A new Bayesian method reduced the ventilation distribution in patients with cystic fibrosis (CF), as published in Respiratory Physiology & Neurobiology. This method used fewer breath cycles than lung clearance index (LCI) and correlated with ventilation magnetic resonance imaging (MRI), and was also important for clinical decision-making.
Determining ventilation heterogeneity is essential because it is an early clinical sign of lung diseases such as CF. Several methods like MRI with hyperpolarized gas ventilation are used for best-assessing lung function, however, shorter test times and more straightforward methods are necessary.
Ventilation heterogeneity assessment requires a multiple breath washout test (MBW), and it is predominantly reported as LCI. Although LCI is a sensitive method for detecting early disease and correlates with the clinical condition, it uses only a little of the gas collected in the MBW. Also, the time for the data collection process is still lengthy for across-the-board use.
On the other hand, MRI with hyperpolarized gas ventilation is a method for assessing the etiology and severity of lung disease. The authors also previously found a correlation between LCI and MRI determinants in patients with CF. Therefore, this situation allows for a new method, like the Bayesian computational software, and correlates it with MRI findings.
Read more about CF testing
“The benefit of the method presented here is that uncertainty in the predictions is readily quantified by the Bayesian methodology, and where there are multiple viable solutions all are given weight relative to their probability of explaining the observed data,” the authors explained.
The results showed that ventilation distribution in the lungs from the MBW aligns with the data from the ventilated MRI. Also, the model predictions showed better results than the ones of LCI.
“This method also has an advantage over LCI in that it incorporates the interdependence of the inferred [functional residual capacity], dead-space and ventilation heterogeneity,” they said. Moreover, the model is robust for up to one-fifth of the termination threshold. That is, washout time can be reduced by approximately three-fourths, considering that both sensitivity and specificity reduce when augmenting the termination threshold.
The authors concluded that this model with an MBW can accurately predict ventilation distribution, enabling a better understanding of clinical findings and reducing test time for future expediency. Likewise, the study confirmed the predictions with ventilation imaging, which adds value to the findings.
“This work is also the first example of a physiological model fitted to patient washout data using Bayesian parameter estimation, which will provide clinicians with all important estimates of uncertainty in physiological inferences,” they said,
Whitfield C, Horsley A, Jensen O, et al. Model-based bayesian inference of the ventilation distribution in patients with cystic fibrosis from multiple breath washout, with comparison to ventilation MRI. Respir Physiol Neurobiol. 2022;302:103919. doi:10.1016/j.resp.2022.103919