Lung damage analyzed by machine vision on tissue sections of mice.

Abstract

The inhalation of environmental toxicants can induce lung damage. Many methods are currently available to analyze lung tissue damage and are based on empirical visual judgment; however, the accuracy of the assessments are influenced by individual differences among pathologists. Here, we establish new methods of analysis for lung tissue sections based on machine vision and verify this new automatic high-flux method with the model of mice inhaling aqueous aerosol with different concentrations of CdCl2 (0, 1, 3, 5mM 2h/day) for 7days through analyses of pulmonary porosity, mucus, pneumonia and co-localized staining. Additionally, the correlation analysis among the concentrations of CdCl2 in aqueous aerosol, the high-flux analyses and empirical visual judgment methods demonstrate the practicality of the new automatic method. The comparison between the high-flux analyses and the empirical visual judgment methods demonstrates the superiority of the new automatic method. In the future, these new automatic high-flux analyses based on machine vision could be conducive to pulmonary histology and pathology research.