Mauren Abreu de Souza  Humberto Remigio Gamba  Helio Pedrini E ditors Multi- Modality Imaging Applications and Computational Techniques Multi-Modality Imaging Mauren Abreu de Souza Humberto Remigio Gamba • Helio Pedrini Editors Multi-Modality Imaging Applications and Computational Techniques 123 Editors MaurenAbreudeSouza HumbertoRemigioGamba GraduateProgramonHealth FederalUniversityofTechnology–Paraná Technology(PPGTS) (UTFPR) PontificalCatholicUniversityof Curitiba,Paraná,Brazil Paraná–PUCPR Curitiba,Paraná,Brazil HelioPedrini InstituteofComputing UniversityofCampinas Campinas,SP,Brazil ISBN978-3-319-98973-0 ISBN978-3-319-98974-7 (eBook) https://doi.org/10.1007/978-3-319-98974-7 LibraryofCongressControlNumber:2018960266 ©SpringerNatureSwitzerlandAG2018 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartof thematerialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation, broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionorinformation storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodology nowknownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthorsandtheeditorsaresafetoassumethattheadviceandinformationinthisbook arebelievedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsor theeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinorforany errorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregardtojurisdictional claimsinpublishedmapsandinstitutionalaffiliations. ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland Preface This book presents different approaches on multimodality imaging with a focus on biomedical applications. In terms of medical imaging, it is possible to divide into two categories: functional (related to physiological body measurements) and anatomical(structural)imagingmodalities. Itisworthmentioningthatthisbookcoverssomeimagingcombinationcoming from the usual popular modalities (such as the anatomical modalities, e.g. X-ray, CT and MRI); but it also includes some promising and new imaging modalities that are still being developed and improved (such as infrared thermography (IRT) andphotoplethysmography imaging(PPGI)),implyinginpotentialapproaches for innovativebiomedicalapplications. Moreover,itincludesavarietyoftoolsoncomputervision,imagingprocessing andcomputergraphics,whichledtothegenerationandvisualizationof3Dmodels, allowing the most recent advances in this area possible. This is an ideal book for students and biomedical engineering researchers covering the biomedical imaging field. The book covers a wide range of topics employing different multimodality imaging techniques for biomedical applications. The book is distributed in nine chapters,asfollows: Chapter1—InfraredThermography Regardingnon-invasiveandnon-contactimagingmodalities,infraredthermogra- phy(IRT)ispresentedinthefirstchapter.Suchmodalityisalsocalledinfrared(IR) imagingorthermalimaging.Themainapproachhereisrelatedtothediagnosisof severaldiseases,includingbreastcancer,rheumaticdiseases,vasculardiseases,etc. Apart from the diagnostic approach, the constant monitoring is also increasing its relevance in a wide range of different medical fields (i.e. the acquisition of vital signs, including temperature, respiratory rate, heart rate and blood perfusion). A recent approach still to be further explored involves the expansion towards 3D infraredimagingapplications. v vi Preface Chapter 2—Photoplethysmography Imaging and Common Optical Hybrid ImagingModalities Stillpresentingnon-invasiveandcontactlessapproaches,thesecondchapterallows obtainingskinperfusionstudies.Insuchmodality,theactivephotoplethysmography imaging (PPGI) provides the mapping of dermal blood perfusion dynamics. The definitionofPPGIconsistsinaclassicalphotoplethysmographyandpulseoximetry (SpO ), which actually involves the remote opto-electronical measurement of 2 arterial and/or venous blood volume changes. Although the results presented are verypromising,theyarestillpreliminary,sincetheystillneedtobestandardizedas aclinicalapplication,especiallyforremovingmovementartefacts. Chapter 3—Multimodal Image Fusion for Cardiac Resynchronization Ther- apyPlanning Cardiac resynchronization therapy (CRT) is due to treat patients with left-sided heartfailure.ThischapterpresentsoptimizationsofpreoperativeCRTplan,inorder to produce increasing rates of such therapies. They had used a variety of imaging modalities,describingtheanatomy,mechanicalactivationandtissuecharacteristics of the left ventricular (LV) under study. The authors developed a full workflow to process, register and fuse computer tomography (CT) images, ultrasound (US) imagesandmagneticresonanceimaging(MRI).Theresultsarerepresentedas3D patient-specificmodels,describingtheanatomyoftheinvolvedheartregions(such