Skip to main content

Table 3 Network performance on outcome measures. Kellgren & Lawrence grades are displayed separately and merged

From: Automating classification of osteoarthritis according to Kellgren-Lawrence in the knee using deep learning in an unfiltered adult population

 

Cases (n = 300)

Sensitivity (%)

Specificity (%)

Youden’s J

AUC (95% CI)

Kellgren-Lawrence

 0

67

97

88

0.85

0.97 (0.93 to 0.99)

 1

23

96

75

0.70

0.88 (0.83 to 0.92)

 2

24

92

61

0.53

0.80 (0.73 to 0.86)

 3

116

92

71

0.63

0.87 (0.83 to 0.90)

 4

70

84

78

0.63

0.87 (0.83 to 0.91)

Grouped Kellgren-Lawrence

 0 to 1

90

88

95

0.83

0.96 (0.94 to 0.98)

 0 to 2

114

83

97

0.81

0.97 (0.95 to 0.98)

 1 to 3

163

80

74

0.53

0.82 (0.77 to 0.87)

 2 to 4

210

97

83

0.80

0.96 (0.94 to 0.98)

 3 to 4

186

96

88

0.84

0.97 (0.95 to 0.99)