Tal Zeevi, MSc
Postgraduate AssociateAbout
Research
Publications
2024
A Hybrid Transformer-Convolutional Neural Network for Segmentation of Intracerebral Hemorrhage and Perihematomal Edema on Non-Contrast Head Computed Tomography (CT) with Uncertainty Quantification to Improve Confidence
Tran A, Desser D, Zeevi T, Abou Karam G, Dierksen F, Dell’Orco A, Kniep H, Hanning U, Fiehler J, Zietz J, Sanelli P, Malhotra A, Duncan J, Aneja S, Falcone G, Qureshi A, Sheth K, Nawabi J, Payabvash S. A Hybrid Transformer-Convolutional Neural Network for Segmentation of Intracerebral Hemorrhage and Perihematomal Edema on Non-Contrast Head Computed Tomography (CT) with Uncertainty Quantification to Improve Confidence. Bioengineering 2024, 11: 1274. DOI: 10.3390/bioengineering11121274.Peer-Reviewed Original ResearchNon-contrast head computed tomographyPerihematomal edemaHead computed tomographyIntracerebral hemorrhageComputed tomographyVolume similarityUniversity Medical Center Hamburg-EppendorfSecondary brain injuryYale cohortInfratentorial locationMulticentre trialCT scanTreatment planningNon-contrastHamburg-EppendorfImaging markersHemorrhagic strokeHemorrhageEdemaCohortBrain injuryDice coefficientDeep learning for prediction of post-thrombectomy outcomes based on admission CT angiography in large vessel occlusion stroke
Sommer J, Dierksen F, Zeevi T, Tran A, Avery E, Mak A, Malhotra A, Matouk C, Falcone G, Torres-Lopez V, Aneja S, Duncan J, Sansing L, Sheth K, Payabvash S. Deep learning for prediction of post-thrombectomy outcomes based on admission CT angiography in large vessel occlusion stroke. Frontiers In Artificial Intelligence 2024, 7: 1369702. PMID: 39149161, PMCID: PMC11324606, DOI: 10.3389/frai.2024.1369702.Peer-Reviewed Original ResearchEnd-to-endComputed tomography angiographyLarge vessel occlusionConvolutional neural networkDeep learning pipelineTrain separate modelsLogistic regression modelsResNet-50Deep learningAdmission computed tomography angiographyNeural networkLearning pipelineAdmission CT angiographyPreprocessing stepDiagnosis of large vessel occlusionsLarge vessel occlusion strokeReceiver operating characteristic areaEnsemble modelAutomated modelPre-existing morbidityCT angiographyReperfusion successNeurological examCross-validationOcclusion strokeMonte-Carlo Frequency Dropout for Predictive Uncertainty Estimation in Deep Learning
Zeevi T, Venkataraman R, Staib L, Onofrey J. Monte-Carlo Frequency Dropout for Predictive Uncertainty Estimation in Deep Learning. 2024, 00: 1-5. DOI: 10.1109/isbi56570.2024.10635511.Peer-Reviewed Original ResearchArtificial neural networkState-of-the-artMedical image dataPredictive uncertainty estimationBiomedical image dataImage dataOptimal artificial neural networkMC dropoutDropout approachSource-codeDrop-connectDeep learningNeural networkSignal spaceMonte-CarloPrediction uncertaintyUncertainty estimationDiverse setComprehensive comparisonPrediction scenariosDeepPosterior predictive distributionRepositoryDecision-makingNetworkDifferentiation of IDH-mutant Glioma Subtypes Using Unsupervised Dimensionality Reduction of MRI Biomarkers
Willms K, Zeevi T, Chadha S, von Reppert M, Lost J, Tillmanns N, Merkaj S, Huttner A, Aneja S, Aboian M. Differentiation of IDH-mutant Glioma Subtypes Using Unsupervised Dimensionality Reduction of MRI Biomarkers. Proceedings Of The International Society For Magnetic Resonance In Medicine ... Scientific Meeting And Exhibition. 2024 DOI: 10.58530/2024/3094.Peer-Reviewed Original ResearchAutomated MR Spectroscopy single-voxel placement in suspected diffuse glioma based on tumor biology
Chadha S, Jacobs S, Zeevi T, Tillmanns N, Merkaj S, Lost J, Lin M, Bousabarah K, Holler W, Memon F, Aneja S, Aboian M. Automated MR Spectroscopy single-voxel placement in suspected diffuse glioma based on tumor biology. Proceedings Of The International Society For Magnetic Resonance In Medicine ... Scientific Meeting And Exhibition. 2024 DOI: 10.58530/2024/5120.Peer-Reviewed Original ResearchTumor biologyDiffuse gliomasSingle-voxel magnetic resonance spectroscopyManagement of diffuse gliomasMagnetic resonance spectroscopyNon-invasive diagnosisVoxel placementMetabolite quantificationSingle-voxelMR imagingRadiology techniciansTumorGliomaPlacementResonance spectroscopyPoor-quality spectraClinicDiagnosisImpact of 18F-FDG PET Intensity Normalization on Radiomic Features of Oropharyngeal Squamous Cell Carcinomas and Machine Learning–Generated Biomarkers
