Figure 5: HER2 scoring concordance.
From: Relevance of deep learning to facilitate the diagnosis of HER2 status in breast cancer
Top: comparison between the initial pathologist assessment (IP) and deep learning (ConvNets); bottom-left: comparison between the initially concordant scores and the independent review; bottom-right: comparison between the initially discordant scores and the independent review. Vertical bars represent single cases and pie charts represent overall agreement between either ConvNets or the independent review and the initial pathologist assessment.
