The Present State of the Automated Micronucleus Test for Lymphocytes

Publisher Website
Google Scholar
Abstract
This minireview presents the state of the art with respect to automated detection of micronuclei (MN) in binucleated lymphocytes. Emphasis is on an image analysis technique, based on the principles of mathematical morphology (pattern recognition), which combines a personal computer with an image processing board and a board for microscope control. The basic idea behind this procedure is that nuclei plus MN and cytoplasms are analysed separately and sequentially by capturing images from gallocyanin-stained nuclei plus MN and naphthol yellow-S stained cytoplasms from one microscope field by using different filters. Major steps in the identification of nuclei and MN are separation of nuclei and MN from background by determination of periphery of the nuclei and MN, and artefact rejection procedures. After changing the filter, a binary image is constructed from cytoplasms and artefacts. Finally, stored information from selected binucleated objects with/without MN is combined with the cytoplasm image to check whether selected objects belong to the same cytoplasm. The procedure described above allows automated detection of binucleated lymphocytes with or without MN. The current capacity to detect 63% of binucleated cells and 57% of the MN within them is quite acceptable. To avoid false positives, artefact rejection procedures need to be improved before the method can be used routinely.

This publication has 8 references indexed in Scilit: