Skip to main content
. 2021 Mar 29;11:7080. doi: 10.1038/s41598-021-86540-4

Figure 1.

Figure 1

Each confusion matrix compares the classification agreement of our model with pathologists’ consensus for each of our three test sets: (a) surgical resection whole-slide images from DHMC, (b) surgical resection whole-slide images from TCGA, and (c) biopsy whole-slide images from DHMC.