Background Chronic respiratory diseases represent a large group of non-communicable diseases that are a leading cause of mortality and morbidity globally. Many of the methods utilized to assess airway simplification in experimental models of the conditions are overly time-consuming and are sensitive to inter-operator biases, necessitating the need for unbiased and efficient tools to supplement analyses. Methods We propose a semi-automated method to quantitate the characteristics of large terminal respiratory airways and alveoli that uses free image-processing software (Fiji). We aimed to develop and test this method in a mouse model of bronchopulmonary dysplasia (BPD), a disease of blunted airway and pulmonary vascular development that remains a leading cause of mortality among preterm infants. Optimal macro parameters were determined with a test set of images from postnatal day 14 (P14) mice exposed to acute postnatal hyperoxia by determining which area and circularity values best correlated with mean linear intercept (L M ). Validation was performed on a separate set of images from P7 mice subjected to the same hyperoxic model of BPD. Results Both alveolar duct (r: 0.7866, p = 0.0359) and alveolar (r: 0.9475, p = 0.0012) area correlated with L M measurements from the test set. Using our method on a validation dataset, we demonstrate that hyperoxia-exposed mice possess fewer, enlarged alveoli that occupy less total area, as well as enlarged alveolar ducts that occupy a greater proportion of the parenchyma. Conclusions We report a semi-automated method of quantitating the characteristics of large and small terminal respiratory airways. This tool expedites analysis and removes operator bias relative to existing methods. We also demonstrate that L M changes in an acute model of hyperoxia-induced BPD result from both alveolar simplification and inadequate primary septation at the level of the alveolar ducts.
Gainer et al. (Tue,) studied this question.