Table 4 Performance of the constructed model combined with existing pathological variables according to multivariable analysis

From: Whole slide image based deep learning refines prognosis and therapeutic response evaluation in lung adenocarcinoma

 

C index

p

3-y AUC

p

5-y AUC

p

Validation cohort 1

 IASLC grade & TNM stage

0.737 (0.706–0.768)

<0.001

0.766 (0.726–0.806)

<0.001

0.802 (0.766–0.839)

<0.001

 WSI-based score & IASLC grade

0.708 (0.677–0.739)

<0.001

0.733 (0.692–0.774)

<0.001

0.746 (0.707–0.786)

<.001

 WSI-based score & TNM stage

0.706 (0.671–0.741)

<0.001

0.731 (0.684–0.777)

<0.001

0.740 (0.696–0.784)

<0.001

 WSI-based score & IASLC grade & TNM stage

0.753 (0.720–0.786)

–

0.782 (0.742–0.823)

–

0.810 (0.772–0.847)

–

Validation cohort 2

 IASLC grade & TNM stage

0.777 (0.716–0.838)

<0.001

0.780 (0.695–0.866)

<0.001

0.877 (0.819–0.936)

<0.001

 WSI-based score & IASLC grade

0.786 (0.729–0.843)

<0.001

0.795 (0.719–0.871)

<0.001

0.829 (0.762–0.895)

<0.001

 WSI-based score & TNM stage

0.763 (0.706–0.820)

<0.001

0.804 (0.727–0.881)

<0.001

0.797 (0.718–0.876)

<0.001

 WSI-based score & IASLC grade & TNM stage

0.811 (0.756–0.866)

–

0.825 (0.747–0.904)

–

0.881 (0.824–0.939)

–

  1. The p values refer to the comparison between the combination of two independent predictors and the combination of three independent predictors in the multivariable Cox regression model.
  2. AUC area under the curve, IASLC International Association for the Study of Lung Cancer.
Back to article page