Avaliação do muco e infiltrado inflamatório em pulmões de ratos submetidos á exposição crônica ao herbicida ácido diclorofenóxiacético (2,4-D), utilizando-se do processamento digital de imagens

Abstract

2,4-Dichlorophenoxyacetic acid, known and marketed as 2,4-D, is one of the most widely used herbicides in the agro-industry for the control of various weeds in plantations. Currently, the areas of health and computing have been studied with the purpose of improving parameters related to medical diagnosis and one of the methods used is Digital Image Processing, characterized by the ability to extract patterns from a set of samples to acquire knowledge of automatically. To date, few studies have been found in the literature on the use of Digital Imaging methods in the histopathological analysis of the respiratory epithelium. The objective of this study was to evaluate histopathologically the pulmonary epithelium in rats submitted to chronic exposure of the 2,4-D herbicide, using histopathological analysis and digital image processing. A total of 40 male adult Wistar rats were divided into four groups: Control Group (GC), Low Concentration Group (GBC), Mean Concentration Group (GMC) and High Concentration Group (GAC) by means of two boxes connected to ultrasonic nebulizers , responsible for spraying the herbicide in different concentrations during a six-month exposure period. After this stage, animals were euthanized for the collection and study of lung tissue for histopathological analysis and digital image processing. Histopathological evaluation of the pulmonary epithelium was performed using the Sigma Stat® version 2.0 program, considering for the normal distribution the analysis of variance followed by Tukey and non-parametric tests, such as Kruskal-Wallis followed by the Dunn test, where the differences considered statistically significant implied a p less than 5% (p <0.05). The Digital Image Processing considered for evaluation the inflammatory infiltrate and mucus in the lungs, the computational analyzes performed through several algorithms of pattern recognition, in orderto verify the existence of patterns in the different data collected. Pre-processing was performed on the images for the segmentation of each cell and the counts of Good and Bad cells were performed through digital image processing. Then the segmented cells were classified as "Good or Bad", following some metrics obtained from the cells that appear in the images and some empirically defined parameters. In the histopathological analysis there was no statistically significant difference between the groups exposed to the herbicide, compared to the control group, but in the computational analysis the existence of intact and healthy alveoli and injured alveoli was demonstrated, demonstrating that the practice of digital image processing is feasible for improve medical diagnosis and research related to inflammation and mucus. The study concludes that alveolar injury is present due to the inhalation exposure of the herbicide 2,4-D in relation to its toxicity classification and exposure period. The presence of a low significance index of lesioned alveoli was observed in healthy alveoli due to the use of digital image processing, which was fundamental to improve the traditional method of histopathological analysis, presenting a clue that leads us to consider the absence of a definite and unique pattern of alveolar injury