2024
Patient-Specific Heart Geometry Modeling for Solid Biomechanics Using Deep Learning
Pak D, Liu M, Kim T, Liang L, Caballero A, Onofrey J, Ahn S, Xu Y, McKay R, Sun W, Gleason R, Duncan J. Patient-Specific Heart Geometry Modeling for Solid Biomechanics Using Deep Learning. IEEE Transactions On Medical Imaging 2024, 43: 203-215. PMID: 37432807, PMCID: PMC10764002, DOI: 10.1109/tmi.2023.3294128.Peer-Reviewed Original ResearchConceptsFinite element analysisDeep learning methodsSpatial accuracyElement analysisDeep learningStress estimationLearning methodsSimulation accuracyDeployment simulationHigh spatial accuracyThin structuresMesh generationVolumetric meshingDeformation energyGeometry modelingVolumetric meshMesh qualityElement qualitySimultaneous optimizationMain noveltyBiomechanics studiesMeshModeling characteristicsAccuracyDownstream analysis
2023
A Multiclass Radiomics Method–Based WHO Severity Scale for Improving COVID-19 Patient Assessment and Disease Characterization From CT Scans
Henao J, Depotter A, Bower D, Bajercius H, Todorova P, Saint-James H, de Mortanges A, Barroso M, He J, Yang J, You C, Staib L, Gange C, Ledda R, Caminiti C, Silva M, Cortopassi I, Dela Cruz C, Hautz W, Bonel H, Sverzellati N, Duncan J, Reyes M, Poellinger A. A Multiclass Radiomics Method–Based WHO Severity Scale for Improving COVID-19 Patient Assessment and Disease Characterization From CT Scans. Investigative Radiology 2023, 58: 882-893. PMID: 37493348, PMCID: PMC10662611, DOI: 10.1097/rli.0000000000001005.Peer-Reviewed Original ResearchConceptsCOVID-19 positive patientsClinical Progression ScaleLung lesionsLesion modelDisease severityGround-glass opacitiesCOVID-19 patientsRadiologist assessmentExpert thoracic radiologistsMulticenter cohortPleural effusionDisease extentRetrospective studyDevelopment cohortPatient assessmentTomography scanCT scanSeverity ScalePatient's diseaseTissue lesionsThoracic radiologistsLesionsPatientsRadiomics modelRadiomic features
2022
DuDoUFNet: Dual-Domain Under-to-Fully-Complete Progressive Restoration Network for Simultaneous Metal Artifact Reduction and Low-Dose CT Reconstruction
Zhou B, Chen X, Xie H, Zhou S, Duncan JS, Liu C. DuDoUFNet: Dual-Domain Under-to-Fully-Complete Progressive Restoration Network for Simultaneous Metal Artifact Reduction and Low-Dose CT Reconstruction. IEEE Transactions On Medical Imaging 2022, 41: 3587-3599. PMID: 35816532, PMCID: PMC9812027, DOI: 10.1109/tmi.2022.3189759.Peer-Reviewed Original ResearchDual-domain self-supervised learning for accelerated non-Cartesian MRI reconstruction
Zhou B, Schlemper J, Dey N, Mohseni Salehi SS, Sheth K, Liu C, Duncan JS, Sofka M. Dual-domain self-supervised learning for accelerated non-Cartesian MRI reconstruction. Medical Image Analysis 2022, 81: 102538. PMID: 35926336, DOI: 10.1016/j.media.2022.102538.Peer-Reviewed Original ResearchConceptsNon-Cartesian MRI reconstructionMRI reconstructionUndersampled dataPrevious baseline methodsSelf-supervised approachSelf-supervised learningHigh-quality reconstructionReconstruction networkAppearance consistencyDataset demonstrateBaseline methodsImage domainDisjoint partitionsSupervised trainingPractical adoptionReconstruction accuracyDomain partitionImproved image qualityImage qualityDDSSSampling patternK-spaceExperimental resultsNetworkMotion robustnessUsing Adversarial Images to Assess the Robustness of Deep Learning Models Trained on Diagnostic Images in Oncology
Joel MZ, Umrao S, Chang E, Choi R, Yang DX, Duncan JS, Omuro A, Herbst R, Krumholz HM, Aneja S. Using Adversarial Images to Assess the Robustness of Deep Learning Models Trained on Diagnostic Images in Oncology. JCO Clinical Cancer Informatics 2022, 6: e2100170. PMID: 35271304, PMCID: PMC8932490, DOI: 10.1200/cci.21.00170.Peer-Reviewed Original Research
2021
Quantitative Automated Segmentation of Lipiodol Deposits on Cone-Beam CT Imaging Acquired during Transarterial Chemoembolization for Liver Tumors: A Deep Learning Approach