astheLV,thecoronaryveins),eventheelectromechanicaldelaysandthepresence of fibrosis. The results obtained with such 3D patient-specific models are helping physicians to select the best surgical procedures, involving the most adequate LV pacingsites. Chapter 4—CFD-Based Postprocessing of CT-MRI Data to Determine the MechanicsofRuptureinAbdominalAorticAneurysms Anapproachemployingacombinationoftwodifferentimagingmodalities(CTand MRI)foranapplicationinvolvingcomputationalfluiddynamics(CFD)ispresented in this chapter. The application involves a case study of diagnosis and surgical intervention’s decision on abdominal aortic aneurysms (AAA). In the presented study,sincetheclinicalmetricisnotenoughfortheprognosesrupture,amechanics- based approach and computational fluid dynamics (CFD) are also undertaken. This is important since a patient-specific geometry and boundary conditions are employed for the analysis. In addition, a fluid structure interaction (FSI)-based analysis of the abnormal aorta is employed. An important conclusion is outlined in this study that the maximum transverse diameter is not the only parameter of AAArupture’srisk.Additionally,themechanicsapproachbasedonmultimodality imagemethodologyproducedabetterdiagnosis. Chapter5—HumanHeadModellingSimulationAppliedtoElectroconvulsive Therapy Regarding the generation of 3D realistic human head models, this chapter recon- structedsuchmodelsbasedonmagneticresonanceimages.Theproblematichereis because it is aimed for electroconvulsive therapy (ECT) applications for treating Preface vii neurological conditions, once ECT uses low frequency and high amplitude of current,duringashortperiodoftime.Then,suchelectricalstimulationmaygenerate heat due to the Joule effect. So, bio-heat transfer equation and Laplace equation wereimplementedforcomputationalinvestigation.Theresultswereanalysedbased ontwopointsofview:thermalconductivity (whichproved tobebrain’ssafe)and electricalconductivity(whichisanimportantfactortobetakenintoaccount). Chapter6—UseofPhotonScatteringInteractionsinDiagnosisandTreatment ofDiseases Regarding the use of invasive imaging modalities, such as photon scattering applications in medicine, it presented two types of scattering events: incoherent (Compton) and coherent (Rayleigh). Therefore, first this chapter presents an overview of Compton cameras for gamma imaging, in the context of proton beam therapy. Potential methods for in vivo proton range verification are also presented. Additionally, the principle of operation of the Compton cameras and imaging reconstruction techniques is included (i.e. back-projection and stochastic approaches). Later, the chapter presents tissue diffraction, which is based on coherent scattering as a diagnostic tool. X-ray diffraction (XRD) is presented in ordertohelpinthedifferentiationofbothhealthandcanceroustissue,basedonthe abilitytodiscriminatetissuetypes.Further,someresultsincludingthegenerationof surface-rendered3Dvolumesarepresented,inordertoallowthe3Ddifferentiation ofthetissuesinvestigated(betweentumourandnormaltissue). Chapter 7—Digital Breast Tomosynthesis: Systems, Characterization and Simulation This chapter also focuses on invasive imaging modalities. Digital breast tomosyn- thesis (DBT) is an imaging application for breast cancer detection, which allows the generation of quasi-3D reconstruction images. First, this chapter presents an introduction and extensive review about digital X-ray tomosynthesis, including detailsaboutgeometries,performanceofthedetectors,automaticexposurecontrol performance,responsefunctionnoiseanalysis,modulationtransferfunction(MTF), etc. Then, the image quality measurements are presented, since these issues are important for clinical tasks, such as the detection of microcalcifications. The last partofthischaptercoversimagesimulationmethodsforDBToptimization.Several aspectsareconsidered(e.g.imageacquisitionparameters,detectorresponse,system geometry, radiation dose and image processing and reconstruction algorithms) consideringacceptablebreastdosesaswell.Regardingtheclinicaltrials,thereare twoapproaches,eitherinvolvingahugegroupofvolunteers(asymptomaticwomen) or based on image simulation methods (virtual trials). The last option consists in fast,radiation-freeandcost-effectiveoption. Chapter8—Out-of-CoreRenderingofLargeVolumetricDataSetsatMultiple LevelsofDetail Regarding the massive high-resolution volume data (especially obtained from the anatomical imaging modalities, such as X-ray, microtomography, ultrasonography and magnetic resonance imaging), there is a constant need for being able to viii Preface computationally deal with all these data. Then, the traditional in-core volume rendering is quite limited. So, this chapter presents an architecture for out-of-core