KRAS Mutation in Stage III Colon Cancer and Clinical Outcome Following Intergroup Trial CALGB 89803 Free

KRAS, one of the first genes found to be mutated in human cancer, encodes a G-protein downstream of receptor tyrosine kinases, including epidermal growth factor receptor (EGFR; ref. 1–3). Population-based studies have shown that approximately 30% to 40% of colon cancers harbor mutations in codons 12 and 13 of KRAS (4–6). Retrospective observational studies (7–12) as well as randomized controlled trials (13–17) have consistently shown that KRAS mutation in stage IV colorectal cancer confers resistance to anti-EGFR targeted treatment (cetuximab or panitumumab). However, whether KRAS mutation in colorectal cancer has a prognostic role, independent of anti-EGFR therapy, has been controversial (18–21). Previous data have not been conclusive, even among several large studies (4, 6, 22–26). In addition, whether KRAS mutational status modifies the effect of irinotecan-based chemotherapy remains uncertain.

We therefore examined the influence of KRAS on cancer recurrence and survival in a large number (N = 508) of stage III colon cancer patients enrolled in a National Cancer Institute (NCI)-sponsored clinical trial of postoperative adjuvant chemotherapy (27). Within this trial (CALGB 89803), patients were randomized to either fluorouracil and leucovorin or fluorouracil, leucovorin, and irinotecan. Moreover, because data on pathologic stage, performance status, postoperative treatment, and follow-up were carefully captured in this trial, the simultaneous impact of disease characteristics and the use of adjuvant therapy could be assessed to control for potential confounding.

Patients in this study were participants in the NCI-sponsored Cancer and Leukemia Group B (CALGB) adjuvant therapy trial for stage III colon cancer comparing therapy with the weekly Roswell Park regimen of 5-fluorouracil (FU) and leucovorin (FU/LV) with the weekly bolus regimen of irinotecan, FU, and leucovorin (IFL; CALGB 89803; ref. 27). From April 1999 to May 2001, 1,264 patients were enrolled in the treatment trial. Patients in the treatment trial (and thus this companion study) were eligible if they had undergone a complete surgical resection of the primary tumor within 56 d prior to study entry, and had regional lymph node metastases (stage III colon cancer) but no evidence of distant metastases. Moreover, patients were required to have a baseline Eastern Cooperative Oncology Group performance status of 0 to 2 (ambulatory; ref. 28) and have adequate bone marrow, renal, and hepatic function. The current analysis was limited to 508 patients for whom archived formalin-fixed paraffin-embedded tumor tissue was available and the KRAS gene was sequenced. All patients signed informed consent, approved by each site's institutional review board.

We compared the baseline characteristics of the patients who were included in this study (with available KRAS data, n = 508) with those who were excluded from this study due to unavailability of tissue data (n = 756). We did not detect any significant or substantial difference between these two groups in terms of age, sex, body mass index, tumor location, T stage, N stage, performance status, bowel perforation, bowel obstruction, or treatment arm. In addition, tumor recurrence or mortality did not substantially differ between these two groups; multivariate hazard ratios (HR; KRAS data available versus unavailable) were 1.05 [95% confidence interval (95% CI), 0.87-1.27] for disease-free survival (DFS), 1.05 (95% CI, 0.86-1.28) for recurrence-free survival (RFS), and 1.06 (95% CI, 0.86-1.32) for overall survival (OS).

In this study, the primary end point was DFS, defined as time from the study enrollment to tumor recurrence, occurrence of a new primary colon tumor, or death from any cause. In addition, we defined RFS as the time from the study enrollment to tumor recurrence or occurrence of a new primary colon tumor. For RFS, patients who died without known tumor recurrence were censored at last documented evaluation by treatment provider. Finally, OS was defined as the time from the study enrollment to death from any cause.

DNA was extracted from paraffin-embedded tissue of colon cancer as previously described (29). We marked a tumor area on a H&E-stained slide, and dissected the tumor area from another tumor tissue section by a sterile needle for subsequent DNA extraction. PCR and pyrosequencing spanning KRAS codons 12 and 13 were done as previously described (29), and validated against Sanger sequencing method (29, 40). In our KRAS pyrosequencing assay, we routinely confirmed the presence of a mutation by two different sequencing primers and by the creation of frameshifted reading of a mutant sequence relative to a wild-type sequence in a pyrogram (29). Microsatellite instability (MSI) was assessed using 10 DNA mononucleotide and dinucleotide microsatellite markers as previously described (30). Tumors showing instability in ≥40% of the loci tested were classified as MSI-high. Tumors showing instability in no or <40% of the loci were classified as microsatellite stable (MSS)/MSI-low.