Malpani R, Petty CW, Yang J, Bhatt N, Zeevi T, Chockalingam V, Raju R, Petukhova-Greenstein A, Santana JG, Schlachter TR, Madoff DC, Chapiro J, Duncan J, Lin M. Quantitative Automated Segmentation of Lipiodol Deposits on Cone-Beam CT Imaging Acquired during Transarterial Chemoembolization for Liver Tumors: A Deep Learning Approach. Journal Of Vascular And Interventional Radiology 2021, 33: 324-332.e2. PMID: 34923098, PMCID: PMC8972393, DOI: 10.1016/j.jvir.2021.12.017.Peer-Reviewed Original ResearchHigh-Resolution Magnetic Resonance Spectroscopic Imaging using a Multi-Encoder Attention U-Net with Structural and Adversarial Loss
Dong S, Hangel G, Bogner W, Trattnig S, Rössler K, Widhalm G, De Feyter HM, De Graaf RA, Duncan JS. High-Resolution Magnetic Resonance Spectroscopic Imaging using a Multi-Encoder Attention U-Net with Structural and Adversarial Loss. Annual International Conference Of The IEEE Engineering In Medicine And Biology Society (EMBC) 2021, 00: 2891-2895. PMID: 34891851, DOI: 10.1109/embc46164.2021.9630146.Peer-Reviewed Original ResearchDuDoDR-Net: Dual-domain data consistent recurrent network for simultaneous sparse view and metal artifact reduction in computed tomography
Zhou B, Chen X, Zhou SK, Duncan JS, Liu C. DuDoDR-Net: Dual-domain data consistent recurrent network for simultaneous sparse view and metal artifact reduction in computed tomography. Medical Image Analysis 2021, 75: 102289. PMID: 34758443, PMCID: PMC8678361, DOI: 10.1016/j.media.2021.102289.Peer-Reviewed Original ResearchConceptsRecurrent networksSevere streak artifactsRecurrent frameworkArtifact reductionSparse viewsImage domainReconstruction qualityCT metal artifact reductionX-ray projectionsMetal artifact reductionArtifact-free imagesMedical diagnosisPrevious methodsProjection dataConsistent layerExperimental resultsMDPET: A Unified Motion Correction and Denoising Adversarial Network for Low-Dose Gated PET
Zhou B, Tsai YJ, Chen X, Duncan JS, Liu C. MDPET: A Unified Motion Correction and Denoising Adversarial Network for Low-Dose Gated PET. IEEE Transactions On Medical Imaging 2021, 40: 3154-3164. PMID: 33909561, PMCID: PMC8588635, DOI: 10.1109/tmi.2021.3076191.Peer-Reviewed Original ResearchConceptsMotion estimationPyramid networkAdversarial networkAccurate motion estimationMotion correctionLow-noise reconstructionGated positron emission tomographyMotion correction methodMotion estimation networkGated PET dataEstimation networkRecurrent layersDenoising NetworkRespiratory motion blurringExperimental resultsLow-noise imagesMotion blurringNoise levelCorrection methodNetworkPET reconstructionPrevious methodsImage qualityImagesEstimationBrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis
Li X, Zhou Y, Dvornek N, Zhang M, Gao S, Zhuang J, Scheinost D, Staib LH, Ventola P, Duncan JS. BrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis. Medical Image Analysis 2021, 74: 102233. PMID: 34655865, PMCID: PMC9916535, DOI: 10.1016/j.media.2021.102233.Peer-Reviewed Original ResearchConceptsFunctional magnetic resonance imagesGraph neural network frameworkMedical image analysisGraph neural networkGraph convolutional layersNeural network frameworkDifferent evaluation metricsSpecific task statesIndependent fMRI datasetsPooling layerConvolutional layersConsistency lossNetwork frameworkNeural networkFMRI datasetsImage analysis methodEvaluation metricsDetection resultsBrain graphsSubjects releaseROI selectionImage analysisCognitive stimuliTask statesFMRI analysisAnatomy-guided multimodal registration by learning segmentation without ground truth: Application to intraprocedural CBCT/MR liver segmentation and registration
Zhou B, Augenfeld Z, Chapiro J, Zhou SK, Liu C, Duncan JS. Anatomy-guided multimodal registration by learning segmentation without ground truth: Application to intraprocedural CBCT/MR liver segmentation and registration. Medical Image Analysis 2021, 71: 102041. PMID: 33823397, PMCID: PMC8184611, DOI: 10.1016/j.media.2021.102041.Peer-Reviewed Original ResearchConceptsMultimodal registrationLiver segmentationLarge-scale manual annotationGround truthMultimodal image registrationMultimodal registration methodSegmentation networkDomain adaptationManual annotationSource modalityImage registrationRegistration frameworkSegmentationImage-guided interventionsRegistration methodMedical imagingDiagnostic medical imagingCorrect transformationLimited FOVStructure informationIntraprocedural CBCTImage qualitySegmenterExperimental resultsPatient dataDeep learning–assisted differentiation of pathologically proven atypical and typical hepatocellular carcinoma (HCC) versus non-HCC on contrast-enhanced MRI of the liver
