Home| Background | Contest Registration | Download |Submission | Contest Rules | Forum | Contact 

Introduction

Human epidermal growth factor receptor 2 (HER2) is a protein that mainly influence the growth of malignant epithelial cells. HER2 overexpression nearly exist in 15-20% cases of invasive breast cancer which tends to attain significance as a prognostic and predictive model. In routine practice, pathologist visually inspects the morphological features to report a score ranging between 0–3 (Table 1).

Staining Score	Staining Pattern	Staining Assessment
0	No membrane staining observed or lies in less then 10 % of the tumor cells	Negative
1	Faint, weak membrane staining in some proportion of the tumor cells	Negative
2	Feeble, non-uniform membrane staining in greater than 10% of the tumor cells (recommended FISH testing )	Equivocal (Weakly Positive)
3	Vigorous, intense membrane staining is observed in more then 10% of the tumor cells	Positive

Table 1: Criteria for scoring HER2 based on cell membrane staining

HER2 is a transmembrane growth factor receptor which causes the division and growth of tumor cells. In routine practice, cytoplasmic staining and normal epithelial cells are not observed and only membrane staining of the invasive tumor regions are being considered while scoring HER2. The tissue regions are normally scored on the basis of morphological patterns and enrichment of cell membrane staining. Generally, there are two methods for testing the HER2, immunohistochemistry (IHC) and fluorescence in-situ hybridisation (FISH). IHC usually done first and those samples scoring 0/1+ are considered as negative, and those scoring 2+ are rated as equivocal with further examined using FISH (status of gene amplification), and samples with 3+ score are considered as strongly positive. The morphological appearance of membrane staining with their corresponding scores can be seen in Fig 1. Further details of the scoring protocol can be found in the following paper:

• E.A. Rakha, S.E. Pinder, J.M.S. Bartlett, M. Ibrahim, J. Starczynski, P.J. Carder, E. Provenzano, A. Hanby, S. Hales, A.H.S. Lee, I.O. Ellis, "Updated UK Recommendations for HER2 assessment in breast cancer," Journal of Clinical Pathology, 68:93-99 (2015) [DOI]

    

Fig 1: a) Score 0 (negative) b) Score 1 (negative) c) Score 2 (equivocal) d) Score 3 (Positive)

The Solution: Automated scoring of HER2

Digital pathology and the adoption of image analysis have grown rapidly in the last few years. The automated whole slide image (WSI) scoring will certainly be a vital step towards computer assisted diagnostic system and it will significantly be helpful for pathologist to overcome their workload. Automated scoring of HER2 also acquire substantial benefits to improve the clinical utility and analytical validity. The essence of the contest is to provide a platform for computer scientists to contribute and evaluate the performance of scientific algorithms for automated scoring of HER2 on Immunohistochemical (IHC) stained slides.