The goal of this correlative study was to determine whether tumoral KRAS mutational status influences clinical outcome of patients with stage III colon cancer. Patient registration and clinical data collection were managed by the CALGB Statistical Center, and analyses were conducted collaboratively between the CALGB Statistical Center and Dana-Farber Cancer Institute. All analyses were based on the study database frozen on March 7, 2008, except for the tumoral KRAS data. All analyses used SAS version 9.1 (SAS Institute) and all P values were two-sided.

In the treatment trial (comparing two chemotherapy regimens), there was no statistical difference in either DFS or OS between the treatment arms (27). The Kaplan-Meier method was used to describe the distribution of survival time according to KRAS status, and the log-rank test was carried out. We used stage-matched (or stratified) Cox proportional hazard models to calculate the HR of events according to tumoral KRAS status, adjusted for age at study entry (as a continuous variable), gender, baseline body mass index (≥30 versus <30 kg/m²), baseline performance status (0 versus 1-2), presence of bowel perforation or obstruction at time of surgery, treatment arm, tumor location (proximal versus distal), and MSI status (high versus low/MSS). Tumor stage (IIIA, IIIB, IIIC, or III unspecified substage) was used as a matching (stratifying) variable (with the “strata” option in the SAS “proc phreg” command) to minimize residual confounding. The proportionality of hazards assumption was satisfied by evaluating time-dependent variables, which were the cross-product of the KRAS variable and survival time (P = 0.10 for DFS, P = 0.06 for RFS, P = 0.24 for OS). Covariates with missing variables, including body mass index (1.2% missing), tumor location (1.0% missing), performance status (0.8% missing), perforation status (1.8% missing), and MSI status (5.5% missing), were coded with separate “missing” indicator variables in adjusted models. We assigned three cases (0.6%) with missing information in obstruction status as “no obstruction.” We confirmed that excluding cases with missing information in any of the covariates did not substantially alter results (data not shown). An interaction was assessed by including the cross product of the KRAS variable and another variable of interest in a multivariate Cox model, and the Wald test was done: P values were conservatively interpreted, considering multiple hypothesis testing. To assess an interaction of KRAS and stage, we dichotomized American Joint Committee on Cancer stage (IIIA-IIIB, N₁ versus IIIC, N₂) as well as assessed an interaction with T stage (T_1-2 versus T_3-4). In addition to obtaining a P value for interaction, we did stratified analysis to assess potential differential effect of KRAS mutation, in which we assessed the effect of KRAS mutation simultaneously in two or more strata (of a variable of interest) in a single Cox regression model (31, 32).

As part of the quality assurance program of the CALGB, members of the Audit Committee visit all participating institutions at least once every three years to review source documents. The auditors verify compliance with federal regulations and protocol requirements, including those pertaining to eligibility, treatment, adverse events, tumor response, and outcome in a sample of protocols at each institution. Such on-site review of medical records was done for a subgroup of 328 patients (26%) of the 1,264 patients under this study.

Study participants were drawn from a multicenter study of postoperative adjuvant chemotherapy in patients with stage III colon cancer who underwent a curative-intent surgical resection. We included 508 cases in this study based on availability of tumor tissue for KRAS sequencing, which detected a KRAS mutation in 178 (35%) patients. Identified KRAS mutations were as follows: 56 cases with codon 12 GGT > GAT (p.G12D, c.35G > A); 52 with codon 13 GGC > GAC (p.G13D, c.38G > A); 32 with codon 12 GGT > GTT (p.G12V, c.35G > T); 21 with codon 12 GGT > TGT (p.G12C, c.34G > T); 9 with codon 12 GGT > GCT (p.G12A, c.35G > C); and 8 with codon 12 GGT > AGT (p.G12S, c.34G > A). Table 1 summarizes the baseline characteristics of study subjects according to KRAS mutational status. Patients with a mutation in KRAS were significantly less likely to possess MSI or receive IFL as compared with FU/LV.

