Jens Petersen · Raúl San José Estépar · Alexander Schmidt-Richberg · Sarah Gerard · Bianca Lassen-Schmidt · Colin Jacobs · Reinhard Beichel · Kensaku Mori (Eds.) 2 0 Thoracic Image 5 2 1 S Analysis C N L Second International Workshop, TIA 2020 Held in Conjunction with MICCAI 2020 Lima, Peru, October 8, 2020, Proceedings Lecture Notes in Computer Science 12502 Founding Editors Gerhard Goos Karlsruhe Institute of Technology, Karlsruhe, Germany Juris Hartmanis Cornell University, Ithaca, NY, USA Editorial Board Members Elisa Bertino Purdue University, West Lafayette, IN, USA Wen Gao Peking University, Beijing, China Bernhard Steffen TU Dortmund University, Dortmund, Germany Gerhard Woeginger RWTH Aachen, Aachen, Germany Moti Yung Columbia University, New York, NY, USA More information about this series at http://www.springer.com/series/7412 ú é é Jens Petersen Ra l San Jos Est par (cid:129) (cid:129) Alexander Schmidt-Richberg (cid:129) Sarah Gerard Bianca Lassen-Schmidt (cid:129) (cid:129) Colin Jacobs Reinhard Beichel (cid:129) (cid:129) Kensaku Mori (Eds.) Thoracic Image Analysis Second International Workshop, TIA 2020 Held in Conjunction with MICCAI 2020 Lima, Peru, October 8, 2020 Proceedings 123 Editors Jens Petersen RaúlSanJoséEstépar University of Copenhagen Harvard Medical School Copenhagen, Denmark Boston, MA, USA Alexander Schmidt-Richberg SarahGerard Philips (Germany) Harvard Medical School Hamburg,Hamburg,Germany Boston, MA, USA Bianca Lassen-Schmidt Colin Jacobs Fraunhofer Institute for Medical Image RadiologyandNuclear Medicine Computing Radboud University Medical Center Bremen, Bremen, Germany Nijmegen, Gelderland, The Netherlands Reinhard Beichel Kensaku Mori University of Iowa Graduate Schoolof Informatics Iowa City,IA, USA Nagoya University Nagoya,Japan ISSN 0302-9743 ISSN 1611-3349 (electronic) Lecture Notesin Computer Science ISBN 978-3-030-62468-2 ISBN978-3-030-62469-9 (eBook) https://doi.org/10.1007/978-3-030-62469-9 LNCSSublibrary:SL6–ImageProcessing,ComputerVision,PatternRecognition,andGraphics ©SpringerNatureSwitzerlandAG2020 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartofthe material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodologynow knownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthorsandtheeditorsaresafetoassumethattheadviceandinformationinthisbookare believedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsortheeditors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregardtojurisdictionalclaimsin publishedmapsandinstitutionalaffiliations. ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland Preface TheSecondInternationalWorkshoponThoracicImageAnalysis(TIA2020)washeld as an entirely online satellite event of the Medical Image Computing and Computer-Assisted Intervention Conference (MICCAI 2020). Building on the history of the workshop and the Pulmonary Image Analysis workshop, a roughly biannual eventatMICCAIgoingbackmorethan10years,theworkshopaimedtobringtogether medical image analysis researchers in the area of thoracic imaging to discuss recent advancesinthisrapidlydevelopingfield.TheCOVID-19pandemichasbroughtmuch attention to lung imaging, and the role of CT imaging in the diagnostic workflow of COVID-19anditsclinicalresolutionhasbeenanimportantresearchtopic.Inaddition tothat,cardiovasculardisease,lungcancer,andchronicobstructivepulmonarydisease, three diseases all visible on thoracic imaging, are among the top causes of death worldwide. Many imaging modalities are currently available to study the pulmonary andcardiacsystem,includingX-ray,CT,PET,ultrasound,andMRI.Weinvitedpapers that dealt with all aspects of image analysis of thoracic data, including but not limited to: image acquisition and reconstruction, segmentation, registration, quantification, visualization, validation, population-based modeling, biophysical modeling (compu- tational anatomy), deep learning, image analysis in small animals, outcome-based research, and novel infectious disease applications (e.g., COVID-19, TB, etc.). We particularlywelcomednovelworkfocusedaroundtheneedfornewmethodologiesfor predisposition, diagnosis, staging, and resolution assessment of COVID-19 infections asanemergingdiseaseaswellasgood-sizedindependentvalidationstudiesontheuse of deep learning models in the area of thoracic imaging, despite having possibly little technical novelty. The 16 papers submitted to the workshop were reviewed in a double-blind manner with atleast three reviewersperpaper, whoseaffiliationsand recentpublications were checked to avoid conflicts of interest. The review process included a blind review followed by a meta-review done by consensus by three committee members that evaluatedthereviewsandthepapersasawhole.Allofthesubmittedpaperswerelong format (8–12 pages). Out of the submitted papers, 15 were accepted for presentation; however, 1 paper was withdrawn after acceptance for inconsistencies in the reported results.Theremainingpapersweregroupedintofourtopics,whicharereflectedinthe structure of these proceedings – Image Segmentation (5), Computer-Aided Diagnosis andLocalization(4),ImageTranslationandInpainting(3),andImageregistration(2). Deep learning techniques continue to expand with 15 out of the 16 submissions using someelementsofdeeplearning.Wewerepleasedtoseethattheworkshophashelped bringfocustoCOVID-19researchwithnolessthanthreeworkscoveringtheanalysis