personalized plate implant and bone 3Dmodels (a) X-ray image in LM planemdashpoint of rotational axis (I) occlusionline (II) mandibular line (III) perpendicular lines (IV) etalon (V) (b) rotational axis (I) occlusion plane (II) mandibular plane (III) andperpendicular planes (IV)

I

II

(a)

IIIIII

(b)

Figure 5 Assembly of personalized plate implant and bone 3D models (a) proximal tibia bone (I) and modied cloverleaf plate (II) (b)customized reconstructive plate (I) mandible body fracture (II) and mandible bone (III)


author of this research was to create a fully adaptablesoftware framework which can be used independently invarious institutions for solving various problems in or-thopaedics and possibly other branches of surgery Per-sonalized medicine and various software solutions thatsupport it are currently strongly applied in medical practiceand the presented POMVC brings additional possibilities tothe field e system can be used in medical educationclinical practice and in all other fields where there is a re-quirement to provide an integrated system for the simula-tion and preoperative planning of surgical interventions inorthopaedics Also manufacturing companies can use thePOMVC to create models of bones and plates by usingconventional or additive manufacturing

Data Availability

e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

e authors declare that there are no conflicts of interest

Acknowledgments

e paper is part of the project III41017mdashldquoVirtual humanosteoarticular system and its application in preclinical andclinical practicerdquo sponsored by the Republic of Serbia for theperiod of 2011ndash2017

References

[1] N Sugano ldquoComputer-assisted orthopaedic surgery androbotic surgery in total hip arthroplastyrdquo Clinics in Ortho-pedic Surgery vol 5 no 1 pp 1ndash9 2013

[2] Z Wang L Zhang L Ma and B Liu ldquoModeling medicalservices with mobile health applicationsrdquo Journal ofHealthcare Engineering vol 2018 Article ID 138503411 pages 2018

[3] S W Young M R Safran and M Clatworthy ldquoApplicationsof computer navigation in sports medicine knee surgery anevidence-based reviewrdquo Current Reviews in MusculoskeletalMedicine vol 6 no 2 pp 150ndash157 2013

[4] S Filippi B Motyl and C Bandera ldquoAnalysis of existingmethods for 3D modelling of femurs starting from two or-thogonal images and development of a script commercialsoftware packagerdquo Computer Methods and Programs inBiomedicine vol 89 no 1 pp 76ndash82 2008

[5] N Vitkovic J Milovanovic N Korunovic et al ldquoSoftwaresystem for creation of human femur customized polygonalmodelsrdquo Computer Science and Information Systems vol 10no 3 pp 1473ndash1497 2013

[6] Q Cheng P X Liu P Lai S Xu and Y Zou ldquoA novel hapticinteractive approach to simulation of surgery cutting based onmesh and meshless modelsrdquo Journal of Healthcare Engi-neering vol 2018 Article ID 9204949 16 pages 2018

[7] V Majstorovic M Trajanovic N Vitkovic and M StojkovicldquoReverse engineering of human bones by using method of an-atomical featuresrdquoCIRP Annals vol 62 no 1 pp 167ndash170 2013

[8] N Vitkovic J Mitic M Manic et al ldquoe parametric modelof the human mandible coronoid process created by methodof anatomical featuresrdquo Computational and Mathematical

II

I

(a)

I

II

(b)

I

(c) (d)

Figure 7 Creation of the personalized plates (a) construction of reference linemdashanatomical points (I) and reference line (II) (b) surfaceextractionmdashanatomical points (I) and contour lines (II) (c) solid model of the plate (I) (d) printed models of the plate implants


Methods in Medicine vol 2015 Article ID 574132 10 pages2015

[9] V Sholukha T Chapman P Salvia et al ldquoFemur shape pre-diction by multiple regression based on quadric surface fittingrdquoJournal of Biomechanics vol 44 no 4 pp 712ndash718 2011

[10] M Rashid K Husain N Vitkovic et al ldquoReverse modeling ofhuman humerus by the method of anatomical features(MAF)rdquo in Proceedings of the Seventh International WorkingConference Total Quality ManagementmdashAdvanced and In-telligent Approaches (TQM 2015) pp 197ndash202 BelgradeSerbia June 2015

