Skip to main content

Table 3 Performance of models in the discovery set and the three external validation sets

From: Multi-scale pathology image texture signature is a prognostic factor for resectable lung adenocarcinoma: a multi-center, retrospective study

  

Clinical model

Full model

Discovery set

C-index(95%CI)

0.798(0.729–0.867)

0.837(0.784–0.890)

AIC

244.905

235.991

P-value

0.0010

External validation set V1

C-index(95%CI)

0.679(0.573–0.784)

0.704(0.608–0.801)

AIC

222.908

219.568

P-value

 < 0.0001

External validation set V2

C-index(95%CI)

0.666(0.572–0.761)

0.728(0.646–0.811)

AIC

313.815

307.537

P-value

 < 0.0001

External validation set V3

C-index(95%CI)

0.669(0.606–0.731)

0.696(0.632–0.760)

AIC

722.453

717.869

P-value

 < 0.0001

  1. Clinical model: TNM stage + differentiation; Full model: TNM stage + differentiation + MPIS. Note that information related to tumor differentiation was unavailable in the external validation set V3. The P-value is calculated to compare the difference between the clinical model and the full model by the likelihood ratio test
  2. CI confidence interval, AIC Akaike information criterion
Back to article page

Journal of Translational Medicine

ISSN: 1479-5876

Contact us