Oestmann PM, Wang CJ, Savic LJ, Hamm CA, Stark S, Schobert I, Gebauer B, Schlachter T, Lin M, Weinreb JC, Batra R, Mulligan D, Zhang X, Duncan JS, Chapiro J. Deep learning–assisted differentiation of pathologically proven atypical and typical hepatocellular carcinoma (HCC) versus non-HCC on contrast-enhanced MRI of the liver. European Radiology 2021, 31: 4981-4990. PMID: 33409782, PMCID: PMC8222094, DOI: 10.1007/s00330-020-07559-1.Peer-Reviewed Original ResearchConceptsNon-HCC lesionsHepatocellular carcinomaHCC lesionsAtypical imagingGrading systemLI-RADS criteriaAtypical imaging featuresPrimary liver cancerTypical hepatocellular carcinomaAtypical hepatocellular carcinomaContrast-enhanced MRISensitivity/specificityLiver transplantMethodsThis IRBRetrospective studyLiver malignanciesImaging featuresLiver cancerAtypical featuresConclusionThis studyLesionsMRIClinical applicationCarcinomaImage-based diagnosis
2020
Layer Embedding Analysis in Convolutional Neural Networks for Improved Probability Calibration and Classification
Zhang F, Dvornek N, Yang J, Chapiro J, Duncan J. Layer Embedding Analysis in Convolutional Neural Networks for Improved Probability Calibration and Classification. IEEE Transactions On Medical Imaging 2020, 39: 3331-3342. PMID: 32356739, PMCID: PMC7606489, DOI: 10.1109/tmi.2020.2990625.Peer-Reviewed Original ResearchConceptsConvolutional neural networkNeural networkClassification taskProbability calibrationTissue classification tasksImage representationBaseline methodsPublic datasetsModel performanceRandom forest modelNetworkBetter performanceForest modelDatasetClassificationTaskCT imagesImagesOriginal model outputMR imagesModel outputInstitutional datasetPerformanceEmbeddingOutputAutomated feature quantification of Lipiodol as imaging biomarker to predict therapeutic efficacy of conventional transarterial chemoembolization of liver cancer
Stark S, Wang C, Savic LJ, Letzen B, Schobert I, Miszczuk M, Murali N, Oestmann P, Gebauer B, Lin M, Duncan J, Schlachter T, Chapiro J. Automated feature quantification of Lipiodol as imaging biomarker to predict therapeutic efficacy of conventional transarterial chemoembolization of liver cancer. Scientific Reports 2020, 10: 18026. PMID: 33093524, PMCID: PMC7582153, DOI: 10.1038/s41598-020-75120-7.Peer-Reviewed Original ResearchConceptsConventional transarterial chemoembolizationLipiodol depositionTransarterial chemoembolizationLiver cancerPeripheral depositionLipiodol depositsTherapeutic efficacyNecrotic tumor areasBaseline MRITherapy optionsTumor responseTreatment responseTumor volumeLiver lesionsLipiodolH postTumor areaH-CTHounsfield unitsBiomarkersChemoembolizationHigh rateTumorsCancerImproved responseMulti-site fMRI analysis using privacy-preserving federated learning and domain adaptation: ABIDE results
Li X, Gu Y, Dvornek N, Staib LH, Ventola P, Duncan JS. Multi-site fMRI analysis using privacy-preserving federated learning and domain adaptation: ABIDE results. Medical Image Analysis 2020, 65: 101765. PMID: 32679533, PMCID: PMC7569477, DOI: 10.1016/j.media.2020.101765.Peer-Reviewed Original ResearchConceptsDeep learning modelsFederated LearningPrivacy-preserving federated learningLearning modelFederated learning approachPrivacy-preserving strategyDomain adaptation methodsData analysis problemsLocal model weightsIterative optimization algorithmEntity dataDomain adaptationLearning approachLearning formulationMulti-site dataRandomization mechanismAdaptation methodNeuroimage analysisDifferent tasksModel weightsModel optimizationOptimization algorithmPrivate informationTraining strategyAnalysis problemSparse Data–Driven Learning for Effective and Efficient Biomedical Image Segmentation
