论文部分内容阅读
Purpose: Micronuclei(MN)scoring is a simple and well-known test which has many applications in studying genetic damages in cell biology,radiation biodosimetry and biomonitoring.The Cytokinesis Block Micronucleus Assay(CBMNA)is a widespread method for measuring MN to examine genetic lesions in cultured human lymphocytes.The MN scoring is usually performed manually by a specialist,which is a time-consuming,tedious and subjective task.Automatic analysis and scoring of MN,facilitates this process and make it objective and more accurate.Methods: In order to create micronuclei,whole blood sample with culture medium were placed on a water phantom at 80 cm of a Co-60(Phoenix Theratron)source.Cytochalasin B was added after irradiation and cells were fixed and stained on clean glass slides.A high resolution digital camera coupled to an Olympus microscope was used for imaging.Totally,200 images were captured for implementing the proposed algorithm.In this work,a software was compiled that utilized image processing techniques and proceeded the following steps: overall segmentation of nuclei,detection of binucleated(BN)objects,extraction of BN cells,detection of MN in the detected BN cells.Results: The software was able to extract BN cells with invisible or vague cytoplasm in peripheral blood smear.To evaluate the accuracy of the proposed automatic technique,compared with manual method,acquired scores from both techniques were compared.Manual scoring was conducted by two experts.In this study,we achieved 98%and 91%sensitivity and specificity respectively for detecting BN cells.Similarly 82%and 99%for detecting MNs.Conclusions: Automation of MN analysis in peripheral blood smear of lymphocytes based on the CBMNA will provide faster and more credible results with minimizing subjective errors and MN identification.