Table 4 Multivariable regression of case-level DLS score using clinicopathologic features as input.

From: Interpretable survival prediction for colorectal cancer using deep learning

Clinicopathologic feature

Validation Set 1

Validation Set 2

 

Coefficient

p

R2

Coefficient

p

R2

T3

0.5454

<0.001

0.18

0.1184

0.276

0.18

T4

0.7775

<0.001

0.4032

<0.001

N1

0.5496

<0.001

0.2912

<0.001

N2

0.5942

<0.001

0.4752

<0.001

N3

1.0311

<0.001

0.3477

0.163

R1

0.1108

0.427

0.3365

0.011

L1

−0.1569

0.032

0.1063

0.074

V1

0.2376

0.033

0.1332

0.054

Grade 2

0.1032

0.467

0.0557

0.605

Grade 3

0.4342

0.004

0.1800

0.112

Grade X

0.5504

0.049

0.1968

0.287

Sex (female)

−0.0091

0.862

0.0179

0.713

Age at diagnosis

−0.0670

0.002

−0.0043

0.833

Intercept

−1.0471

<0.001

−1.4258

<0.001

  1. For the overall model, p < 0.001 (t test). Each coefficient represents the relative increase of the DLS score associated with that variable. Bold indicates statistically significant input variables (p < 0.05).
Back to article page