Skip to main content

Table 3 Calibration assessment of the EGFR mutation predictive model

From: Development and validation of a predictive model for estimating EGFR mutation probabilities in patients with non-squamous non-small cell lung cancer in New Zealand

Group a

N

Predicted EGFR mutation

Observed EGFR mutation

Number

Mean

(min-max)

Number

Proportion

(95% CI)

Development group

1

254

17

0.07

(0.04–0.08)

18

0.07

(0.04–0.11)

2

228

24

0.11

(0.09–0.11)

14

0.06

(0.04–0.10)

3

282

41

0.14

(0.11–0.15)

41

0.15

(0.11–0.19)

4

234

67

0.28

(0.16–0.38)

80

0.34

(0.28–0.40)

5

178

101

0.57

(0.39–0.62)

96

0.54

(0.47–0.61)

Validation group

1

100

7

0.07

(0.04–0.08)

7

0.07

(0.03–0.14)

2

100

10

0.10

(0.09–0.11)

9

0.09

(0.05–0.16)

3

133

20

0.15

(0.11–0.15)

15

0.11

(0.07–0.18)

4

59

14

0.23

(0.16–0.29)

13

0.22

(0.13–0.34)

5

97

48

0.50

(0.31–0.62)

46

0.47

(0.38–0.57)

  1. a The five groups were created by the ranks of the predicted probabilities