Table 2 Comparison of ARGO performance in the whole vs. the top 50 reportsa.

From: Electronic case report forms generation from pathology reports by ARGO, automatic record generator for onco-hematology

 

PRECISION (P) = TP/(TP + FP)

RECALL (R) = TP/(TP + FN)

F1-SCORE = 2*(P*R)/(P + R)

All reports, N = 239

Top reports, N = 50a

Diff

All reports, N = 239

Top reports, N = 50a

Diff

All reports, N =239

Top reports N = 50a

Diff

DATA-FIELD

TP, N

FP, N

TP + FP

P, %

TP, N

FP, N

TP + FP, N

P, %

%

TP, N

FN, N

TP + FN, N

R, %

TP, N

FN, N

TP + FN, N

R, %

%

F1, %

F1, %

%

MYC

20

0

20

100.0

13

0

13

100.0

0.0

20

15

35

57.1

13

4

17

69.2

12.1

72.7

81.8

9.1

BCL2

130

2

132

98.5

28

0

28

100.0

1.5

130

55

185

71.4

28

5

33

84.8

13.4

82.2

91.8

9.6

BCL6

115

1

116

99.1

27

0

27

100.0

0.9

115

51

166

61.5

27

5

32

84.4

22.9

75.9

91.5

15.6

CD10

95

3

98

96.9

25

0

25

100.0

3.1

95

75

170

55.9

25

7

32

78.1

22.2

70.9

87.7

16.8

CD20

164

1

165

99.4

36

0

36

100.0

0.6

164

51

215

76.3

36

3

39

92.3

16.0

86.3

96.9

10.6

Cyclin D1

58

0

58

100.0

5

0

5

100.0

0.0

58

23

81

71.6

5

3

8

62.5

-9.1

83.5

76.3

-7.2

Mean (std)

1.0 (1.2)

Mean (std)

12.9 (11.7)

Mean (std)

9.1 (8.6)

  1. ARGO: Automatic Record Generator for Onco-hematology, TP: True Positive, FP: False Positive, FN: False Negative, CD: Cluster of Differentiation, Diff: difference, std: standard deviation.
  2. aTop 50 reports (internal series) with the highest optical resolution.