Skip to main content

Table 4 The consistency between the training set and the validation set pathologists interpreting breast cancer Ki-67LI in the homogeneous and heterogeneous groups through SRC and AI

From: Artificial intelligence-assisted interpretation of Ki-67 expression and repeatability in breast cancer

Variables

ICC(95%CI)

P

Training group

Validation group

Homogeneity

Heterogeneity

Heterogeneity

Heterogeneity

VA

0.848 (0.785,0.901)

0.698 (0.634,0.760)

  

<0.001

MC

0.941 (0.914,0.963)

0.672 (0.605,0.738)

  

<0.001

SRC

0.964 (0.946,0.977)

0.877 (0.844,0.907)

0.985 (0.978,0.990)

0.990 (0.986,0.993)

<0.001

AI

0.984 (0.976,0.990)

0.959 (0.947,0.970)

0.986 (0.979,0.990)

0.991 (0.986,0.994)

<0.001

  1. P < 0.05 is statistically significant
  2. VA microscopic visual assessment, MC microscopic manual counting, SRC standard comparison card, AI artificial intelligence, LI Label Index
Back to article page