Quantitative immunocytochemistry using an image analyzer. I. Hardware evaluation, image processing, and data analysis.

Abstract

In this review we describe how video-based image analysis systems are used to measure immunocytochemically labeled tissue. The general principles underlying hardware and software procedures are emphasized. First, the characteristics of image analyzers are described, including the densitometric measure, spatial resolution, gray scale resolution, dynamic range, and acquisition and processing speed. The errors produced by these instruments are described and methods for correcting or reducing the errors are discussed. Methods for evaluating image analyzers are also presented, including spatial resolution, photometric transfer function, short- and long-term temporal variability, and measurement error. The procedures used to measure immunocytochemically labeled cells and fibers are then described. Immunoreactive profiles are imaged and enhanced using an edge sharpening operator and then extracted using segmentation, a procedure which captures all labeled profiles above a threshold gray level. Binary operators, including erosion and dilation, are applied to separate objects and to remove artifacts. The software then automatically measures the geometry and optical density of the extracted profiles. The procedures are rapid and efficient methods for measuring simultaneously the position, geometry, and labeling intensity of immunocytochemically labeled tissue, including cells, fibers, and whole fields. A companion paper describes non-biological standards we have developed to estimate antigen concentration from the optical density produced by antibody labeling (Nabors et al., 1988).