Skip to main content
. 2019 Sep 21;179(1):197–206. doi: 10.1007/s10549-019-05446-y

Table 3.

Associations between PAM50 subtypes and hypoxia signatures with breast cancer outcomes: Multiple regression survival analysis*

Disease-free survival
(N = 295 relapse events)*
Breast cancer survival
(N = 212 breast cancer deaths)*
PAM50 subtypea HR (95% CI) HR (95% CI)
 Luminal A (ref) 1.0 1.0
 Basal 1.24 (0.87, 1.78) 1.01 (0.65, 1.55)
 Her2 0.98 (0.65, 1.49) 0.91 (0.56, 1.49)
 Luminal B 1.60 (1.19, 2.13) 1.68 (1.20, 2.35)
Model comparison: clinical vs (PAM50 + clinical)
Likelihood ratio test: Chi square statistic 11.7 12.9
 Degrees of freedom 3 3
 P-value 0.009 0.005
VEGF13 signatureb
 Low (ref) 1.00 1.00
 Medium 1.33 (0.99, 1.78) 1.27 (0.90, 1.79)
High 1.48 (1.08, 2.02) 1.41 (0.98, 2.03)
Model comparison: (PAM50 + clinical) vs (PAM50 + clinical +VEGF13)
 Likelihood ratio test: chi square statistic 6.2 3.6
 Degrees of freedom 2 2
 P-value 0.04 0.16
VEGF15 signatureb
 Low (ref) 1.00 1.00
 Medium 0.92 (0.68, 1.24) 0.92 (0.65, 1.31)
 High 1.33 (0.99, 1.78) 1.33(0.94,1.87)
Model comparison: (PAM50 + clinical) vs (PAM50 + clinical +VEGF15)
Likelihood ratio test: chi square statistic 6.7 4.8
 Degrees of freedom 2 2
 P-value 0.03 0.09

*Subjects who were classified as normal-like subtype were excluded from this analysis

aModel adjusted for age at diagnosis, tumor grade, tumor stage

bModel adjusted for age at diagnosis, tumor grade, tumor stage, PAM50 subtype