Skip to main content

Table 1 Molecular subtyping accuracy across folds. Sample size and performance statistics within the held-out test set across fivefold cross-validation

From: A deep learning image-based intrinsic molecular subtype classifier of breast tumors reveals tumor heterogeneity that may affect survival

 

No. of patches

No. of WSIs

No. of patients

Patch-level accuracy (%)

WSI-level accuracy (%)

Patient-level accuracy (%)

Fold1

7505

95

92

60.47

70.53

71.74

Fold2

7501

94

89

56.97

67.02

67.42

Fold3

7581

95

88

57

67.37

69.32

Fold4

7564

95

86

59.56

65.26

65.12

Fold5

7420

93

88

59.1

60.22

62.5

Average

7514.2

94.4

88.6

58.62

66.1

67.27