We assessed the influence of KRAS mutational status on clinical outcome in the 508 patients with stage III colon cancers. With median follow-up of 6.2 years among surviving participants, there were 196 events for DFS analysis, 180 events for RFS analysis, and 149 events for OS analysis. In Kaplan-Meier analysis, there were no significant differences in survival time distributions between patients with KRAS mutations and those with wild-type KRAS (log-rank P = 0.89 for DFS; Fig. 1; log-rank P = 0.84 for RFS; log-rank P = 0.56 for OS). DFS at 5 years was 62% for KRAS-mutated and 63% for KRAS-wild-type patients. RFS at 5 years was 64% for KRAS-mutated and 66% for KRAS-wild-type patients. Finally, OS at 5 years was 75% for KRAS-mutated and 73% for KRAS-wild-type patients.

In a univariate Cox regression analysis, when compared with KRAS-wild-type patients, KRAS-mutated patients did not experience a significant difference in DFS (HR, 0.98; 95% CI, 0.73-1.31), RFS (HR, 0.97; 95% CI, 0.71-1.32), or OS (HR, 0.90; 95% CI, 0.64-1.27; Table 2). These findings persisted in multivariate analysis that adjusted for clinical, pathologic, or molecular predictors of patient outcome, and no substantial confounding was identified.

We assessed whether the effect of KRAS mutational status on patient outcome was modified by adjuvant chemotherapy (Table 3). In both treatment arms (FU/LV and IFL), the presence of a mutation in KRAS was not associated with any significant difference in patient survival. Moreover, statistical tests for interaction failed to show any significant interaction between chemotherapy assignment and KRAS mutational status (P for interaction = 0.64, 0.67, and 0.60 for DFS, RFS, and OS, respectively).

We also assessed whether the effect of adjuvant chemotherapy arm on patient survival was modified by KRAS mutational status (Table 4). In both KRAS-wild-type and KRAS-mutated cases, there were no significant differences in DFS, RFS, or OS between the two treatment arms.

Finally, we examined whether there was significant modifying effect on the relation between KRAS mutation and clinical outcome by any of the other covariates (age, gender, body mass index, baseline performance status, tumor location, T stage, N stage, stage III substage, status of bowel perforation or obstruction, and MSI status). There was no evidence of significant effect modification by any of the variables examined (all P_interaction > 0.23).

In this study of 508 patients with stage III colon cancer treated with surgery and adjuvant chemotherapy, KRAS mutational status was not associated with any significant influence on cancer recurrence or death. These results were not materially altered in multivariate analyses that adjusted for other predictors for patient outcome. Moreover, the effect of KRAS mutation on patient survival did not significantly differ according to clinical features, chemotherapy arm, or MSI status, and the effect of adjuvant chemotherapy arm did not differ according to KRAS status. In separate independent cohort studies (6, 33), we previously showed that KRAS mutation was not significantly associated with survival of colon cancer patients in univariate analysis as well as multivariate analysis that adjusted for tumor stage, MSI, BRAF mutation, and other related molecular features. Thus, together with our previous data, our current data do not support a substantial prognostic role of KRAS mutation in colon cancer.

Although KRAS mutation does not seem to be a significant prognostic marker in colon cancer, its importance in colorectal carcinogenesis has been well documented. KRAS is one of the most commonly mutated oncogenes in human cancer. KRAS mutation activates the RAS-RAF pathway as well as the phosphoinositide 3-kinase-AKT pathway, leading to cellular growth and proliferation (2). Indeed, KRAS and PIK3CA mutations are associated with each other in colorectal cancer (34, 35), and KRAS and PIK3CA mutations seem to interact in survival analysis (33). Recently, a link between KRAS mutation and epigenomic aberrations in colorectal cancer has been suggested (31, 36–38). Specifically, KRAS mutation has been associated with low-level CpG island methylator phenotype (31, 36, 38, 39), and this relation has been shown in another independent dataset (22). In contrast to somatic mutations including those in KRAS, epigenomic aberrations are potentially reversible. Although a mechanistic link between epigenomics and KRAS mutation remains uncertain, analysis of KRAS mutation in colon cancer may shed light on epigenomic aberrations in cancer and provide targeted therapeutic opportunities.