[11] H K Uhthoff P Poitras and S D Backman ldquoInternal platefixation of fractures short history and recent developmentsrdquoJournal of Orthopaedic Science vol 11 no 2 pp 118ndash1262006

[12] Y C Lai Y W Tarng C J Hsu W N Chang S W Yangand J H Renn ldquoComparison of dynamic and locked com-pression plates for treating midshaft clavicle fracturesrdquo Or-thopedics vol 35 no 5 pp e697ndashe702 2012

[13] R Frigg M Wagner and A Frenk ldquoLocking compressionplates (LCP) amp less invasive stabilization system (LISS)rdquoEuropean Cells and Materials vol 16 no 5 p 5 2008

[14] M J Gardner D L Helfet and D G Lorich ldquoHas lockedplating completely replaced conventional platingrdquo AmericanJournal of Orthopedics (Belle Mead NJ) vol 33 no 9pp 440ndash446 2004

[15] C R Berkin and D V Marshall ldquoree-sided plate fixationfor fractures of the tibial and femoral shafts a follow-up noterdquoJournal of Bone and Joint Surgery vol 54 no 5 pp 1105ndash1113 1972

[16] AO Foundation ldquoInternal fixation and platesrdquo June 2018httpswwwaofoundationorg

[17] T Spiriev V Nakov L Laleva and C Tzekov ldquoOsiriXsoftware as a preoperative planning tool in cranial neuro-surgery a step-by-step guide for neurosurgical residentsrdquoSurgical Neurology International vol 8 no 1 p 241 2017

[18] J M Toto E I Chang R Agag K Devarajan S A Patel andN S Topham ldquoImproved operative efficiency of free fibulaflap mandible reconstruction with patient-specific computer-guided preoperative planningrdquoHead and Neck vol 37 no 11pp 1660ndash1664 2015

[19] J Parthasarathy ldquo3D modeling custom implants and itsfuture perspectives in craniofacial surgeryrdquo Annals of Max-illofacial Surgery vol 4 no 1 pp 9ndash18 2014

[20] Y Liu Y Fan X Jiang and D A Baur ldquoA customizedfixation plate with novel structure designed by topologicaloptimization for mandibular angle fracture based on finiteelement analysisrdquo BioMedical Engineering OnLine vol 16no 1 p 131 2017

[21] Y H Ding C H Liu and Y X Tang ldquoMVC pattern based onJAVArdquo Applied Mechanics and Materials vol 198-199pp 537ndash541 2012

[22] N Vitkovic M M Mitkovic M B Mitkovic et al ldquoReverseengineering of the mitkovic type internal fixator for lateraltibial plateaurdquo FactaUniversitatis Series Mechanical Engi-neering vol 13 no 3 pp 259ndash268 2015

[23] M Ristic M Manic D Misic M Kosanovic andM Mitkovic ldquoImplant material selection using expert sys-temrdquo FactaUniversitatis Series Mechanical Engineeringvol 15 no 1 pp 133ndash144 2017

[24] S Brown ldquoMultiple linear regression analysis a matrix ap-proach with MATLABrdquo Alabama Journal of Mathematicspp 1ndash4 2009 SpringFall httpajmonlineorg2009brownpdf


International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom

arguments are morphometric parameters that can be readfrommedical images By using this approach it is possible tocreate a 3D geometrical model which corresponds to thepatientrsquos bone ie personalized models Possible disad-vantages of this approach are the inability to read all pa-rameters (for the same reasons as in the first approach)insufficient number of parameters involved in predictivefunctions as well as inadequately defined parameters

