Book chapter
Machine learning and in silico methods
Inhaled Medicines: Optimizing Development through Integration of In Silico, In Vitro and In Vivo Approaches, pp.375-390
Elsevier
01/01/2021
DOI: 10.1016/b978-0-12-814974-4.00013-4
Abstract
Abstract This chapter reviews the techniques and strategies for identifying subpopulations (clusters) characterized by distinct lung features via machine learning and using cluster information to guide in silico computational fluid and particle dynamics (CFPD) analysis for the design of future inhaled drug delivery methods. We first review the collaborative efforts of collecting imaging, genetic, clinical and biological data sets for large cohorts of healthy, asthma and chronic obstructive pulmonary disease (COPD) subjects to investigate the heterogeneous nature of lung disease. We then focus on imaging-based phenotyping due to its quantitative nature that sensitively captures lung structural and functional alternations at both local (segmental/parenchymal) and global (lobar/lung) scales. Machine learning is then applied to identify imaging clusters for asthma and COPD patients. We select cluster archetypes to perform CFPD analysis and use CFPD-derived variables to interpret the link between cluster-specific alterations and particle depositions in the human lungs. Finally, we discuss the prospect of employing machine learning, physics-based learning and deep learning complementarily toward precision medicine.
Details
- Title: Subtitle
- Machine learning and in silico methods
- Creators
- Ching-Long LinEric A. Hoffman - University of IowaStavros C. Kassinos - University of Cyprus
- Resource Type
- Book chapter
- Publication Details
- Inhaled Medicines: Optimizing Development through Integration of In Silico, In Vitro and In Vivo Approaches, pp.375-390
- Publisher
- Elsevier
- DOI
- 10.1016/b978-0-12-814974-4.00013-4
- Language
- English
- Date published
- 01/01/2021
- Academic Unit
- Roy J. Carver Department of Biomedical Engineering; Radiology; Mechanical Engineering; Internal Medicine
- Record Identifier
- 9984573938902771
Metrics
1 Record Views