Haider S, Zeevi T, Sharaf K, Gross M, Mahajan A, Kann B, Judson B, Prasad M, Burtness B, Aboian M, Canis M, Reichel C, Baumeister P, Payabvash S. Impact of 18F-FDG PET Intensity Normalization on Radiomic Features of Oropharyngeal Squamous Cell Carcinomas and Machine Learning–Generated Biomarkers. Journal Of Nuclear Medicine 2024, 65: jnumed.123.266637. PMID: 38514087, DOI: 10.2967/jnumed.123.266637.Peer-Reviewed Original ResearchOropharyngeal squamous cell carcinomaSquamous cell carcinomaHuman papillomavirusRadiomic featuresIntraclass correlation coefficientCell carcinomaLentiform nucleusHuman papillomavirus statusReceiver-operating-characteristic analysisReceiver-operating-characteristic curvePET radiomic featuresUnivariate logistic regressionF-FDGPrimary tumorTraining cohortValidation cohortRadiomic biomarkersUnivariate analysisInterindividual comparabilityPredictive valueDegree of reproducibilityMedian areaRadiomic markersLogistic regressionAUCReliable Prostate Cancer Risk Mapping From MRI Using Targeted and Systematic Core Needle Biopsy Histopathology
Zeevi T, Leapman M, Sprenkle P, Venkataraman R, Staib L, Onofrey J. Reliable Prostate Cancer Risk Mapping From MRI Using Targeted and Systematic Core Needle Biopsy Histopathology. IEEE Transactions On Biomedical Engineering 2024, 71: 1084-1091. PMID: 37874731, PMCID: PMC10901528, DOI: 10.1109/tbme.2023.3326799.Peer-Reviewed Original ResearchMagnetic resonance imagingIndividual patientsBiopsy locationProstate biopsy dataBiopsy histopathologyHistopathology scoresPathology scoresBiopsy dataMRI biomarkersTreatment planPatientsResonance imagingProstate regionBiomarkersTherapy treatment plansPathologyRepresentative sampleScoresImaging analysisPrevious studiesHistopathologyProstateCancerCliniciansRadiomics-Based Prediction of Collateral Status from CT Angiography of Patients Following a Large Vessel Occlusion Stroke
Avery E, Abou-Karam A, Abi-Fadel S, Behland J, Mak A, Haider S, Zeevi T, Sanelli P, Filippi C, Malhotra A, Matouk C, Falcone G, Petersen N, Sansing L, Sheth K, Payabvash S. Radiomics-Based Prediction of Collateral Status from CT Angiography of Patients Following a Large Vessel Occlusion Stroke. Diagnostics 2024, 14: 485. PMID: 38472957, PMCID: PMC10930945, DOI: 10.3390/diagnostics14050485.Peer-Reviewed Original ResearchCollateral statusCollateral scoreLarge vessel occlusionAcute LVO strokeRadiomics modelTest cohortCT angiography of patientsAdmission computed tomography angiographyAnterior circulation territoryAngiography of patientsLong-term outcomesReceiver operating characteristic areaRadiomics-based predictionCollateral arterial circulationOperating characteristics areaAdmission CTACirculation territoryCT angiographyClinical outcomesRadiomic featuresTreatment triageOcclusion strokeVessel occlusionPatientsArterial circulationUncertainty-aware deep-learning model for prediction of supratentorial hematoma expansion from admission non-contrast head computed tomography scan
Tran A, Zeevi T, Haider S, Abou Karam G, Berson E, Tharmaseelan H, Qureshi A, Sanelli P, Werring D, Malhotra A, Petersen N, de Havenon A, Falcone G, Sheth K, Payabvash S. Uncertainty-aware deep-learning model for prediction of supratentorial hematoma expansion from admission non-contrast head computed tomography scan. Npj Digital Medicine 2024, 7: 26. PMID: 38321131, PMCID: PMC10847454, DOI: 10.1038/s41746-024-01007-w.Peer-Reviewed Original ResearchDeep learning modelsHematoma expansionIntracerebral hemorrhageICH expansionComputed tomographyNon-contrast head CTNon-contrast head computed tomographyHigh risk of HEHead computed tomographyHigh-confidence predictionsRisk of HENon-contrast headReceiver operating characteristic areaModifiable risk factorsMonte Carlo dropoutOperating characteristics areaPotential treatment targetHead CTVisual markersIdentified patientsAutomated deep learning modelDataset of patientsRisk factorsHigh riskPatientsPeri-hematomal edema shape features related to 3-month outcome in acute supratentorial intracerebral hemorrhage
Dierksen F, Tran A, Zeevi T, Maier I, Qureshi A, Sanelli P, Werring D, Malhotra A, Falcone G, Sheth K, Payabvash S. Peri-hematomal edema shape features related to 3-month outcome in acute supratentorial intracerebral hemorrhage. European Stroke Journal 2024, 9: 383-390. PMID: 38179883, PMCID: PMC11318427, DOI: 10.1177/23969873231223814.Peer-Reviewed Original ResearchNIH Stroke ScaleGlasgow Coma ScaleAssociation of baselineIntracerebral hemorrhagePerihematomal edemaHematoma volumeAcute supratentorial intracerebral hemorrhageNon-traumatic intracerebral hemorrhageSecondary brain injuryAcute ICH patientsIndependent prognostic factorSupratentorial intracerebral hemorrhageAdmission NIH Stroke ScalePotential treatment targetRankin scoreStroke ScaleIndependent predictorsPatient agePrognostic factorsComa ScaleMultivariable analysisICH patientsPrognostic valueBrain injuryFavorable outcome