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

Current diagnostic assessments of lung disease rely primarily on medical imaging and global pulmonary function tests, which either provide static structural information or collapse complex regional behavior into global indices. As a result, important information about regional mechanical heterogeneity and early pathological changes may remain inaccessible. In this work, we introduce a conceptual diagnostic framework for the lung based on mechanical system identification and evaluate its feasibility using simulation-based analysis. Rather than directly imaging internal lung structure, the lung–thorax system is treated as an identifiable viscoelastic dynamical system whose internal mechanical properties can be inferred from its response to controlled external excitation. A multi-degree-of-freedom mechanical representation of the lung was developed to capture the dominant low-frequency behavior of the chest wall and major lung regions. Sensitivity and Fisher-information analysis confirmed the structural identifiability of regional stiffness parameters (FIM eigenvalues λ1 = 1.75 × 10−9 and λ2 = 8.91 × 10−10). Inverse fitting experiments accurately recovered simulated stiffness perturbations (e.g., k01 = 240 → 239.5; k02 = 154 → 159.5) from noisy frequency response data, while classification experiments achieved the complete separation of simulated pathological configurations in an idealized synthetic scenario, supporting theoretical discriminability rather than clinical performance claims. These findings demonstrate the theoretical feasibility of a diagnostic paradigm in which regional lung mechanical alterations can in principle be identified through mechanical system identification rather than direct imaging, thereby suggesting a complementary approach for a non-invasive assessment of regional lung mechanics from externally measured responses. By quantifying regional stiffness and mechanical heterogeneity, this framework may also support the personalization and monitoring of pulmonary rehabilitation strategies in chronic respiratory disease.