A Simulation-Based Mechanical System-Identification Framework for Non-Invasive Lung Diagnostics and Personalized...

Area of Science:

• Biomedical Engineering
• Pulmonary Medicine
• Systems Biology

Background:

• Current lung disease diagnostics (imaging, pulmonary function tests) lack regional mechanical detail.
• This limits the detection of early pathological changes and mechanical heterogeneity.
• Existing methods provide static structural data or global indices, missing regional insights.

Purpose of the Study:

• To introduce and evaluate a conceptual diagnostic framework for the lung based on mechanical system identification.
• To assess the feasibility of inferring internal lung mechanical properties from external responses.
• To explore a non-invasive approach for assessing regional lung mechanics.

Main Methods:

• Treated the lung-thorax system as an identifiable viscoelastic dynamical system.
• Developed a multi-degree-of-freedom mechanical model of the lung.
• Used sensitivity and Fisher-information analysis to confirm parameter identifiability.
• Performed inverse fitting with simulated noisy frequency response data.

Main Results:

• Confirmed structural identifiability of regional stiffness parameters.
• Accurately recovered simulated stiffness perturbations from noisy data.
• Achieved complete separation of simulated pathological configurations in classification experiments.
• Demonstrated theoretical feasibility of identifying regional mechanical alterations.

Conclusions:

• A diagnostic paradigm based on mechanical system identification is theoretically feasible.
• This approach offers a complementary, non-invasive method for assessing regional lung mechanics.
• Quantifying regional stiffness and heterogeneity can aid personalized pulmonary rehabilitation.