Skip to main content

Table 1 Summary the model training statistics

From: A broad scope knowledge based model for optimization of VMAT in esophageal cancer: validation and assessment of plan quality among different treatment centers

 

Lung left

Lung right

Heart

Spine

Liver

Structures in model

70

70

70

70

65

Estimation model goodness of fit

 DVH average fit

0.99

0.99

0.99

0.99

0.99

 GED average fit

0.99

0.99

0.99

0.99

0.99

 Coeff. of determination

0.83

0.92

0.73

0.49

0.88

 Whole estimation model fit

0.83

0.91

0.73

0.49

0.87

 Average chi square

1.07

1.09

1.04

1.05

1.05

Model goodness of estimation

 MSE original and estimate

0.001

0.001

0.005

0.01

0.002

 Dose bins outside bound.[%]

41

42

35

28

41

 

Spleen

Stomach

Left kidney

Right kidney

Small bowel

 Structures in model

60

64

34

31

40

Estimation model goodness of fit

 DVH average fit

0.99

0.99

0.99

0.99

0.99

 GED average fit

0.99

0.99

0.97

0.99

0.99

 Coeff. of determination

0.88

0.95

0.79

0.84

0.95

 Whole estimation model fit

0.87

0.95

0.76

0.82

0.94

 Average chi square

1.02

1.09

1.08

1.11

1.15

Model goodness of estimation

 MSE original and estimate

25

0.002

0.01

0.004

0.003

 Dose bins outside bound.[%]

0.005

39

29

23

44