Studying patient outcome has been an important area in cancer research. Accumulating evidence suggests KRAS mutational status is a critical biomarker to predict response or resistance to anti-EGFR targeted therapy in patients with metastatic colorectal cancer. Retrospective observational studies (7–12) as well as randomized controlled trials (13–17) have consistently shown that KRAS mutation in stage IV colorectal cancer confers resistance to cetuximab or panitumumab treatment. Thus, KRAS mutation testing is rapidly emerging as a routine clinical test for patients with metastatic colorectal cancers who are potential candidates for treatment with either cetuximab or panitumumab (1, 2, 40).

In contrast to anti-EGFR targeted therapy, the role of KRAS mutation in predicting response to other therapies remains unclear. For example, a couple of previous studies have examined the relationship between KRAS mutation and response to bevacizumab, and have shown that KRAS mutation does not predict response or resistance to bevacizumab in colon cancer (25, 41).

Although the “predictive” role for KRAS mutational testing in defining sensitivity to anti-EGFR targeted therapy in stage IV colorectal cancer is now widely accepted, the “prognostic” role for KRAS as an independent predictor of survival in patients with colorectal cancer remains less conclusive (18–20). Previous meta-analyses (RASCAL and RASCAL II; refs. 42, 43) showed that KRAS mutation was associated with worse outcome in colorectal cancer. However, these meta-analyses substantially suffered from publication bias; especially as most studies used were relatively small (N <150 in most studies; N <290 in all included studies). Compared with small studies with significant results, small studies with null results were more likely unpublished, and thus more likely excluded from these meta-analyses. Larger studies (e.g., N >290) have tended to show no independent prognostic significance of KRAS mutation in colorectal cancer. A large population-based study of 569 colorectal cancer patients reported that KRAS mutation was independently associated with worse survival (22), whereas most other large studies found no independent prognostic role of KRAS mutation (4, 6, 23–25, 44), including a recent study on 1,379 stage II-III colon cancers (26). Our current findings were limited to only stage III colon cancers. Nonetheless, our results are consistent with most previous large studies on colon cancers including stage III and other stages (4, 6, 23–26, 44). Moreover, although one small study of 35 patients suggested that KRAS mutational status influenced irinotecan sensitivity (45), KRAS mutational status did not seem to modify the influence of irinotecan-based adjuvant therapy in our trial.

There are several advantages in examining associations of molecular markers with outcome of patients in a NCI-sponsored clinical trial of adjuvant chemotherapy. All patients had stage III colon cancer, thus reducing the impact of heterogeneity by disease stage. Moreover, treatment and follow-up care were all standardized within the clinical trial, and the date and nature of recurrence were prospectively recorded. In addition, detailed information on other prognostic variables was routinely collected at study entry.

We recognize that patients who enroll in randomized trials may differ from the population-at-large. To participate, patients must meet eligibility criteria, be selected as appropriate candidates, and be motivated to participate. In addition, patients were particularly selected for this study on the basis of availability of colon cancer tissue specimens. Nonetheless, the demographic data of the patients in this study did not suggest significant selection bias. Moreover, because the study included patients from both community and academic centers across North America, our findings should reflect the general population of stage III patients in North America. In addition, although data on KRAS mutational status were available on a subset of patients enrolled in the trial, baseline characteristics and patient survival did not differ for patients with and without available archived tumor tissue in this trial.

In conclusion, we found that KRAS mutational status did not significantly predict clinical outcome in this study of stage III colon cancer patients. Although KRAS mutational testing should be routinely utilized to assess for appropriate use of anti-EGFR therapy in advanced colorectal cancer, KRAS status is unlikely to meaningfully predict patient prognosis.

L. Saltz has received research funding from Pfizer, Roche, Genentech, Bristol Myers Squibb, Imclone, and Amgen, and is a consultant to Pfizer, Roche, Genentech, Bristol Myers Squibb, Imclone, Amgen, Genzyme and Genomic Health; A. Benson has received research funding from Pfizer, Imclone, Bristol Myer Squibb, Amgen, and Sanofi Aventis and is a scientific advisor for Pfizer, Imclone, Bristol Myer Squibb, Amgen, and Sanofi Aventis.

