Combining patient-specific airway models with acinar asymmetry in simulations of multiple breath washout (MBW)

Background: Patient-specific airway models enable accurate simulation of lung ventilation and gas transport. These can be used to test mechanisms behind MBW indices such as LCI and Scond. Similarly, asymmetric branching models of the human acinus can reproduce more realistic washout curves and Sacin values. We have developed a computationally efficient method to combine these two models to perform patient-specific MBW modelling

Method: Distal conducting airways in the patient-specific airway tree were replaced by symmetric trees with same effective volume and resistance. A reduced asymmetric acinar model was added.

Results: The models were extensively tested on a healthy subject-specific geometry in a range of scenarios. We tested various averaging methods on the same tree with constrictions artificially applied. Replacing sub-trees by Horsfield order (bottom-up) that satisfied a constraint measuring their flow variance proved a robust method of averaging. Scond and LCI generally reduced as more airways were included in the averaged section. Inclusion of asymmetric acinar models gave rise to more realistic Sacin values and phase-II profiles and increased LCI, but did not significantly impact Scond.

Conclusions: These techniques enable realistic simulation of all aspects of MBW by modelling transport at all scales. This method reduces the computational complexity while retaining patient-specificity.