Discriminating between ductal carcinoma in situ and invasive tumour components using digital image analysis

D. Omanovic, T. Steiniche, G. Tindbæk, A. Schønau
Visiopharm A/S, Hoersholm, Denmark

Objective: To investigate, with a digital image analysis (DIA) system, the importance of discriminating between ductal carcinoma in-situ (DCIS) and invasive tumour components (ITC) when using tissue-based biomarkers for diagnosing breast cancer
Method: Digital images of serial whole tissue sections, one double-stained with CK7/19 and p63 and the other stained with the analytical marker Ki-67 or HER2, were aligned using the VirtualDoubleStaining™ technique. Using novel Visiopharm image analysis algorithms ITC regions were identified based on the CK7/19 stain and DCIS regions were identified based on the joint CK7/19 and p63 stain. These regions were transferred to the analytical slide, where the biomarker expression was quantitated within the ITC regions only or within the ITC regions and the DCIS regions. The two set of results were compared to manual scores.
Results: The presented DIA system has been found to robustly discriminate between DCIS and ITC facilitating the quantification of biomarker expression within ITC only. Preliminary results show that this could improve the diagnostic capabilities when using tissue-based biomarkers for diagnosing breast cancer
Conclusion: Using DIA to reliably detect and eliminate DCIS components from the consecutive analysis of biomarker expression could prove to be an important tool when diagnosing breast cancer using tissue-based biomarkers.