Table 1 Multivariable survival analysis of clinicopathological and molecular features in patients with stage I-III CRC

From: Multiregional transcriptomics identifies congruent consensus subtypes with prognostic value beyond tumor heterogeneity of colorectal cancer

Variable

Patientsa

Five-year relapse-free survivalb

n (%)

HR [95% CI]

p-value

cox.zph p-value

Total

398 (100)

  

0.2

Sex

Female

204 (51)

Reference

0.2

Male

194 (49)

1.5 [1.0-2.2]

0.05

Age (continuous)

398 (100)

1.0 [1.0-1.1]

0.0007

0.009

Tumor location

Right

176 (44)

Reference

1.0

Left

125 (31)

0.9 [0.6-1.5]

0.7

Rectum

97 (24)

1.0 [0.6-1.7]

0.9

TNM stage

I

95 (24)

Reference

0.4

II

176 (44)

1.2 [0.7-2.2]

0.5

III

127 (32)

2.1 [1.1-3.9]

0.02

Adjuvant chemotherapy

No

327 (82)

Reference

0.4

Yes

64 (16)

1.2 [0.6-2.2]

0.7

Unknown

7 (2)

2.4 [0.7-8.1]

0.1

MSI status

MSS

320 (80)

Reference

0.6

MSI

78 (20)

0.8 [0.3-1.7]

0.5

KRAS status

Wild-type

258 (65)

Reference

0.06

Mutation

140 (35)

1.1 [0.7-1.8]

0.6

BRAFV600E status

Wild-type

328 (82)

Reference

0.4

Mutation

70 (18)

2.6 [1.3-5.3]

0.01

CTL-score (continuous)

398 (100)

1.2 [0.5-3.2]

0.7

0.6

CAF-score (continuous)

398 (100)

1.1 [0.8-1.5]

0.7

1.0

CMS heterogeneity

Homogeneous

257 (65)

Reference

0.5

Heterogeneous

115 (29)

1.6 [1.1-2.4]

0.03

Undetermined

26 (7)

2.2 [1.0-4.7]

0.05

cCMS

cCMS1

70 (18)

Reference

0.5

cCMS2

224 (55)

3.7 [1.6-8.7]

0.002

cCMS3

76 (19)

2.3 [1.0-5.2]

0.04

cCMS4

28 (7)

4.3 [1.6-11.3]

0.003

  1. CAF cancer-associated fibroblasts, CMS consensus molecular subtypes, cCMS congruent consensus molecular subtypes, CTL cytotoxic lymphocytes, MSI microsatellite instability, MSS microsatellite stable, TNM tumor-node-metastasis.
  2. aExcluding patients with synchronous tumors, heterogeneous cCMS classification, pre-surgical chemoradiation and residual tumor status 1 or 2.
  3. bHazard ratios (HR) and 95% confidence intervals (CI) are from a multivariable Cox proportional hazards model, p values are two-sided and from Wald tests, and cox.zph p values are from tests of the proportional hazards assumption. Statistically significant p values are highlighted in bold. Results were similar in a stratified analysis according to the variable breaking the proportional hazards assumption (patient age; Supplementary Data 11).
Back to article page