Summary:
UCLA researchers have developed a novel method to characterize lung tissue biomechanics and applied it to the creation of a computational model to improve understanding and visualization of lung diseases.
Background:
Chronic lower respiratory disease (ex. COPD) was the sixth leading cause of death in the United States in 2021, taking nearly 140,000 individuals. COPD is characterized by damaged lung tissue with altered biomechanical properties and ventilation profiles. Current staging methods using pulmonary function tests are insufficient to determine COPD phenotypes, contributing to nonoptimal treatment identification and implementation. Recent advancements in computed tomography (CT) imaging have led to improved characterization of lung heterogeneity and pulmonary lung function. In addition, lung biomechanics and airflow dynamics modeling have been investigated which can provide patient-specific models. However, such strategies have not been integrated into the patient selection and treatment process. Another key limitation of existing techniques is the inherent assumption of breathing cycle regularity; however, COPD patients cannot breathe regularly, severely impacting the actual efficacy of these techniques. Although biomechanical properties and airflow dynamics of COPD patients are direct indicators of the phenotype, little work has been done to non-invasively measure or model them in a subject-specific manner.
Innovation:
UCLA researchers in the Department of Radiation Oncology have developed a novel model for improving COPD phenotype characterization and response monitoring. This model accomplishes high accuracy and non-invasive lung biomechanical analysis through a novel patient-specific flow structure interaction (FSI) model. This model was generated using data acquired during free breathing, avoiding regular breathing assumptions plaguing existing models. Instead of looking at how stiff lung tissue is in a straight-forward way like before, this new method focuses on how it can stretch and bounce back, which is crucial for understanding the special properties of diseased fibrotic tissue in COPD. . This improved characterization will enable researchers to develop personalized treatments that lead to better treatment response and improved quality of life.
Potential Applications:
• Personalized medicine.
• COPD and other chronic respiratory illness diagnosis and treatment.
• Pulmonary research.
Advantages:
• Non-invasive, hyperelastic modeling.
• Fibrotic tissue biomechanical analysis without sorting artefacts.
• More accurate COPD phenotype characterization.
State of Development:
The inventors have developed and experimentally tested a fluid mechanical model for pulmonary analysis. They have also performed patient studies and demonstrated increased model accuracy when compared to standard non-invasive methods.
Related Papers:
1. Stiehl, Brad, Michael Lauria, Dylan O'Connell, Katelyn Hasse, Igor Z. Barjaktarevic, Percy Lee, Daniel A. Low, and Anand P. Santhanam. "A quantitative analysis of biomechanical lung model consistency using 5DCT datasets." Medical physics 47, no. 11 (2020): 5555-5567.
2. Stiehl, Bradley, Michael Lauria, Kamal Singhrao, Jonathan Goldin, Igor Barjaktarevic, Daniel Low, and Anand Santhanam. "Scalable quorum-based deep neural networks with adversarial learning for automated lung lobe segmentation in fast helical free-breathing CTs." International Journal of Computer Assisted Radiology and Surgery 16, no. 10 (2021): 1775-1784.
Reference:
UCLA Case No. 2022-328
Lead Inventor:
Anand Santhanam, Associate Professor in UCLA’s Department of Radiation Oncology.