PerspectiveIn the Journals

KRAS status predicted time to recurrence in resected stage III colon cancer

Show Citation

November 12, 2014

KRAS exon 2 mutations independently predicted shorter time to recurrence in patients with resected stage III distal colon cancer who received adjuvant therapy, according to results of a post-hoc analysis of data from a phase 3 trial.

The prognostic potential of KRAS mutations in patients with colon adenocarcinoma had not been established, so researchers strived to address that question as part of an ancillary study of the PETACC8 trial. The trial evaluated adjuvant FOLFOX chemotherapy with or without cetuximab (Erbitux; Bristol-Myers Squibb, Lilly) in patients with stage III colon cancer.

In the subanalysis, Helene Blons, PhD, of Georges Pompidou European Hospital in Paris, and colleagues assessed the prognostic value of KRAS exon 2 mutations on time to recurrence and DFS.

No differences in time to recurrence and DFS were observed between the two treatment arms in the PETACC8 trial, so both were pooled for analysis in the ancillary study. However, patients with BRAF-mutated cancers were excluded from the ancillary study.

Blons and colleagues identified KRAS mutations in 638 (38.5%) of the 1,657 eligible tumors. They determined KRAS mutations were associated with shorter time to recurrence (P<.001).

Compared with wild-type tumors, those with codon 12 mutations were significantly associated with shorter time to recurrence (HR=1.67; 95% CI, 1.35-2.04) independent of other covariates. However, codon 13 mutations were not significantly associated with shorter time to recurrence (HR=1.23; 95% CI, 0.85-1.79).

Researchers also determined KRAS genotype has different impact on recurrence (P=.02) and DFS (P=.042) based on tumor location.

A subgroup analysis showed KRAS only affected time to recurrence and DFS in distal tumors. Researchers observed a significantly increased risk for relapse among patients with KRAS codon 12 mutations (HR=1.96; 95% CI, 1.51-2.56) and a borderline significantly increased risk for relapse among patients with codon 13 mutations (HR=1.59; 95% CI, 1-2.56).

“Future clinical trials in the adjuvant setting should consider both the tumor location and KRAS mutations as important stratification factors,” Blons and colleagues wrote.

Disclosure: The researchers report research funding and honoraria from, advisory roles with or lecture roles with Amgen, Bayer, Bristol-Myers Squibb, Genomic Health, Lilly, Merck, Merck KGaA, Merck-Serono, Myriad Genetics, Pfizer, Roche and Sanofi.


itj+ Infographic

itj+ Perspective

Andrew H. Ko

Andrew H. Ko

The predictive significance of KRAS is firmly established in terms of guiding selection of colorectal cancer therapy; specifically, multiple lines of clinical evidence have demonstrated definitively that mutations in this oncogene predict a lack of benefit from epidermal growth factor receptor (EGFR) monoclonal antibodies. Furthermore, recent in-depth analyses of available data sets have shown that expanded KRAS testing, beyond mutations located in exon 2 (the ones currently evaluated in standard assays), should also be performed to identify additional subsets of patients in whom EGFR inhibitors should be avoided.
The prognostic value of KRAS, on the other hand, has not been quite as conclusive in colorectal cancer, either in the metastatic setting or for earlier stages of the disease, as various studies examining this question have produced mixed results to date.
The retrospective analysis performed by Blons and colleagues on a large cohort of colon cancer patients enrolled on the PETACC8 trial adds one more checkbox in favor of KRAS mutations representing a negative prognostic factor, at least in terms of risk for relapse for patients with stage III disease (and limited only to mutations in exon 2). Quite intriguingly, these investigators found a specific interaction between KRAS mutations and tumor location in the colon; it was only in distally located tumors, not proximal ones, that mutated KRAS was observed to be a negative prognostic factor for relapse. This offers a useful reminder that, beyond and in addition to the sophisticated molecular subtyping that we now do routinely, we should not lose sight of very simple distinguishing features of colon cancer — such as tumor location — that may also be of pivotal importance and interact with genotype in crucial ways. Variables such as these could at some point be useful in influencing treatment decisions or serve as the basis of stratification in future colon cancer trials, as the authors suggest. However, further corroboration is required. Truly, we are discovering quite clearly that colon cancers are not all created equal, in more ways than we might even have imagined.

Andrew H. Ko, MD
HemOnc Today Editorial Board member
University of California, San Francisco

Disclosure: Ko reports no relevant financial disclosures.