• DocumentCode
    3239466
  • Title

    Automated assessment of breast cancer lymphatic infiltration by H&E image

  • Author

    Yung-Lung Kuo ; Chien-Chuan Ko ; Ming-Ji Lee

  • Author_Institution
    Dept. of Comput. Sci. & Inf. Eng., Far East Univ., Tainan, Taiwan
  • fYear
    2012
  • fDate
    14-16 Aug. 2012
  • Firstpage
    202
  • Lastpage
    205
  • Abstract
    The occurrence of invasive breast cancer and lymphatic infiltration is highly relative. Although medical screening and imaging technology has been considerable progress, there are still many patients with metastasis phenomenon of axillary lymph gland at the first inspection of breast cancer. In order to avoid breast cancer cells transferring by lymphatic system, physicians recommend breast cancer patients lymphadenectomy. However, it is estimated that about 70% in early breast cancer patients undergoing unnecessary lymphadenectomy. At present, lymphatic infiltration examination is performed via microscopes and concludes the results by physician experience and expertise. It is time consuming, labor intensive, and prone to inter-observer discrepancy. Therefore, in this article, we present a multistage model of automatic system of lymphatic infiltration assessment. And the experimental results not only prove the agreement of the lymphatic infiltration between the proposed scheme and expert pathologist but also confirmed that this system can greatly reduce the clinician´s burden.
  • Keywords
    cancer; medical image processing; H&E image; automated assessment; axillary lymph gland; breast cancer lymphatic infiltration; imaging technology; interobserver discrepancy; invasive breast cancer; lymphatic infiltration; lymphatic system; medical screening; Accuracy; Breast cancer; Feature extraction; Glands; Image segmentation; breast cancer; lymphatic infiltration; scale-invariant feature transform;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Information Security and Intelligence Control (ISIC), 2012 International Conference on
  • Conference_Location
    Yunlin
  • Print_ISBN
    978-1-4673-2587-5
  • Type

    conf

  • DOI
    10.1109/ISIC.2012.6449741
  • Filename
    6449741