Abstract
Background
KRAS and BRAF V600E mutations are important predictive and prognostic markers, respectively, in colon cancer, but little is known about patient and clinical factors associated with them.
Methods
Two thousand three hundred twenty-six of 3397 patients in the N0147 phase III adjuvant trial for stage III colon cancer completed a patient questionnaire. Primary tumors were assessed for KRAS and BRAF V600E mutations and defective mismatch repair (dMMR) status. Logistic regression models and categorical data analysis were used to identify associations of patient and tumor characteristics with mutation status. All statistical tests were two-sided.
Results
KRAS (35%) and BRAF V600E (14%) mutations were nearly mutually exclusive. KRAS mutations were more likely to be present in patients without a family history of colon cancer and never smokers. Tumors with KRAS mutations were less likely to have dMMR (odds ratio [OR] = 0.21; 95% confidence interval [CI] = 0.15 to 0.31; P < .001) and high-grade histology (OR = 0.73; 95% CI = 0.59 to 0.92; P < .001) but were more often right-sided. Among KRAS-mutated tumors, those with a Gly13Asp mutation tended to have dMMR and high-grade histology. Tumors with BRAF V600E mutations were more likely to be seen in patients who were aged 70 years or older (OR = 3.33; 95% CI = 2.50 to 4.42; P < .001) and current or former smokers (OR = 1.64; 95% CI = 1.26 to 2.14; P < .001) but less likely in non-whites and men. Tumors with BRAF V600E mutations were more likely to be right-sided and to have four or more positive lymph nodes, high-grade histology, and dMMR.
Conclusions
Specific patient and tumor characteristics are associated with KRAS and BRAF V600E mutations.
Colon cancer is the fourth most common malignancy in the United States, with more than 100000 new case patients diagnosed each year (1). Its pathogenesis involves the accumulation of genetic and epigenetic modifications that regulate proliferation, apoptosis and angiogenesis (2). The activation of the epidermal growth factor receptor (EGFR) signaling cascade is a well-described pathway leading to colon tumorigenesis (3); moreover, mutations within the KRAS and BRAF proto-oncogenes located downstream to EGFR within this pathway lead to its constitutive activation (4,5). KRAS mutations are present in approximately 35% to 40% of colon cancers, with roughly two-thirds of these mutations in codon 12 and one-third in codon 13 (6). The presence of a KRAS mutation is predictive for resistance to anti-EFGR monoclonal antibodies (mAbs) in advanced colon cancer (7,8). However, the biological and functional consequences of KRAS mutations at codon 12 may be different than those at codon 13. It has been suggested that patients whose tumors harbor a KRAS Gly13Asp mutation may benefit from anti-EGFR mAb therapy (9–11). BRAF activating mutations, specifically V600E, occur in less than 10% of patients with sporadic colon cancer (12) and are a strong negative prognostic marker (13); however, their predictive value for efficacy of anti-EGFR mAb treatment is less certain (14–16).
To date, few studies have evaluated the associations of epidemiologic factors with these mutations. Hence, using prospectively collected patient questionnaire (PQ) data from a large phase III clinical trial (N0147), along with the corresponding KRAS and BRAF V600E mutational status of the patients’ primary colon tumors, we described patient and tumor characteristics associated with these mutations.
Methods
The clinical trial N0147 was designed by the North Central Cancer Treatment Group in collaboration with the National Cancer Institute and the National Cancer Institute–sponsored cooperative groups. This trial evaluated different chemotherapy regimens with or without the addition of cetuximab, an anti-EGFR mAb, for the adjuvant treatment of patients with node-positive colon cancer. Patients were required to have complete en bloc resection of histologically proven stage III (any T, N1–2, M0 tumors) colon adenocarcinoma that was at least 12cm from the anal verge. Before enrollment, each participant signed an institutional review board–approved, protocol-specific informed consent in accordance with federal and institutional guidelines. The results of this trial were reported and published separately by Alberts et al. (17). Patients were classified based on race as whites and as non-whites, including American Indian, Alaskan native, Asian, black or African American, and Native Hawaiian or other Pacific Islander.
Patient Questionnaire
The PQ form (Supplementary Methods, available online) was part of the N0147 trial from its inception in February 2004 until discontinuation in August 2008, and its completion was required for all patients at the time of enrollment. These forms were scanned, aggregated and merged with clinical characteristics collected at the time of randomization, as well as KRAS, BRAF V600E, and mismatch repair (MMR) results.
