Skip to main content

Table 2 Tabular results of the quadratic weighted Cohen κ scores for the different readers and the deep learning (DL) model

From: Detection of femoropopliteal arterial steno-occlusion at MR angiography: initial experience with artificial intelligence

Side

Classes

Reader 1 versus Reader 2

Reader 1 versus

DL model

Reader 2 versus

DL model

Left

2

0.90

0.72

0.74

 

3

0.91

0.78

0.73

 

4

0.92

0.80

0.73

Right

2

0.87

0.70

0.69

 

3

0.87

0.69

0.69

 

4

0.89

0.66

0.68

  1. Different class scenarios were generated for various class separations, including binary class (0 versus 1−3), three-class (0 versus 1 or 2 versus 3), and four-class (0−3) cases