Table 1 Clinicopathologic characteristics in training set, validation set and test set.

From: Deep learning to automatically evaluate HER2 gene amplification status from fluorescence in situ hybridization images

 

Training set

Validation set

Test set

p value (training set vs. validation set)

p value (training set vs. test set)

p value (validation set vs. test set)

Age (years)

 ≤ 50 

93

31

36

0.657

0.530

0.900

 > 50 

87

33

40

   

Tumor size (cm)

 < 2 

27

8

10

0.284

0.446

0.764

 ≥ 2 

38

19

20

   

Tumor grade

 1/2

63

20

27

0.304

0.898

0.434

 3

74

33

33

   

Node status

 Negative

68

32

25

0.302

0.708

0.603

 Positive

41

13

13

   

Molecular subtypes

 Luminal A

39

19

15

0.052

0.225

0.403

 Luminal B

90

21

29

   

 HER2-overexpression

21

6

12

   

 Triple-negative breast cancer

30

18

19

   
Back to article page