Identification of asynchronous effect via pressure-volume loop reconstruction in mechanically ventilated breathing waveforms

Patient-specific lung-mechanics during mechanical ventilation (MV) can be modelled via using fully ventilated/controlled waveforms. However, patient asynchrony due to spontaneous breathing (SB) effort commonly exists in patients on full MV support, leading to variability in breathing waveforms and reducing the accuracy of identified, model-based, and patient-specific lung mechanics. This study aims to extract ventilated breathing waveforms from affected asynchronous breathing cycles using an automated virtual patient model-based approach. In particular, change of lung elastance over a pressure-volume (PV) loop is identified using hysteresis loop analysis (HLA) to detect the occurrence of asynchrony, as well as its type and pattern. The identified HLA parameters are then combined with a nonlinear mechanics hysteresis loop model (HLM) to extract and replicate the ventilated waveforms from the coupled asynchronous breaths. The magnitude of asynchrony can then be quantified using an energy dissipation metric, Easyn, comparing the area difference of PV loops between model-reconstructed and original breathing cycles. A proof-of-concept study is conducted using clinical data from 2700 breathing cycles of two patients exhibiting asynchrony during MV. The reconstruction root mean square errors are within 5-10% of the clinical data for 90% of the cycles, indicating good and robust reconstruction accuracy. Estimation of Easyn shows significant asynchrony magnitude for Patient 1 with Easyn greater than 10% for over 50% breaths, while asynchrony occurrence for Patient 2 is lower with 90% breaths at Easyn < 10%, which is a minimal asynchrony magnitude. These results match direct observation, thus validating the ability of the virtual patient model and methods presented to be used for a real-time monitoring of asynchrony with different types and magnitudes, which in turn would justify prospective clinical tests.