One of the possible applications of geometrical models ofhuman bones is for the internal fixation Internal fixation isa surgical technique in which implants (plates rods pins etc)are placed inside of the human body For bone fixation andreduction plate implants are widely used ere are severaltypes of these implants but most common are dynamiccompression plates (DCP) and locking compression plates(LCP) DCP [11 12] improved healing processes by eliminatingthe need for external immobilization and providing morestability to the fixation assembly In order to fulfil theirfunction DCPs have to be mounted onto the periosteum (thetissue that surrounds bone surface) and should be pressed ontothe bone to achieve stability [12ndash15] As a consequence bloodsupply to the bone can be interrupted and cortical bone porosiscan appear at the site of implant placement A new plate designthe limited contact-dynamic compression plate (LC-DCP) [12]was introduced in order to reduce cortical necrosis In todayrsquospractice locking compression plates (LCP) are most common[14] ese plates provide locking (locking screws) and non-locking function which means that they can combine prop-erties of both the mentioned plates (DCP and LC-DCP) eprecontouring of plate shape ie the personalization of plateshape is a common operation which can be done before orduring the surgical interventione need for precontouring isgenerally determined by a type of plate and bone disease Forexample reconstructive plates [16] require precontouring inorder to achieve mechanical stability but for LCP with lockingscrews it is not required because locking screws provide sta-bility to the fixation system Regardless of the precontouringrequirement the possibility to adjust a plate shape is a goodsoftware feature because it gives the surgeonmore flexibility inplanning the orthopaedic intervention and the research pre-sented in this paper provides that functionality

Software platforms used for preoperative planning andguiding in orthopaedics have been developed for manyyears and there are many different solutions ere arecommercial solutions like Materialise Mimics and Otho-View (httpswwwmaterialisecomenmedical) or Osirix(httpswwwosirix-viewercom) [17] which uses medicalimages in appropriate format (DICOM) to plan a surgicalintervention or to analyse the patientrsquos health state ere isa possibility for surgeons to use software which is distributedwith the scanning systems like Vitrea (httpwwwvitalimagescomvitrea-visionvitrea-advanced-visualization) or Vesalius3D (httpswwwvesalius-3dcom) which enable direct processing of the 2D scannedimages Open-source or free solutions are also used likeRadiAnt (httpswwwradiantviewercom) or MicroDicom(httpwwwmicrodicomcom) which can be used for thereconstruction of DICOM images and creation of a 3Dmodel of the scanned tissue All of this software is based on

the direct implementation of medical images obtained frommedical scanning devices [18 19] In these cases whenscanned data of a human organ are incomplete the properreconstruction of their missing part or the creation ofa personalized plate implant model is very complex andtherefore the accuracy is questionable [4 6 20]

In order to improve pre- intra- and postoperativeprocedures a new software framework based on Model-View-Controller (MVC) architectural pattern [21] was de-veloped e software framework is named PersonalizedOrthopaedic Model-View-Controller (POMVC) frame-work e POMVC framework represents integrated soft-ware environment which uses and improves Method ofAnatomical Features (MAF) [5 7 8] procedures and al-gorithms for the creation of personalized 3D models ofhuman bones and plate implants Integrated means that theframework includes individual software components usedfor the creation of 3D models and connects them in onefunctional platform By the authorsrsquo knowledge there is nosimilar kind of framework currently applied for the creationof bone and plate models or for preoperative planningPossible benefits of POMVC framework application can becreation of complete bone models andor plate implantseven in the cases when the input bone data is incompleteshorter time of surgical intervention due to better pre-operative planning manufacturing of personalized platemodel by the application of additive manufacturing or high-speed machining generation of bone and plate model da-tabases which can be used for further research and others

e paper is structured as follows the first section of thepaper describes bone samples used for the research In thenext section of the paper descriptions of the MAF and plate3D models are presented ereafter the POMVC frame-work and its components are described in detail In the lastsection of the paper the application of the POMVCframework for a real clinical case is presented

2 Materials and Methods

In order to demonstrate the methods and the POMVCframework human tibia and mandible samples were usede samples were scanned by 64-slice CT (MSCT)(Aquillion 64 Toshiba Japan) with a resolution of 512 times

512 px and a slice thickness of 05mm e same sampleswere used as the ones presented in [7 8]

21 Research Concept In this section of the paper firsta short introduction to the methods developed by the au-thors of the research which represent the foundation for theresearch presented in this study is provided Next a de-scription of the software framework its components andapplication in real medical case is shown

