Figure 3 | Scientific Reports

Figure 3

From: Comparative analysis of machine learning approaches to classify tumor mutation burden in lung adenocarcinoma using histopathology images

Figure 3

Visualization of LungCNN-Histo predictions for annotated regions. Shown are representative regions of interest (1 mm × 1 mm) for regions with predominance of each of the feature classes (see legend). Black regions represent areas of tissue where there was lack of consensus for the three pathologist annotations. Middle columns for each panel represent consensus pathologist annotations and the right panel depicts the model classification across patches (with the non-consensus patches masked in black to match for easier visual reference to the pathologist annotations.) H&E hematoxylin and eosin, ROIs regions of interest.

Back to article page