Thursday, 21 February 2019

Quantitative immunohistochemical analysis


Jump to: navigation, search



Olivier DEBEIR - Christine DECAESTECKER - Adrien FOUCART - Yves-Rémi VAN EYCKE

Project description

Example of immunohistochemical (IHC) stained slide

The quantification of immunostaining by means of digital image analysis allows the characterization of protein expression revealed by immunohistochemistry (IHC) on normal or pathological tissue slices. In contrast to other methods of evaluation of protein expression, this approach enables morphological controls and detailed tissue and cell localization to be carried out, avoiding the problems due to cell and tissue heterogeneity. This constitutes a promising approach for the identification of tissue-based biomarkers useful for diagnostic, prognostic and/or therapeutic purposes.

This project is is supported by the F.N.R.S. (Belgian National Fund for Scientific Research) and carried out in close collaboration with the Pathology Department of the Erasme Hospital (Prof. I. Salmon). It is thus centered on problems commonly encountered by pathologists in their research activities and aims to provide original and pragmatic solutions based on an engineering approach.

IHC biomarker processing workflow

It consists in:

  • Standardization of the image acquisition process
  • Automated methods for whole slide imaging (WSI) quality assessment (including image normalization process)
  • Robust morphological and IHC feature extraction
  • Multiresolution registration of serial slide images
  • IHC staining colocalization measurement
  • Data base construction (including clinical features) and integrated data analysis
  • Identification and validation of biomarkers by means of multivariate data analysis and machine learning

In addition: deep learning and data augmentation techniques to automate image annotation of histological slides and so allow compartmentalization of IHC quantification


Diapath platform

Resulting from our close collaboration with the Pathology Department of the Erasme Hospital, DIAPATH (Digital Image Analysis in Pathology) is a unit of the Center for Microscopy and Molecular Imaging (CMMI). This center is a biomedical imaging facility supported by the "Région Wallonne" and the European Union ("FEDER-Convergence" funds). The DIAPATH unit aspires to become a leader in histological image analysis services for academic and industrial partners.



We developed a tool able to help pathologists in the task of characterizing the proliferative activity on high-grade gliomas. Ki67 is a common biomarker used in clinical practice to assess the proliferative activity of tumors. Faced with large images, pathologists require tools that can help them identify tumor regions that exhibit high proliferating activity, called "hot-spots" (HSs). The new tool is capable of helping pathologists to identify Ki67 HSs in whole tumor section images that are produced by slide scanners. We showed that this tool strongly improves the consistency among pathologists in this task of HS identification.

In order to compare the performance and the robustness of different clustering algorithms, we produce a set of 50 different datasets. These datasets are publicly available on the page CLUSTERING METHODS APPLIED IN THE DETECTION OF KI67 HOT-SPOTS

Representative publications

via Di-fusion

Books chapters

Moles Lopez, Xavier, Debeir, Olivier, Salmon, Isabelle, Decaestecker, Christine,
Whole slide imaging and analysis for biomarker evaluation in digital pathology
A. Méndez-Vilas (Ed.), Microscopy: advances in scientific research and education, Vol. 2, pp. 776-787, 2014
Bibtex : info:hdl:2013/175598
Note : Language of publication: na