211 Method of Anatomical Features (MAF) MAF isa method developed by the authors of this research and itintroduces a new approach for the creation of geometricalmodels of human bones e main objective of MAF ap-plication is to provide complete 3D geometrical models of





X ones size d1( 1113857( 1113857d1d2d3d41113858 1113859

A XprimelowastX

K inv(A)

B A

XprimelowastXcoord

M XlowastB

(1)






























Excel

CAD soware








Data Model



+created on



Model

View


laquodeployraquo

laquodeployraquo

laquodeployraquo




















parameters


Performcalculations

in Excel






Validatemodel

Modelcorrection

Validateassembly

Finishingprocesses

Modelscorrection

Yes

No

Yes

No

Parametricfunctions

Valuesof



valuesMedicalimages

End


Databaseof created





and plates

Start












3 Conclusion


I

II

III

IV

V

(a)

I

II

III

IV

(b)


I

II

(a)

IIIIII

(b)




Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia





X ones size d1( 1113857( 1113857d1d2d3d41113858 1113859

A XprimelowastX

K inv(A)

B A

XprimelowastXcoord

M XlowastB

(1)






























Excel

CAD soware








Data Model



+created on



Model

View


laquodeployraquo

laquodeployraquo

laquodeployraquo




















parameters


Performcalculations

in Excel






Validatemodel

Modelcorrection

Validateassembly

Finishingprocesses

Modelscorrection

Yes

No

Yes

No

Parametricfunctions

Valuesof



valuesMedicalimages

End


Databaseof created





and plates

Start












3 Conclusion


I

II

III

IV

V

(a)

I

II

III

IV

(b)


I

II

(a)

IIIIII

(b)




Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




















Excel

CAD soware








Data Model



+created on



Model

View


laquodeployraquo

laquodeployraquo

laquodeployraquo




















parameters


Performcalculations

in Excel






Validatemodel

Modelcorrection

Validateassembly

Finishingprocesses

Modelscorrection

Yes

No

Yes

No

Parametricfunctions

Valuesof



valuesMedicalimages

End


Databaseof created





and plates

Start












3 Conclusion


I

II

III

IV

V

(a)

I

II

III

IV

(b)


I

II

(a)

IIIIII

(b)




Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia










Excel

CAD soware








Data Model



+created on



Model

View


laquodeployraquo

laquodeployraquo

laquodeployraquo




















parameters


Performcalculations

in Excel






Validatemodel

Modelcorrection

Validateassembly

Finishingprocesses

Modelscorrection

Yes

No

Yes

No

Parametricfunctions

Valuesof



valuesMedicalimages

End


Databaseof created





and plates

Start












3 Conclusion


I

II

III

IV

V

(a)

I

II

III

IV

(b)


I

II

(a)

IIIIII

(b)




Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



















parameters


Performcalculations

in Excel






Validatemodel

Modelcorrection

Validateassembly

Finishingprocesses

Modelscorrection

Yes

No

Yes

No

Parametricfunctions

Valuesof



valuesMedicalimages

End


Databaseof created





and plates

Start












3 Conclusion


I

II

III

IV

V

(a)

I

II

III

IV

(b)


I

II

(a)

IIIIII

(b)




Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia







parameters


Performcalculations

in Excel






Validatemodel

Modelcorrection

Validateassembly

Finishingprocesses

Modelscorrection

Yes

No

Yes

No

Parametricfunctions

Valuesof



valuesMedicalimages

End


Databaseof created





and plates

Start












3 Conclusion


I

II

III

IV

V

(a)

I

II

III

IV

(b)


I

II

(a)

IIIIII

(b)




Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia










3 Conclusion


I

II

III

IV

V

(a)

I

II

III

IV

(b)


I

II

(a)

IIIIII

(b)




Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia





3 Conclusion


I

II

III

IV

V

(a)

I

II

III

IV

(b)


I

II

(a)

IIIIII

(b)




Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



Data Availability




Acknowledgments


References









II

I

(a)

I

II

(b)

I

(c) (d)























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia






















RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


Software Framework for the Creation and Application of ...

Documents