The Development of Workflows for Image Analysis of Tissue Stained with Specific Antibody Markers
Honorable Mention - Pathology InformaticsMark, R, Verardo ; Dako North America, Inc.;
Content:
The foundation of image analysis applications developed for routine diagnostic pathology markers are algorithms for quantitative image analysis. These applications are dependent upon tissue specimens stained using a standardized and automated workflow model. The workflow model is comprised of antibody and visualization kits, automated immunohistochemistry staining protocols, and image analysis applications. Data output, or image measurements, are captured in standardized reports for target antibodies. Marker-specific workflows can be developed as tools addressing a major issue in the pathology laboratory, standardization.
Technology:
Automated staining was carried out using a Dako Autostainer Plus Link (Dako, Carpinteria, CA, USA) instrument and digital microscopy was performed with the Dako ACIS® III Automated Cellular Imaging System (Dako, Carpinteria, CA, USA). A marker-specific imaging application (not currently available worldwide) was used for image analysis.
Design:
Tissue samples were provided by the Cooperative Human Tissue Network, which is funded by the National Cancer Institute. Tissue was stained with a specific antibody using an automated workflow model. Specifically, Ki-67 was detected on breast cancer specimens using the Dako Ki-67 RTU Antibody (Clone MIB-1) (Dako, Carpinteria, CA, USA) and visualized with the Dako Flex Detection Kit (low pH TRS) (Dako, Carpinteria, CA, USA). Images of the tissue, captured by digital microscopy, were scored or graded by a board-certified pathologist who identified marker-specific regions of interest. An algorithm was developed that was antibody- and stain-specific and a pathologists report was created.
Results:
An automated workflow model was successfully used to develop an antibody-specific application that allowed standardization of immunohistochemistry staining and imaging resulting in reproducible, quantitative measurements that could then be incorporated into a final report. Important components in the development of this application included staining protocols, image analysis algorithms, and interactions with a board certified pathologist.
Conclusion:
The development of an automated and efficient marker-specific workflow can aid in quantitative measurements from images captured by digital microscopy. An example of this model has been developed for Ki-67. Standardized antibody detection, image capture and image analysis are key components to producing an effective, reproducible tool for routine diagnostic and prognostic pathology markers.
