Automated assessment in HER-2/neu immunohistochemical expression of breast cancer

HER-2/neu, a protein often giving higher aggressiveness in breast cancers, has been shown that if the gene is expanded for some reason, the Her-2 protein produced by the cells will be over-expressed to enhance the cancer cells reproduced ability, the prognosis will be also relatively less, too. The HER-2 immunohistochemical stained provides a simple and reliable method for pathologist in clinical diagnosis and prognosis. However, it is labor intensive and time consuming in checking a lot of immunohistochemical images. In addition, there is often an inconsistency between different observers caused by individual experience and knowledge. Therefore automated computer-aided evaluation had been suggested to provide a consistent and objective evaluation of HER2 expression and reduce the burden of pathologist in clinic. In this paper, we present an automated method about the quantitative assessment of HER2 expression of immunohistochemical stained images. The proposed system efficiently extracts nuclei of interest including positive stained nuclei and negative stained nuclei. By applying a series of image processing, including color pixel classification, nuclei segmentation, cell membrane extracting, measures of cell membrane staining intensity and completeness, HER-2 expression can be assessed. This evaluation provides pathologist significantly with a good reference for diagnosis and prognosis.