Output list
Conference presentation
Removing the Foreground and Background of Chest X-rays using a Generative Adversarial Network (GAN)
Date presented 10/15/2022
Biomedical Engineering Society (BMES) Annual Meeting, 10/12/2022–10/15/2022, San Antonio, Texas
Conference presentation
Deep-Learning CT-based COPD Latent Traits Associate with SPECT-based Ventilation Patterns
Date presented 10/2021
Biomedical Engineering Society (BMES) Annual Meeting, 10/06/2021–10/09/2021, Orlando, Florida, USA
Conference presentation
Machine Learning of CT-based Imaging Clusters in Asthma and Chronic Obstructive Pulmonary Disease (COPD)
Date presented 08/2019
Workshop on Application of Biomechanics in Clinical Treatment and Rehabilitation & 1st International Conference on "Medicine in Novel Technology and Devices", 08/11/2019–08/12/2019, Beijing, China
Invited talk.
Conference presentation
Deep Learning CT Imaging-based Analysis of COPD Cohorts
Date presented 07/13/2019
International Environmental Lung Disease Symposium, 07/12/2019–07/13/2019, Iowa City, Iowa
Invited talk.
Conference presentation
CFD Lung Models for Drug Delivery
Date presented 03/13/2019
American Society for Clinical Pharmacology & Therapeutics (ASCPT) Pre-Conference “PBPK Modeling for the Development and Approval of Locally Acting Products”, 03/13/2019, Washington, DC, USA
Invited talk
Conference presentation
Automated Cluster Classification into 4 Identified Clusters
Date presented 11/08/2018
Deep Learning Workshop, COPDGene Denver Meeting, 11/18/2018, Denver, Colorado
Deep Learning Workshop, COPDGene Denver Meeting, 11/08/2018, Denver, CO
Invited talk
Conference presentation
Lung Motionography in Emphysema Patients Based on Mass Preserving Non-Rigid Image Registration of Inspiration-Expiration CTs: Correlation with PFTs and Air trapping
Date presented 06/2017
World Congress Thoracic Imaging, 06/18/2017–06/21/2017, Boston, Massachusetts
Conference poster
Multiscale Imaging-Based Cluster Analysis of a Cohort of Current Smokers
Date presented 03/22/2017
Annual Multiscale Modeling (MSM) Consortium Meeting, 03/22/2017–03/24/2017, Bethesda, Maryland
Classification of patients with chronic obstructive pulmonary disease (COPD) is usually based on the severity of airflow limitation, e.g. pre- and post- bronchodilator FEV1, which may not reflect the phenotypic heterogeneity nature of the disease. Recently, we have developed a multiscale-imaging based cluster analysis (MICA) and applied it to analyze 248 asthmatics from Severe Asthma Research Program (SARP). MICA yielded four stable imaging clusters which exhibit strong associations with clinical characteristics (1). In this study, we further applied MICA to a cohort of current smokers from the SubPopulations and InteRmediate Outcome Measures in COPD Study (SPIROMICS). We obtained four statistically stable clusters with distinct imaging-based structural and functional alterations. We further demonstrated that these imaging clusters exhibit significant associations with distinct clinical phenotypes used for diagnosis of COPD. Our MICA provides a link between individual and population scales.
Conference presentation
Acinar Scale Relative Regional Air Volume Change Maps Reflecting Mechanics of LAAs in COPD Patients
Date presented 11/29/2016
Radiological Society of North America (RSNA) Scientific Assembly and Annual Meeting, 11/27/2016–12/02/2016, Chicago, Illinois
PURPOSE
Low attenuation areas (LAAs) on CT scans have been used to evaluate emphysema and air trapping. However, the regional ventilation changes of LAAs while breathing may vary due to the various causes of emphysema or air trapping. Therefore, the purpose of this study was to investigate regional air volume change at acinar scale of the lung using mass preserving image registration technique and compare with the -950HUinsp, -856HUexp and PFTs in COPD.
METHOD AND MATERIALS
18 emphysema patients (12 centriacinar, 6 distal acinar emphysema) and 10 normal subjects were included in the study. VIDA Apollo software (Coralville, IA) and mass preserving image registration technique were used to compute relative regional air volume change (RRAVC) between inspiration and expiration CT scans. Then, low ventilation area (LVA) was defined as percent lung volume of RRAVC < 0.8. -950HUinsp, -856HUexp and LVA0.8 in total lung were correlated with FEV1, FEV1/FVC and compared between normal and emphysema patients.
RESULTS
LVA0.8 and -856HUexp showed positive correlation with FEV1 (r=-0.89 and p=0.016, r=-0.91, p=0.014) while -950HUinsp did not show a correlation with FEV1 in distal acinar emphysema patients. -950HUinsp and -856HUexp correlated well with FEV1/FVC in centriacinar emphysema (r=-0.61, p=0.036; r=-0.65, p=0.021). In the RRAVC map, LVA0.8 (colored blue) was well-matched with low attenuation (emphysema) regions, demonstrating decreased ventilation (air volume change) when compared with adjacent normal lung.
CONCLUSION
RRAVC map correlates well with FEV1 and demonstrates various ventilation patterns in the LAAs on CT in COPD and the proposed LVA0.8 may provide additional functional information at an acinar scale, supplementing LAAs in quantitative CT scans.
CLINICAL RELEVANCE/APPLICATION
Relative regional air volume change map using mass preserving registration technique may be useful for the explanation of different pathophysiology of the LAAs in COPD.
Conference presentation
Application of Image Registration Based Local Displacement Measurement (Lung Motionography) for the Assessment of Lung Fibrosis
Date presented 11/29/2016
Radiological Society of North America (RSNA) Scientific Assembly and Annual Meeting, 11/27/2016–12/02/2016, Chicago, Illinois
PURPOSE
Elastography has been used for the assessment of fibrosis in the liver and breast. However, ultrasonography is not easy to be applied for the fibrosis in the lung parenchyma due to the air. In this study, we applied image registration based local displacement information from expiration to inspiration to grade the degree of fibrosis in the idiopathic interstitial lung disease such as usual interstitial pneumonia and nonspecific interstitial pneumonia.
METHOD AND MATERIALS
10 normal and 18 idiopathic interstitial lung disease (13 IPF and 5 NSIP) subjects were included in our study. VIDA Apollo software (Coralville, Iowa) and mass preserving image registration technique were used to compute displacement vectors of local lung regions at an acinar scale. Three-dimensional displacements and dorsal basal displacements were normalized by the cubic root of global lung volume change from expiration to inspiration CT scans. Displacements and volume changes in the upper and lower lobes and the whole lung are compared between three groups using analysis of variance test (ANOVA).
RESULTS
IPF and NSIP were not differentiated by volume changes of the whole lung or upper and lower lobes, whereas lower lobe air volume change were smaller in both IPF and NSIP than normal subjects (p=0.02, p=0.001). In the whole lung, dorsal basal displacement was smaller in ILD and normal subjects (p=0.035), while three-dimensional displacement was not different between the groups. Three-dimensional and dorsal basal displacement was smaller in the lower lobes of IPF subjects than NSIP (p=0.044) and normal (p=0.006) subjects.
CONCLUSION
Lung motionography using image registration based dorsal basal displacement in the lower lobe may be used for the understanding of the structure-function relationships in fibrotic lung disease.
CLINICAL RELEVANCE/APPLICATION
Image registration based local displacement information may help us to assess the degree of lung fibrosis and to make a diagnosis in the fibrotic lung disease.