The following data fields were recorded by the PQ form and included in the analysis: activity level (almost no activity [mainly sitting]; mild activity [walking for short periods, but not lifting or carrying], moderate activity [walking quite a lot, but not lifting or carrying], heavy activity [usually running, lifting, or carrying]), aspirin/nonsteroidal anti-inflammatory drug use, smoking history (never smoker: less than 100 cigarettes in a lifetime; former smoker: more than 100 cigarettes in a lifetime but has quit; current smoker: more than 100 cigarettes in a lifetime and still smoking), alcohol intake (never: less than one drink per month [one drink is defined as one beer, mixed drink, or a 5-oz. glass of wine]; former: one or more drinks per month but has quit; current: one or more drinks per month and continues to drink), body mass index (obese: ≥30kg/m2; overweight: 25–29.9kg/ m2; normal: 18.5–24.9kg/ m2; underweight: <18.5kg/m2), family history of colorectal cancer (in parents, children, or siblings), and history of chronic gastrointestinal diseases (Gastroesophageal reflux or Barrett’s esophagus, Helicobacter pylori infection, Celiac sprue, Crohn’s disease, ulcerative colitis, microscopic colitis, or irritable bowel syndrome).
KRAS and BRAF V600E Mutation Status
Assessment of KRAS and BRAF V600E (NCBI Entrez Gene 673) mutational status was performed centrally at the Mayo Clinic, Rochester, Minnesota, in a Clinical Laboratory Improvement Amendments–compliant laboratory. DNA required for the mutation testing was extracted from paraffin-embedded tumor tissue. KRAS testing was performed with the DxS mutation test kit KR-03/04 (DxS, Manchester, UK), together with the Light-Cycler 480 (Roche Applied Sciences, Basel, Switzerland), which assesses seven different potential mutations in codons 12 and 13 (Gly12Ala, Gly12Asp, Gly12Arg, Gly12Cys, Gly12Ser, Gly12Val, and Gly13Asp). Tumors with any of the aforementioned KRAS mutations were classified as mutant KRAS, whereas the rest were classified as wild-type KRAS. Assessment for the BRAF V600E mutation was performed using a Mayo-developed multiplex allele-specific polymerase chain reaction–based assay. After amplification, polymerase chain reaction products were analyzed on an ABI 3130xl instrument (Life Technologies, Applied Biosystems, Grand Island, NY) and scored for the presence or absence of the V600E variant only. Tumors with the BRAF V600E mutation were classified as mutant BRAF (vs wild-type).
Expression of DNA MMR Proteins
Immunohistochemical analyses of the protein expression of three DNA MMR proteins (ie, MLH1, MSH2, and MSH6) using commercially available antibodies were performed and graded as previously described (18). The MMR status was defined as proficient if all three proteins were detected; otherwise, the tumor was classified as dMMR (deficient).
Statistical Analysis
The primary goal of this analysis was to identify epidemiologic and clinicopathologic factors associated with KRAS and BRAF V600E mutational status. Logistic regression models were used to detect associations of these characteristics with each of the mutations. The distributions of patient characteristics, risk, and tumor and clinicopathologic factors were initially evaluated using summary statistics for continuous variables (eg, means, medians), categorical data analysis [eg, Kruskal Wallis (19), χ2 (19), or Fisher exact test (20)], and univariate logistic regression models to further categorize and define the final covariables used for multivariable analysis. Statistical tests were two-sided, and P values of .05 were considered significant. P values less than .003 (ie, 0.05/16) could be considered significant when accounting for tests of association of 16 individual factors with each of KRAS and BRAF V600E. However, in this study no adjustment was considered for the P values obtained from the univariate models identifying candidate variables to be used in multivariable models because the latter inherently adjusted for all covariables within a model. Furthermore, P values were not adjusted for analyses of KRAS submutations because these were secondary endpoints. Statistically significant (ie, unadjusted P < .05) characteristics based on univariate models using the 16 individual factors were then included in multivariable models using stepwise and backwards model selection procedures. Both modeling procedures resulted in a final multivariable model for each mutation and provided estimates of odds ratios (ORs) and confidence intervals (CIs) for each statistically significant patient, risk, tumor, and clinicopathologic characteristic associated with the mutational status.
Data collection and quality assurance were managed by the Clinical Trials Support Unit until September 2011; thereafter, all data collection, quality assurance, and statistical analyses were conducted by the North Central Cancer Treatment Group Statistics and Data Center. Analyses were performed using intention-to-treat principles. The data analysis for this article was generated using SAS/STAT software (version 9.3 of the SAS System for LINUX (SAS Institute, Cary, NC).