volumerendering,keepingthelevelsofdetail.Therefore,severalexperimentswere conductedinordertoshowtheimprovementsofsuchapproach,especiallyinterms of memory storage, computational required time during the rendering process and evenframerate. Chapter 9—Geometric and Topological Modelling of Organs and Vascular StructuresfromCTData Still related to the generation and visualization of 3D models, the automatic segmentation of computer tomography (CT) images is a topic of interest. So, this chapterdealswiththeseveralmodellingproblems,mainlyforsurgeryplanningand trainingapplications.Forexample,someoftheissuespresentedaretransformation and extraction of DICOM data, organs with branching or complex structures, polygontriangulationandfileformats,amongotherissues.Ingeneral,theproposed solutionsarerelatedtoallowthealgorithmstobeusedtogetherinordertoprovide surface-based reconstructions not only of organs but also at vascular structures, whichcanevenbevisualizedwithtransparency. The common approach among all these chapters is that either they employ the combination of functional and/or anatomical imaging, leading to multimodality imaging,ortheyapplydifferentimagingmodalitiesinordertogenerate3Dimaging data.Thepromisingbiomedicalapplicationspresentedhereare(1)theacquisition of vital signs from infrared thermography images (including temperature, respi- ratory rate, heart rate and blood perfusion), (2) dermal blood perfusion dynamics (involving photoplethysmography and optical hybrid imaging modalities), (3) 3D generation of abdominal aortic aneurysms (AAA) (including computational fluid dynamics(CFD),fluidstructureinteraction(FSI)andmechanicalapproaches),(4) cardiac resynchronization therapy (CRT) involving the 3D generation of not only the heart but also specific regions of it (e.g. left ventricular (LV) and coronary veins) and (5) the generation of 3D realistic human head models (from MRI data) for electroconvulsive therapy applications; regarding the use of invasive radiation, such as (6) X-ray diffraction (XRD) for 3D differentiation between tumour and normal tissues and (7) digital breast tomosynthesis (DBT) for the generation of quasi-3D reconstruction of the breast, for cancer detection; and finally, some of the computational techniques that are allowing all these lately improvements, for example, (8) dealing with the out-of-core volume rendering of all the massive highresolutionvolumedatainvolvingmedicalimagingand(9)thegenerationand visualizationof3Dmodelsmainlyforsurgeryplanningandtrainingapplications. Curitiba,Paraná,Brazil MaurenAbreudeSouza Curitiba,Paraná,Brazil HumbertoRemigioGamba Campinas,SP,Brazil HelioPedrini Contents 1 InfraredThermography..................................................... 1 Carina Barbosa Pereira, Xinchi Yu, Stephan Dahlmanns, VladimirBlazek,SteffenLeonhardt,andDanielTeichmann 2 PhotoplethysmographyImagingandCommonOpticalHybrid ImagingModalities .......................................................... 31 Vladimir Blazek, Stephan Dahlmanns, Carina Barbosa Pereira, Xinchi Yu, Nikolai Blanik, Steffen Leonhardt, andClaudiaRosaBlazek 3 Multimodal Image Fusion for Cardiac Resynchronization TherapyPlanning............................................................ 67 SophieBruge,AntoineSimon,NicolasCourtial,JulianBetancur, AlfredoHernandez,FrançoisTavard,ErwanDonal,MathieuLederlin, ChristopheLeclercq,andMireilleGarreau 4 CFD-Based Postprocessing of CT-MRI Data to Determine theMechanicsofRuptureinAbdominalAorticAneurysms ........... 83 TejasCanchi,EddieY. K.Ng,AshishSaxena,andSriramNarayanan 5 Human Head Modelling Simulation Applied toElectroconvulsiveTherapy............................................... 103 Marília Menezes de Oliveira, Bo Song, Tony Ahfock, Yan Li, andPaulWen 6 Use of Photon Scattering Interactions in Diagnosis andTreatmentofDisease ................................................... 135 Robert Moss, Andrea Gutierrez, Amany Amin, Chiaki Crews, Robert Speller, Francesco Iacoviello, Paul Shearing, SarahVinnicombe,andSelinaKolokytha 7 Digital Breast Tomosynthesis: Systems, Characterization andSimulation ............................................................... 159 AnastasiosKonstantinidis,SelinaKolokytha,andAndriaHadjipanteli ix x Contents 8 Out-of-Core Rendering of Large Volumetric Data Sets atMultipleLevelsofDetail ................................................. 191 Paulo Henrique Junqueira Amorim, Thiago Franco de Moraes, JorgeVicenteLopesdaSilva,andHelioPedrini 9 GeometricandTopologicalModellingofOrgansandVascular StructuresfromCTData.................................................... 217 JoãoFradinhoOliveira,JoséBlasPagador,JoséLuisMoyano-Cuevas, FranciscoMiguelSánchez-Margallo,andHugoCapote Index............................................................................... 249