Onofrey JA, Staib LH, Huang X, Zhang F, Papademetris X, Metaxas D, Rueckert D, Duncan JS. Sparse Data–Driven Learning for Effective and Efficient Biomedical Image Segmentation. Annual Review Of Biomedical Engineering 2020, 22: 1-27. PMID: 32169002, PMCID: PMC9351438, DOI: 10.1146/annurev-bioeng-060418-052147.Peer-Reviewed Original Research
2019
Deep learning for liver tumor diagnosis part II: convolutional neural network interpretation using radiologic imaging features
Wang CJ, Hamm CA, Savic LJ, Ferrante M, Schobert I, Schlachter T, Lin M, Weinreb JC, Duncan JS, Chapiro J, Letzen B. Deep learning for liver tumor diagnosis part II: convolutional neural network interpretation using radiologic imaging features. European Radiology 2019, 29: 3348-3357. PMID: 31093705, PMCID: PMC7243989, DOI: 10.1007/s00330-019-06214-8.Peer-Reviewed Original ResearchMeSH KeywordsAdultAgedAlgorithmsBile Duct NeoplasmsBile Ducts, IntrahepaticCarcinoma, HepatocellularCholangiocarcinomaDeep LearningFemaleHumansImage Interpretation, Computer-AssistedLiver NeoplasmsMachine LearningMagnetic Resonance ImagingMaleMiddle AgedNeural Networks, ComputerPredictive Value of TestsProof of Concept StudyRetrospective StudiesConceptsDeep learning systemConvolutional neural networkLearning systemRelevance scoresFeature mapsPre-trained CNN modelsFeature relevance scoresMulti-phasic MRINeural network interpretationEvidence-based decision supportDeep NeuralDeep learningCNN modelLesion classifierLearning prototypeNeural networkOriginal imageSystem prototypeDecision supportLesion classificationNetwork interpretationImage voxelsIncorrect featuresLesion classesTest setDeep learning for liver tumor diagnosis part I: development of a convolutional neural network classifier for multi-phasic MRI
Hamm CA, Wang CJ, Savic LJ, Ferrante M, Schobert I, Schlachter T, Lin M, Duncan JS, Weinreb JC, Chapiro J, Letzen B. Deep learning for liver tumor diagnosis part I: development of a convolutional neural network classifier for multi-phasic MRI. European Radiology 2019, 29: 3338-3347. PMID: 31016442, PMCID: PMC7251621, DOI: 10.1007/s00330-019-06205-9.Peer-Reviewed Original ResearchAdultAgedBile Duct NeoplasmsBile Ducts, IntrahepaticCarcinoma, HepatocellularCholangiocarcinomaDeep LearningFemaleHumansImage Interpretation, Computer-AssistedLiver NeoplasmsMagnetic Resonance ImagingMaleMiddle AgedNeural Networks, ComputerReproducibility of ResultsROC CurveSensitivity and SpecificityUnited States
2018
Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma.
Abajian A, Murali N, Savic LJ, Laage-Gaupp FM, Nezami N, Duncan JS, Schlachter T, Lin M, Geschwind JF, Chapiro J. Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma. Journal Of Visualized Experiments 2018 PMID: 30371657, PMCID: PMC6235502, DOI: 10.3791/58382.Peer-Reviewed Original ResearchConceptsIntra-arterial therapyN patientsHepatocellular carcinomaTrans-arterial therapiesIntra-arterial treatmentCohort of patientsStandard of careLikelihood of responseClinical research questionsSurgical resectionNew patientsTreatment responseUnivariate associationsPatientsTraining patientsInterventional radiologyTherapyCarcinomaTreatmentImage-guided therapyOutcomesFinal modelImaging dataResectionResponseLearning Generalizable Recurrent Neural Networks from Small Task-fMRI Datasets
Dvornek NC, Yang D, Ventola P, Duncan JS. Learning Generalizable Recurrent Neural Networks from Small Task-fMRI Datasets. Lecture Notes In Computer Science 2018, 11072: 329-337. PMID: 30873514, PMCID: PMC6411297, DOI: 10.1007/978-3-030-00931-1_38.Peer-Reviewed Original ResearchConceptsRecurrent neural networkNeural networkTask fMRI datasetsMedical image analysis problemsSuch deep networksImage analysis problemsTask fMRI scanTypical control subjectsDeep networkDeep learningTraining lossSmall datasetsLarge datasetsNumber of approachesAutism spectrum disorderAnalysis problemDatasetNetworkTraining runsImage analysisGeneralizable modelNon-imaging variablesSpectrum disorderFMRI analysisModel performance