Skip to main content
. 2015 Aug 25;113(6):979–988. doi: 10.1038/bjc.2015.296

Table 4. Likelihood ratio tests evaluating the fit of OS and RFS data for patients with microsatellite stable, proximal colon cancer when adding APC mutation status to Cox proportional-hazard models with CIMP-H, BRAF, KRAS, PIK3CA, TP53 or CIN status (labeled X) or vice versa.

  Overall survival
Recurrence-free survival
  Events/N P (add APC) P (add X) AIC (APC)-AIC (X) Events/N P (add APC) P (add X) AIC (APC)-AIC (X)
CIMP status 58/149 0.003* 0.542 −9.3 43/123 0.021* 0.731 −6.7
BRAF-mt 79/220 0.036* 0.323 −3.4 46/130 0.014* 0.464 −5.6
KRAS-mt 79/220 0.012* 0.827 −6.3 46/130 0.002* 0.470 −9.4
PIK3CA-mt 79/220 0.010* 0.181 −4.8 46/130 0.002* 0.487 −8.8
TP53-mt 79/220 0.007* 0.403 −6.7 46/130 0.001* 0.178 −8.4
CIN status 71/194 <0.001* 0.904 −11.2 44/123 0.001* 0.096 −8.1

Abbreviations: CIMP=CpG island methylator phenotype; CIN=chromosomal instability; MSS=microsatellite stable; OS=overall survival; RFS=recurrence-free survival.

*P<0.05.

Differences in Aikake Information Criteria (AIC) are shown for the comparison of models with APC mutation against the relevant molecular feature of interest; negative values of the AIC difference indicate a better fit for the APC mutation models.