anddetectionofthisdisease.Theimagingmodalitiesusedwereagoodmixtureof2D X-ray,3DCT,4DCT,DECT,andMRI,demonstratingthecomplementaryinformation broughttogetherbydifferent modalitiesusedtostudythethoracicsystem.Webelieve that thoracic image analysis keeps on playing a crucial role in the understanding of vi Preface chronic and infectious diseases with high morbidity and mortality as reflected by the works submitted to this workshop. We want to express our gratitude to all the authors for submitting papers to TIA 2020 and everyone involved in the organization and peer-review process. September 2020 Jens Petersen Raúl San José Estépar Organization Organizing Committee Kensaku Mori Nagoya University, Japan Raúl San José Estépar Brigham and Women’s Hospital, Harvard Medical School, USA Sarah Gerard Brigham and Women’s Hospital, Harvard Medical School, USA Reinhard Beichel The University of Iowa, USA Colin Jacobs Radboud University Medical Center, The Netherlands Alexander Philips Research Laboratories, Germany Schmidt-Richberg Bianca Lassen-Schmidt Fraunhofer Institute for Digital Medicine MEVIS, Germany Jens Petersen University of Copenhagen, Denmark Program Committee Raúl San José Estépar Brigham and Women’s Hospital, Harvard Medical School, USA Sarah Gerard Brigham and Women’s Hospital, Harvard Medical School, USA Alexander Philips Research Laboratories, Germany Schmidt-Richberg Jens Petersen University of Copenhagen, Denmark Additional Reviewers Abraham George Smith Olivier Nempont Alexander Schmidt-Richberg Pietro Nardelli Christian Buerger Raghavendra Selvan Colin Jacobs Rene Werner Jens Petersen Sarah Gerard Lasse Hansen Silas Ørting Matthias Wilms Tobias Klinder Mattias Heinrich Contents Image Segmentation Two-Stage Mapping-Segmentation Framework for Delineating COVID-19 Infections from Heterogeneous CT Images. . . . . . . . . . . . . . . . . . . . . . . . . 3 Tong Li, Zhuochen Wang, Yanbo Chen, Lichi Zhang, Yaozong Gao, Feng Shi, Dahong Qian, Qian Wang, and Dinggang Shen Multi-cavity Heart Segmentation in Non-contrast Non-ECG Gated CT Scans with F-CNN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Rafael Moreta-Martínez, Gonzalo Vegas Sánchez-Ferrero, Lasse Andresen, Jakob Qvortrup Holsting, and Raúl San José Estépar 3D Deep Convolutional Neural Network-Based Ventilated Lung Segmentation Using Multi-nuclear Hyperpolarized Gas MRI . . . . . . . . . . . . 24 Joshua R. Astley, Alberto M. Biancardi, Paul J. C. Hughes, Laurie J. Smith, Helen Marshall, James Eaden, Jody Bray, Nicholas D. Weatherley, Guilhem J. Collier, Jim M. Wild, and Bilal A. Tahir Lung Cancer Tumor Region Segmentation Using Recurrent 3D-DenseUNet. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Uday Kamal, Abdul Muntakim Rafi, Rakibul Hoque, Jonathan Wu, and Md. Kamrul Hasan 3D Probabilistic Segmentation and Volumetry from 2D Projection Images. . . 48 Athanasios Vlontzos, Samuel Budd, Benjamin Hou, Daniel Rueckert, and Bernhard Kainz Computer-Aided Diagnosis and Localization CovidDiagnosis: Deep Diagnosis of COVID-19 Patients Using Chest X-Rays. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Kushagra Mahajan, Monika Sharma, Lovekesh Vig, Rishab Khincha, Soundarya Krishnan, Adithya Niranjan, Tirtharaj Dash, Ashwin Srinivasan, and Gautam Shroff Can We Trust Deep Learning Based Diagnosis? The Impact of Domain Shift in Chest Radiograph Classification . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Eduardo H. P. Pooch, Pedro Ballester, and Rodrigo C. Barros x Contents A Weakly Supervised Deep Learning Framework for COVID-19 CT Detection and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Ophir Gozes, Maayan Frid-Adar, Nimrod Sagie, Asher Kabakovitch, Dor Amran, Rula Amer, and Hayit Greenspan Deep Reinforcement Learning for Localization of the Aortic Annulus in Patients with Aortic Dissection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Marina Codari, Antonio Pepe, Gabriel Mistelbauer, Domenico Mastrodicasa, Shannon Walters, Martin J. Willemink, and Dominik Fleischmann Image Translation and Inpainting Functional-Consistent CycleGAN for CT to Iodine Perfusion Map Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Pietro Nardelli, Rubén San José Estépar, Farbod N. Rahaghi, and Raúl San José Estépar MRI to CTA Translation for Pulmonary Artery Evaluation Using CycleGANs Trained with Unpaired Data . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Maialen Stephens, Raúl San José Estepar, Jesús Ruiz-Cabello, Ignacio Arganda-Carreras, Iván Macía, and Karen López-Linares Semi-supervised Virtual Regression of Aortic Dissections Using 3D Generative Inpainting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Antonio Pepe, Gabriel Mistelbauer, Christina Gsaxner, Jianning Li, Dominik Fleischmann, Dieter Schmalstieg, and Jan Egger Image Registration Registration-Invariant Biomechanical Features for Disease Staging of COPD in SPIROMICS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Muhammad F. A. Chaudhary, Yue Pan, Di Wang, Sandeep Bodduluri, Surya P. Bhatt, Alejandro P. Comellas, Eric A. Hoffman, Gary E. Christensen, and Joseph M. Reinhardt Deep Group-Wise Variational Diffeomorphic Image Registration . . . . . . . . . 155 Tycho F. A. van der Ouderaa, Ivana Išgum, Wouter B. Veldhuis, and Bob D. de Vos Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165