Transactions of Japanese Society for Medical and Biological Engineering
Online ISSN : 1881-4379
Print ISSN : 1347-443X
ISSN-L : 1347-443X
Proceedings
Image processing approach for differentiation between flat epithelial atypia and columnar cell change/hyperplasia of the breast
YOSHIKO YAMASHITASHU ICHIHARASUZUKO MORITANIMASAKI HASEGAWAAKARI IWAKOSHI
Author information
JOURNAL FREE ACCESS

2015 Volume 53 Issue Supplement Pages S231_03

Details
Abstract
Flat epithelial atypia (FEA) has received attention as one of precursor lesion of breast cancer in recent years. It is practically important to distinguish between FEA and columnar cell change/hyperplasia (CCC/CCH) with similar histology. However, medical decisions are quite often subjective decision-making. The effectiveness evaluation of image analysis methods for diagnosis of FEA and CCC/CCH has not yet been reported. We focused on the characteristic that many of FEA nuclei have a more rounded configuration than CCC ones, and investigated whether it is possible to distinguish FEA from CCC/CCH based on the degree of circularity of nuclei. We converted Hematoxylin and Eosin (H&E) stained glass slides into Whole Slide Images (WSI) by Hamamatsu NanoZoomer 2.0- HT. Pathologists manually selected the regions of interest (ROI) in WSI. For each ROI, the original image was converted to a color space free from batch effects. We extracted nuclei. Then, for each extracted nucleus, the degree of circularity was measured. We confirmed that the degree of circularity of nuclei was an effective index to distinguish FEA from CCC/CCH.
Content from these authors
© 2015 Japanese Society for Medical and Biological Engineering
Previous article Next article
feedback
Top