Table 1 Classification performance of the TNTCM across different datasets

From: A full-automated tumor budding annotation approach in hematoxylin and eosin-stained whole slide images of colorectal cancer

DATASETS

SPECIFICITY (95% CI)

SENSITIVITY (95% CI)

AUC (95% CI)

PUTH-IV

0.951 (0.934, 0.968)

0.935 (0.915, 0.955)

0.988 (0.979, 0.997)

TCGA-COAD-EV

0.889 (0.809, 0.968)

0.848 (0.758, 0.939)

0.921 (0.853, 0.990)

TCGA-READ-EV

0.733 (0.620, 0.847)

0.929 (0.862, 0.995)

0.929 (0.862, 0.995)

  1. Bold values highlight the optimal results (i.e., the best performance) for each metric across the compared methods.
Back to article page