Semiautomated Computer-assisted Image Analysis to Quantify 3,3'-diaminobenzidine Tetrahydrochloride-immunostained Small Tissues

This work aimed to develop a technique to measure stained areas in images from sample tissue sections, namely when the structure of interest does not fill the entire image field of the microscope. We propose a semiautomated computer-assisted image analysis (SACAIA) method in which brightfield color images of 3,3'-diaminobenzidene tetrahydrochloride (DAB)-stained antigens are converted to their blue component and boundaries are delineated to extract the object of interest. The number of pixels of a defined color (elicited by DAB) is counted and used to measure the stained area relative to the total area of the tissue under study. The percentages of area stained with adenosine A(1) receptor were 40.76+/-2.08 and 42.44+/-2.26% for manual analysis and SACAIA, respectively (P=0.582). A strong linear correlation of A(1) receptor quantification was found (r=0.98, P<0.001, and 95% CI=0.97 to 0.99 for manual method; r=0.99, P<0.001, and 95% CI=0.98 to 0.99 for SACAIA method). The extent to which misclassification affected staining quantification was evaluated by Bland-Altman analysis, indicating that this method can be applied accurately to quantify the immunohistochemical staining area (occupied by a specific antigen) in small sample tissues that do not fill the entire image field of the microscope.