Automated Immunohistochemical Stains Analysis for Computer-Aided Diagnosis of Parathyroid Disease

摘要

Parathyroid disease has a huge impact on overall health and quality of life. Immunohistochemistry (IHC) is a biological technique, which is useful in diagnosis and prognosis of the parathyroid disorders. The use of IHC as a diagnostic tool brings a substantial methodological problem related to evaluation of stain intensity in micrographs. This paper introduces an image processing approach for automatic IHC stain analysis in micrographs of parathyroid tissue. The introduced approach can be used for computer-aided diagnosis of parathyroid disease as well as for medical research studies in this field. The main novelty of this approach lays in the combination of color deconvolution procedure with a parathyroid cell nuclei localization algorithm, which is based on custom image filtering and circular objects recognition. Accuracy of the proposed approach was verified by comparison with results of experts' evaluation in experiments conducted on micrographs of healthy tissue, adenomas, and hyperplasias with various IHC markers.