Novo método para análise automatizada da organização do tecido tendíneo corado por Hematoxilina-Eosina (H&E)

Introduction: Tendons are fibrous organs, histologically formed by tenocytes and an extracellular matrix with a hierarchical collagenous structure. In cases of injuries, tendinopathies, changes in tendon morphology arise, such as: hypercellularity, changes in phenotype and loss of uniformity in the spatial orientation of tenocytes. Whereas this important correlation between the tendinous histological changes and their biomechanical properties, the scarcity of accurate methods of tendon analysis and the need for automation in the study of tendinopathies, the development of analysis methods that explore tendinous microarchitecture is promising. Objective: Validate a new method for automated analysis of the tendon tissue organization stained with hematoxylin-eosin. Methodology: This is a methodological study for routine validation of automated analysis of the organization of tendon tissue. Histological slides from Wistar rats submitted to partial tenotomy of the right calcaneus tendon were used, keeping the left one as a control. After 35 days of repair, the animal was euthanized for collection and microscopic analysis of the tendons stained with hematoxylin-eosin. Images of four histological fields (20x objective) were captured from the middle third of each tendon. The images were processed using the ImageJ software, using an automated computational routine. Subsequently, an analysis of the efficiency of the method in detecting the tenocyte nucleus and its use in the comparison between control and injured tendons was performed, considering three variables: cellularity, aspect ratio (AR) and circular standard deviation (CSD). The CSD was calculated using circular statistics and the data were compared using a t test, assuming p <0.05. All calculations were performed using Microsoft Excel 2010 software (Microsoft Corporation, USA). Results: The method showed an efficiency of 87.27% to 96.96% in the isolation of tenocytes. Concerning the comparison between control and injured tendon images, the method showed an average CSD of 6.19 ± 0,96 in the control group and 10.46 ± 1,71 in the injured group (p <0.05); mean cellularity of 61.50 ± 12,23 in the control group and 111.00 ± 4,08 in the injured group (p <0.05); average aspect ratio of 4.01 ± 0,43 in the control and 3.33 ± 0,35 in the injured group (p <0.05). Conclusion: The results of the present study showed that the new method developed for automated analysis of the organization of tendon tissue is efficient in detecting the tenocytes nuclear profiles, as well as in the comparison between healthy and injured tendons.