Fig. 6From: Performance comparison between multi-center histopathology datasets of a weakly-supervised deep learning model for pancreatic ductal adenocarcinoma detectionClassification performance in a 12-Fold CV overt the CPTAC dataset where each test set was formed by patients of a single nationality. Panel A Test accuracy obtained per country in the 12-Fold CV. Panel B F1-score obtained per country in the 12-Fold CV. Panel C Number of samples per nationality in the CPTAC datasetBack to article page