We thank the CALGB Pathology Coordinating Office at the Ohio State University for banking and preparing the materials for the study. The following institutions participated in this study:

• Baptist Cancer Institute CCOP, Memphis, TN - Lee S. Schwartzberg, M.D., supported by CA71323

• Christiana Care Health Services, Inc. CCOP, Wilmington, DE - Stephen Grubbs, M.D., supported by CA45418

• Dana-Farber Cancer Institute, Boston, MA - Eric P. Winer, M.D., supported by CA32291

• Dartmouth Medical School - Norris Cotton Cancer Center, Lebanon, NH - Marc S. Ernstoff, M.D., supported by CA04326

• Duke University Medical Center, Durham, NC - Jeffrey Crawford, M.D., supported by CA47577

• Georgetown University Medical Center, Washington, DC - Minetta C. Liu, M.D., supported by CA77597

• Cancer Centers of the Carolinas, Greenville, SC - Jeffrey K. Giguere, M.D., supported by CA29165

• Hematology-Oncology Associates of Central New York CCOP, Syracuse, NY - Jeffrey Kirshner, M.D., supported by CA45389

• Long Island Jewish Medical Center, Lake Success, NY - Kanti R. Rai, M.D., supported by CA11028

• Massachusetts General Hospital, Boston, MA - Jeffrey W. Clark, M.D., supported by CA12449

• Memorial Sloan-Kettering Cancer Center, New York, NY - Clifford A. Hudis, M.D., supported by CA77651

• Missouri Baptist Medical Center, St. Louis, MO - Alan P. Lyss, M.D., supported by CA114558-02

• Mount Sinai Medical Center, Miami, FL - Rogerio C. Lilenbaum, M.D., supported by CA45564

• Mount Sinai School of Medicine, New York, NY - Lewis R. Silverman, M.D., supported by CA04457

• Nevada Cancer Research Foundation CCOP, Las Vegas, NV - John A. Ellerton, M.D., supported by CA35421

• North Shore-Long Island Jewish Health System, New Hyde Park, NY - Daniel Budman, M.D., supported by CA35279

• Rhode Island Hospital, Providence, RI - William Sikov, M.D., supported by CA08025

• Roswell Park Cancer Institute, Buffalo, NY - Ellis Levine, M.D., supported by CA02599

• Southeast Cancer Control Consortium Inc. CCOP, Goldsboro, NC - James N. Atkins, M.D., supported by CA45808

• State University of New York Upstate Medical University, Syracuse, NY - Stephen L. Graziano, M.D., supported by CA21060

• The Ohio State University Medical Center, Columbus, OH - Clara D. Bloomfield, M.D., supported by CA77658

• University of California at San Diego, San Diego, CA - Barbara A. Parker, M.D., supported by CA11789

• University of California at San Francisco, San Francisco, CA - Alan P. Venook, M.D., supported by CA60138

• University of Chicago, Chicago, IL - Gini Fleming, M.D., supported by CA41287

• University of Illinois MBCCOP, Chicago, IL - Lawrence E. Feldman, M.D., supported by CA74811

• University of Iowa, Iowa City, IA - Daniel A. Vaena, M.D., supported by CA47642

• University of Maryland Greenebaum Cancer Center, Baltimore, MD - Martin Edelman, M.D., supported by CA31983

• University of Massachusetts Medical School, Worcester, MA - William V. Walsh, M.D., supported by CA37135

• University of Minnesota, Minneapolis, MN - Bruce A Peterson, M.D., supported by CA16450

• University of Missouri/Ellis Fischel Cancer Center, Columbia, MO - Michael C. Perry, M.D., supported by CA12046

• University of Nebraska Medical Center, Omaha, NE - Anne Kessinger, M.D., supported by CA77298

• University of North Carolina at Chapel Hill, Chapel Hill, NC - Thomas C. Shea, M.D., supported by CA47559

• University of Tennessee Memphis, Memphis, TN - Harvey B. Niell, M.D., supported by CA47555

• University of Vermont, Burlington, VT - Hyman B. Muss, M.D., supported by CA77406

• Wake Forest University School of Medicine, Winston-Salem, NC - David D Hurd, M.D., supported by CA03927

• Walter Reed Army Medical Center, Washington, DC - Thomas Reid, M.D., supported by CA26806

• Washington University School of Medicine, St. Louis, MO-Nancy Bartlett, M.D., supported by CA77440

• Weill Medical College of Cornell University, New York, NY - John Leonard, M.D., supported by CA07968