Results
Epidemiologic and Clinicopathologic Factors
A total of 3397 patients were registered onto the clinical trial, 2502 of whiom enrolled during the time when PQ data were collected at registration (this was discontinued on August 18, 2008). The completed PQ form was available for 2326 (93%) patients. Of these, 2222 (96%) and 2166 (93%) tumors yielded KRAS and BRAF V600E mutation status, respectively. There were 783 (35%) tumors that had KRAS mutations, of which 191 (24%) were KRAS Gly13Asp, whereas 310 (14%) tumors had a BRAF V600E mutation. Of 2231 tumors analyzed for MMR status, 279 (13%) were characterized as dMMR. The distribution and frequencies of the various patient, tumor, and clinicopathologic characteristics are summarized in Table 1. Risk factors are detailed in Table 2, and specific KRAS mutations are enumerated in Supplementary Table 1 (available online).
Table 1.
Distributions of patients, tumor, and clinicopathologic characteristics by KRAS and BRAF mutation status*
| Characteristic | Wild-type KRAS (n = 1439) | Mutant KRAS (n = 783) | P | Wild-type BRAF (n = 1856) | Mutant BRAF (n = 310) | P |
|---|---|---|---|---|---|---|
| Age, y | .07† | <.001† | ||||
| Mean (SD) | 58.5 (11.4) | 57.7 (10.9) | 57.0 (11.2) | 65.2 (8.4) | ||
| Median | 59.0 | 59.0 | 58.0 | 66.0 | ||
| Range | (19.0–86.0) | (23.0–85.0) | (19.0–86.0) | (41.0–86.0) | ||
| Age, y | .03‡§ | <.001‡ | ||||
| <70 | 1183 (82%) | 671 (86%) | 1602 (86%) | 205 (66%) | ||
| ≥70 | 256 (18%) | 112 (14%) | 254 (14%) | 105 (34%) | ||
| Sex | .5511‡ | <.001‡ | ||||
| Male | 767 (53%) | 407 (52%) | 1025 (55%) | 114 (37%) | ||
| Female | 672 (47%) | 376 (48%) | 831 (45%) | 196 (63%) | ||
| No. of positive nodes | .07‡ | <.001‡ | ||||
| <4 | 840 (58%) | 488 (62%) | 1138 (61%) | 158 (51%) | ||
| ≥4 | 599 (42%) | 295 (38%) | 718 (39%) | 152 (49%) | ||
| Grade‖ | <.001‡ | <.001‡ | ||||
| Low | 1045 (73%) | 621 (79%) | 1450 (78%) | 166 (54%) | ||
| High | 394 (27%) | 162 (21%) | 406 (22%) | 144 (47%) | ||
| T-stage | .26‡ | .04‡§ | ||||
| T1–T2 | 191 (13%) | 120 (15%) | 269 (15%) | 32 (10%) | ||
| T3 | 1097 (76%) | 572 (73%) | 1393 (75%) | 235 (76%) | ||
| T4 | 151 (10%) | 90 (12%) | 193 (10%) | 43 (14%) | ||
| Race | .46‡ | <.001‡ | ||||
| White | 1253 (87%) | 673 (86%) | 1585 (85%) | 292 (94%) | ||
| Non-white | 186 (13%) | 110 (14%) | 271 (15%) | 18 (6%) | ||
| Tumor site¶ | <.001‡ | <.001‡ | ||||
| Left, distal | 721 (51%) | 306 (40%) | 959 (52%) | 43 (14%) | ||
| Right, proximal | 698 (49%) | 466 (60%) | 870 (48%) | 263 (86%) | ||
| BMI, kg/m2 | .23‡ | .68‡ | ||||
| <25 | 403 (28%) | 238 (30%) | 542 (29%) | 87 (28%) | ||
| ≥25 | 1036 (72%) | 545 (70%) | 1314 (71%) | 223 (72%) | ||
| BRAF status | — | — | — | — | ||
| Wild type | 1095 (78%) | 756 (100%) | ||||
| Mutant | 307 (22%) | 1 (0%) | ||||
| KRAS status | — | — | — | — | ||
| Wild type | 1095 (59%) | 307 (100%) | ||||
| Mutant | 756 (41%) | 1 (0%) | ||||
| MMR | <.001‡ | <.001‡ | ||||
| Proficient | 1179 (83%) | 738 (95%) | 1705 (93%) | 163 (53%) | ||
| Deficient | 235 (17%) | 38 (5%) | 125 (7%) | 144 (47%) |
* BMI = body mass index; MMR = mismatch repair; SD = standard deviation.
† Two-sided Kruskal Wallis test.
‡ Two-sided χ2 test.
§ P value would not be statistically significant if adjusted for tests of association of 16 individual factors requiring P < .003 to be considered statistically significant.
‖ Low grade: grade 1 and 2 histology. High grade: grade 3 and 4 histology.
¶ Left: splenic flexure, descending colon, sigmoid colon. Right: cecum, ascending colon, hepatic flexure, transverse colon. Patients that have sites on both the left and right were not assigned a side.
Table 2.
