Skip to main content

Table 4 Diagnostic performance of models in the training and testing cohorts

From: The solid component within part-solid nodules: 3-dimensional quantification, correlation with the malignant grade of nonmucinous pulmonary adenocarcinomas, and comparisons with 2-dimentional measures and semantic features in low-dose computed tomography

Models

AUC (95% CI)

Sensitivity (95% CI)

Specificity (95% CI)

Z

P a

Training cohort

     

Semantic

0.809 (0.753–0.856)

0.784 (0.713–0.845)

0.753 (0.642–0.844)

2.755

0.006

2-dimensional

0.821 (0.767–0.868)

0.728 (0.653–0.795)

0.818 (0.714–0.897)

3.097

0.002

CTRV− 250HU

0.890 (0.843–0.927)

0.870 (0.809–0.918)

0.792 (0.685–0.876)

-

-

Testing cohort

     

Semantic

0.710 (0.603–0.802)

0.763 (0.634–0.864)

0.714 (0.513–0.868)

1.990

0.047

2-dimensional

0.760 (0.656–0.845)

0.661 (0.526–0.779)

0.750 (0.551–0.893)

2.108

0.035

CTRV− 250HU

0.832 (0.737–0.904)

0.847 (0.730–0.928)

0.786 (0.590–0.917)

-

-

  1. a Comparison of AUC with 3-dimensional model. AUC, area under curve; CI, confidence interval