Distribution analysis of alveolar size by digital imaging

To estimate the size distribution of the alveoli of the lung, an automated method has been developed by using a digital imaging technique, which acquires, processes, converts, and edits lung histologic images from the microscope. The binary images are analyzed by measuring the spacing between alveolar walls along a grid of parallel lines. The resultant intercept data are fit with a bimodal distribution consisting of two lognormal distributions. We were thus able to distinguish the smaller distribution that is comprised primarily of alveolar intercepts. The effect of the binary image-editing features, and the goodness-of-fit of the distribution curve are discussed. The analysis provides a dimensional improvement compared to the information that is normally obtained from lung histologic sections.