Table 4 Inter-reader Agreement (Kappa Coefficients) for Diagnostic Performance

From: Development and validation of an interpretable model integrating multimodal information for improving ovarian cancer diagnosis

  

Readers

Internal test dataset

External test dataset

Inter-reader

O-RADS

A vs. B

0.760 (0.640, 0.873)

0.649 (0.576, 0.720)

A vs. C

0.711 (0.601, 0.817)

0.588 (0.521, 0.653)

A vs. D

0.837 (0.739, 0.929)

0.693 (0.627, 0.763)

A vs. E

0.924 (0.844, 0.982)

0.667 (0.602, 0.736)

B vs. C

0.800 (0.679, 0.901)

0.669 (0.597, 0.742)

B vs. D

0.826 (0.711, 0.914)

0.742 (0.673, 0.812)

B vs. E

0.802 (0.694, 0.909)

0.760 (0.685, 0.828)

C vs. D

0.835 (0.741, 0.932)

0.681 (0.603, 0.760)

C vs. E

0.782 (0.679, 0.878)

0.760 (0.687, 0.826)

D vs. E

0.911 (0.829, 0.982)

0.796 (0.732, 0.856)

OvcaFinder

A vs. B

0.886 (0.798, 0.951)

0.869 (0.812, 0.921)

A vs. C

0.902 (0.834, 0.968)

0.868 (0.815, 0.925)

A vs. D

0.983 (0.949, 1.000)

0.910 (0.862, 0.954)

A vs. E

0.983 (0.949, 1.000)

0.896 (0.841, 0.940)

B vs. C

0.952 (0.889, 1.000)

0.863 (0.804, 0.917)

B vs. D

0.902 (0.840, 0.967)

0.894 (0.839, 0.941)

B vs. E

0.902 (0.825, 0.967)

0.906 (0.854, 0.952)

C vs. D

0.918 (0.855, 0.983)

0.892 (0.840, 0.940)

C vs. E

0.918 (0.855, 0.983)

0.891 (0.834, 0.939)

D vs. E

0.967 (0.916, 1.000)

0.933 (0.893, 0.973)

  1. Data in parentheses are 95% confidence intervals; O-RADS Ovarian-Adnexal Reporting and Data System.