Determining Alveolar Dynamics by Automatic Tracing of Area Changes Within Microscopy Videos

Abstract

Alveolar microscopy allows in vivo inspection of changes in alveolar geometry during mechanical ventilation in animal models. It is currently an open question how the size of alveoli relates to cyclic mechano-stimulation during ventilation. Determining the changes in alveolar geometry from recorded videos would provide new insights into alveolar dynamics. But the evaluation of thousands of video frames by hand is not suitable and produces unreliable results. Other groups regardless of the applied imaging acquisition technique therefore restrict analysis to single frames or use global measures extracted from binary images. Here we report on an automated alveolar tracking system that determines relative changes in alveolar size during mechanical ventilation. The observed changes in alveolar size can be correlated with mechanical properties of the respiratory system. Finally variations in alveolar geometry may be used to evaluate and to guide respirator settings in the treatment of critically ill lungs.