Dissertation
An imaging-based subject-specific whole-lung deposition model and its applications
University of Iowa
Doctor of Philosophy (PhD), University of Iowa
Autumn 2023
DOI: 10.25820/etd.006943
Abstract
Particles, both harmful and therapeutic, can be inhaled into the respiratory tracts and deposited throughout the airways. Regional particle deposition in the human respiratory tract is an important biomarker for assessing inhaled drug delivery and susceptibility to environmental pollutants. In human airways, the transport and deposition of particles are determined by particle characteristics, airflow rate, and airway geometry. With large quantitative computed tomography (QCT) lung images acquired by multi-center studies, it is essential to develop an efficient computational method to simulate and predict particle deposition over a population and apply machine learning to examine the relationships between clinical, functional and computational fluid dynamics (CFD) endpoints/biomarkers in subpopulations (clusters, subtypes or subgroups). Thus, the first objective of the research is to develop a CT-based subject-specific whole-lung deposition model. The three-dimensional (3D) airway structure was reduced into onedimensional (1D) along the axial direction to reduce the computational expense while keeping accuracy. A volume-filling algorithm was applied to extend the CT-resolved branches to a complete whole-lung 1D conducting airway tree. A dichotomous model of acinar airways was attached to the end of each terminal bronchiole. The regional ventilation was calculated using the sum of squared tissue volume difference (SSTVD) using the image-registration method. The airflow and pressure distribution were estimated using a 1D computational fluid and particle dynamics (CFPD) lung model. Particle deposition formulae for diffusion, sedimentation, and impaction mechanisms were implemented to estimate the deposition fractions in each airway segment. The 1D CFPD prediction were validated against CT and single-photon emission computed tomography (SPECT) chronic obstructive pulmonary disease (COPD) data and showed a good agreement compared with in-silico and in-vivo results.
The second objective is to apply the above approach to post-COVID-19 patients. A total of 245 subjects were recruited for the study, with 145 participants who tested positive for SARSCOVID-19 and 100 healthy non-smokers participants who were recruited as the healthy control group. A contrastive learning model using CT scans at TLC and RV volumes was implemented to differentiate post-COVID-19 subjects from the healthy control group and identify postCOVID-19 clusters. The particle deposition fraction in each cluster was analyzed to assess the impact of the disease. The results showed distinctive features in individual post-COVID-19 clusters: Cluster 1 and Cluster 2. Cluster 1 was a female-dominant group with small airway disease and high heterogeneity of ventilation, and Cluster 2 was characterized by older individuals with fibrotic lung. Cluster 1 subjects exhibited elevated resistance in the distal region, leading to a greater proportion of particle deposition in the proximal airways. Conversely, Cluster 2 subjects relocated resistance to the proximal region, enabling particles to penetrate deeper into the lungs.
The third objective is to examine the deposition patterns between two obstructive diseases by using imaging-based clustering classification. Patients with asthma (n=6) and COPD (n=8) with varying severity levels were examined for this study. Airway structure change and CFPD metrics were analyzed and compared against CT/SPECT data. Compared with spirometry classification, imaging-based clustering method had higher sensitivity capturing the difference in biomarkers in sub-phenotypes of asthma and COPD.
Details
- Title: Subtitle
- An imaging-based subject-specific whole-lung deposition model and its applications
- Creators
- Xuan Zhang
- Contributors
- Ching-Long Lin (Advisor)Eric A Hoffman (Committee Member)Casey M Harwood (Committee Member)Shaoping Xiao (Committee Member)
- Resource Type
- Dissertation
- Degree Awarded
- Doctor of Philosophy (PhD), University of Iowa
- Degree in
- Mechanical Engineering
- Date degree season
- Autumn 2023
- Publisher
- University of Iowa
- DOI
- 10.25820/etd.006943
- Number of pages
- xiv, 114 pages
- Copyright
- Copyright 2023 Xuan Zhang
- Grant note
- This work was supported in part by NIH U01-HL114494 and R01-HL168116, FDA U01-FD005837 and ED P116S210005.
- Language
- English
- Date submitted
- 12/03/2023
- Description illustrations
- Illustrations, tables, graphs, charts
- Description bibliographic
- Includes bibliographical references (pages 105-114).
- Public Abstract (ETD)
The lungs serve as gateways for particles during the act of breathing. These particles can be helpful, like medicine, or harmful like pollutants. Understanding the precise location of these particles within our lungs is vital for comprehending the functioning of inhaled medications and the potential hazards posed by environmental pollutants.
To study this, 3D CT images are collected to provide detailed views of the lung. For a more efficient way to study the particle transportation in human lung for a large group of people, the 3D structure of the airways is simplified into a 1D line. Using this simplified model, we could mimic airflow and see where particles land in the airways.
The first goal was to build a reliable model based on CT scans that could predict where particles end up in the lungs. This model was cross-checked against SPECT ventilation imaging data to ensure its accuracy.
Next, they studied lungs of those who had COVID-19 to see if the disease changed how particles settled. We compared the lungs of 145 COVID-19 patients with 100 healthy individuals. Through this, we identified certain deposition patterns unique to post-COVID lungs.
Lastly, we studied how particles behave in different lung diseases. We examined 14 patients with conditions like asthma and COPD to see if their airways affected particle behavior.
In summary, this research develops a whole-lung deposition model and aims to give a clearer understanding of how particles move and settle in our lungs, helping with drug development and environmental health assessment.
- Academic Unit
- Mechanical Engineering
- Record Identifier
- 9984546749602771
Metrics
10 File views/ downloads
32 Record Views