Skip to main content

Table 4 Predictive performance of the model

From: A gadoxetic acid-enhanced MRI-based model using LI-RADS v2018 features for preoperatively predicting Ki-67 expression in hepatocellular carcinoma

Appearance

AUC

95%CI

Sensitivity

95%CI

Specificity

95%CI

Mosaic architecture

0.727

0.639, 0.804

88.89

75.9, 96.3

56.58

44.7, 67.9

Infiltrated appearance

0.619

0.526, 0.706

77.78

62.9, 88.8

46.05

34.5, 57.9

Targetoid HBP

0.575

0.482, 0.665

75.56

60.5, 87.1

39.47

28.4, 51.4

Model A

0.781

0.696, 0.851

75.56

60.5, 87.1

77.63

66.6, 86.4

Model B

0.746

0.659, 0.821

88.89

75.9, 96.3

56.58

44.7, 67.9

Model C

0.669

0.578, 0.752

60.00

44.3, 74.3

71.05

59.5, 80.9

Model D

0.776

0.691, 0.847

84.44

70.5, 93.5

63.16

51.3, 73.9

  1. Model A: mosaic architecture and infiltrated appearance
  2. Model B: mosaic architecture and targetoid HBP
  3. Model C: infiltrated appearance and targetoid HBP
  4. Model D: mosaic architecture, infiltrated appearance and targetoid HBP
  5. Abbreviations: AUC the areas under the receiver operating characteristic curves (AUC), CI confidence interval, HBP hepatobiliary phase