Morphological analysis of mouse pulmonary acini extracted from synchrotron micro-CT images with a multiscale-based segmentation algorithm

A semi-automatic segmentation method for extracting the complete structure of pulmonary acinus from synchrotron micro-CT images was developed on the basis of the multiscale and hierarchical nature of lung structures. To verify the potential of the method for the statistical analysis of acinar morphology, we here measured and analyzed the morphological characteristics of acini with the method specialized for acinus structure to the micro-CT images of mouse lungs. The acinus morphology that we focused on was the cluster of acini belonging to the same terminal bronchiole, which is obtained by our multiscale-based segmentation algorithm. The volume and surface area of each acinus were measured by counting the voxels in the acinus and the triangular patches created on the wall surface of the acinus, respectively. A treestructure skeletonization algorithm was used for the analysis of the airway path in the acinus. Multiple clusters of isolated acini belonging to the same terminal bronchiole were obtained without floating voxels. The number of clusters was 3 and each cluster had 8, 9, and 9 acini. The volume and surface area of acinus were 0.125±0.09 mm³ (mean±SD.) and 10.09±6.70 mm³, respectively. The range of the volume of acinus was from 0.05 to 0.45 mm³ and the peak value was 0.1 mm³. In each cluster of acini, the maximum of acinar volume was found at the end of terminal bronchiole. The pathway lengths increased with the generation of airways. As for the average and the peak of acinar volume, the values reasonably agree well with the histology- based estimates of previous researches.