Qualitopix - automatic quality assessment

S. Harder, A. Ottosen, A. Schønau, M. Vyberg
Visiopharm A/S, Hoersholm, Denmark

Objective: Staining sufficiency in immunohistochemistry (IHC) is essential for patient diagnostics. Therefore, proficiency testing should be performed by External Quality Assessment (EQA) organizations using expert pathologists. We propose to use digital image analysis for automatic quality assessment, to assist pathologists and to assure an objective assessment.
Method: Image analysis can be used to extract useful information from a virtual IHC slide. The useful information with respect to quality assessment is e.g., completeness of membrane staining (assessing human epidermal growth factor receptor 2 [HER 2]) [1] or percentage and intensity of tumour nuclei staining (assessing estrogen receptor [ER]). Our image analysis method captures completeness of membrane staining using connectivity [2] and percentage and intensity of tumour nuclei staining using H-score [3].
Results: A Receiver Operating Characteristic (ROC) analysis of the quality assessments performed by pathologists from NordiQC and our automatic method reveal an area under the curve of 0.98 for HER2 assessments and 0.93 for ER assessments. The assessments were classified into sufficient and insufficient.
Conclusion: Our image analysis method shows promising results for automatic quality assessment of both HER2 and ER, rendering the method useful as a support for EQA organizations or as an objective quality assessment for individual laboratories. [1] RA Walker et al., “HER2 testing in the UK: further update to recommendations,” Journal of Clinical Pathology, 2008. [2] A Brügmann et al., “Digital image analysis of membrane connectivity is a robust,” Breast Cancer Res Treat, 2011. [3] NL Andersen et al., “A Digital Approach to Immunohistochemical Quantification of Estrogen Receptor Protein in Breast Carcinoma Specimens,” Appl Immunohistochem Mol Morphol, 2017.