Machine scoring of Her2/neu immunohistochemical stains

John W Bishop, Raphael Marcelpoil, Joachim Schmid

Research output: Contribution to journalArticle

7 Scopus citations

Abstract

OBJECTIVE: To establish the feasibility of a machine scoring method for her2/neu immunohistochemistry in samples of breast carcinoma. STUDY DESIGN: A total of 65 consecutive cases of breast carcinoma with immunohistochemical staining for her2/neu by the Herceptest™ (Dako Corp., Carpinteria, California, U.S.A.) method (DAB chromogen with hematoxylin counterstain) were analyzed using an Extended Slide Wizard™ (Tripath Imaging, Inc., Burlington, North Carolina, U.S.A.) workstation running prototype software. Representative fields of view from the positive control, negative control and up to 10 fields from the stained tumor sample were captured interactively with a phased alternating line 3 CCD color camera. To determine the amount of specific membrane staining, chromogen separation of nuclear counterstain and membrane-positive stain was performed based on their respective absorption coefficients in the three color channels. The amount of specific membrane staining was scored based on a training set covering the range from 0 to 3+ staining scores according to Dako. Manual scores of 2+ were tested for amplification by fluorescence in situ hybridization. RESULTS: The automated scoring results correlated highly with the manual scores obtained per the Herceptest™ (Dako) instructions (R2 > .92). The results were obtained in real time in the interactive mode. CONCLUSION: Machine scoring of immunohistochemical stains is practical, rapid and inherently reproducible, especially for samples with 1+ and 2+ manual scores.

Original languageEnglish (US)
Pages (from-to)257-262
Number of pages6
JournalAnalytical and Quantitative Cytology and Histology
Volume24
Issue number5
StatePublished - Oct 2002
Externally publishedYes

Keywords

  • Breast cancer
  • Computer-assisted diagnosis
  • Her2 proto-oncogene protein
  • Immunohistochemistry

ASJC Scopus subject areas

  • Anatomy
  • Histology
  • Cell Biology

Fingerprint Dive into the research topics of 'Machine scoring of Her2/neu immunohistochemical stains'. Together they form a unique fingerprint.

  • Cite this