Semi-automatic FISH quantification on digital slides

Background HER2 is a transmembrane glycoprotein, a member of epidermal growth factor receptor family. It was documented that the amplification and the over expression of this gene plays an important role in the pathogenesis and in the progression of breast cancer [1]. Nowadays this is one of the most important biomarker and target for breast cancer therapy. 10-30% of invasive breast carcinomas are HER2 positive, the gene over expression occurring in invasive ductal adenocarcinomas and in invasive lobular carcinomas as well. Therefore the assessment of the HER2 receptor status of the formalin-fixed paraffin embedded cancer specimens has a key importance for specifying the appropriate therapy. For prognostic and predictive testing the immuncytochemical and FISH stains are routinely used in the clinical diagnostic. According to the guideline of the American Society of Clinical Oncology/College of American Pathologists [2] when the immunoquantification is not clear FISH stain should be apply to support the diagnosis. By using dual or multicolor probes, the chromosomal aberration and gene sequence modification can be detected in parallel. Several methods are available for the evaluation of gene status in FISH stained samples, such as the analysis of the image histogram [3]. Netten et al. [4] used image processing tools to develop their own algorithm, which automatically detects the region of interest, counts the cell nuclei and finds the spots in each nucleus. Furthermore, there is an algorithm which quantifies 2D and 3D FISH images as well [5]. All of these methods work on images which were recorded by microscope. The conventional microscopic diagnosis process allows to make quantification on the recorded sub-regions of the sample. With the expansion of the digital pathology, we intended to develop a FISH quantification platform, which combines the innovative fluorescence whole slide scanning technology with image processing applications.