Distributions of patient questionnaire risk characteristics by KRAS and BRAF mutation status*
| Characteristic | Wild-type KRAS (n = 1439) | Mutant KRAS (n = 783) | P | Wild-type BRAF (n = 1856) | Mutant BRAF (n = 310) | P |
|---|---|---|---|---|---|---|
| Smoking status Never | .01†‡ | <.001† | ||||
| Former/current | 637 (45%) 792 (56%) |
390 (50%) 390 (50%) |
886 (48%) 960 (52%) |
115 (38%) 192 (62%) |
||
| Physical activity§ | .55† | .054† | ||||
| Almost no activity | 63 (4%) | 43 (6%) | 91 (5%) | 13 (4%) | ||
| Mild activity | 585 (41%) | 319 (41%) | 736 (40%) | 141 (46%) | ||
| Moderate activity | 602 (42%) | 312 (40%) | 766 (42%) | 128 (42%) | ||
| Heavy activity | 170 (12%) | 99 (13%) | 240 (13%) | 25 (8%) | ||
| Family history of CRC | <.001† | .28† | ||||
| Yes | 218 (15%) | 75 (10%) | 240 (13%) | 47 (15%) | ||
| No | 1221 (85%) | 708 (90%) | 1616 (87%) | 263 (85%) | ||
| Alcohol history | .41† | .22† | ||||
| Never | 438 (31%) | 253 (32%) | 569 (31%) | 106 (34%) | ||
| Former/Current | 991 (69%) | 529 (68%) | 1278 (69%) | 203 (66%) | ||
| Chronic GI conditions | .64† | .02†‡ | ||||
| Yes | 228 (16%) | 130 (17%) | 287 (15%) | 65 (21%) | ||
| No | 1211 (84%) | 653 (83%) | 1569 (85%) | 245 (79%) | ||
| Diabetes mellitus | .71† | .36† | ||||
| Yes | 130 (9%) | 67 (9%) | 162 (9%) | 32 (10%) | ||
| No | 1309 (91%) | 716 (91%) | 1694 (91%) | 278 (90%) | ||
| Exposure to NSAIDs/ASA | .10† | .16† | ||||
| Yes | 791 (55%) | 402 (51%) | 992 (53%) | 179 (58%) | ||
| No | 648 (45%) | 381 (49%) | 864 (47%) | 131 (42%) |
* ASA = acetylsalicylic acid (aspirin); CRC = colorectal cancer; GI = gastrointestinal; NSAID = nonsteroidal anti-inflammatory drug.
† Two-sided χ2 test.
‡ P value would not be statistically significant if adjusted for tests of association of 16 individual factors requiring P < .003 to be considered statistically significant.
§ Almost no activity (mainly sitting), mild activity (walking for short periods, but not lifting or carrying), moderate activity (walking quite a lot, but not lifting or carrying), heavy activity (usually running, lifting, or carrying).
The majority of patients were aged less than 70 years (83%), were white (87%), and had a BMI greater than 25kg/m2 (71%); 75% had T3 primary tumors, and 46% of tumors were located distally. Compared with patients with mutant KRAS tumors, those with wild-type KRAS tumors were more likely to be aged 70 years or older (185 vs. 14%; P = .03), to be former/current smokers (56% vs 50%; P = .01), and to have a first-degree relative with a history of colorectal cancer (15% vs 10%; P < .001). Patients with wild-type KRAS tumors were also more likely to have dMMR status (17% vs 5%; P < .001), high-grade histology (27% vs 21%; P < .001), and a distal tumor (51% vs 40%; P < .001) than those with KRAS-mutated tumors (dMMR: OR = 0.21, 95% CI = 0.15 to 0.31, P < .001; high-grade histology: OR = 0.73, 95% CI = 0.59 to 0.92, P < .001). When compared with tumors with other KRAS mutations, tumors with a KRAS Gly13Asp mutation were more likely to have high-grade histology (28% vs 18%; P = .006) and dMMR status (9% vs 4%; P = .003) status.
When compared with those without BRAF V600E-mutated tumors, patients with BRAF V600E-mutated tumors were more likely to be aged 70 years or older (34% vs 14%; OR = 3.33; 95% CI = 2.50 to 4.42; P < .001), female (63% vs 45%; P < .001), white (94% vs 85%; P < .001), and a current/former smoker (62% vs 52%; OR = 1.64; 95% CI = 1.26 to 2.14; P < .001). Tumors with BRAF V600E mutations were more likely to be associated with dMMR status (47% vs 7%; P < .001), four or more positive nodes (49% vs 39%; P < .001), high-grade histology (47% vs 22%; P < .001), proximal location (86% vs 48%; P < .001), and T4 depth of invasion (14% vs 10%; P = .04).
Associations Between Factors and KRAS or BRAF Mutations
Univariate logistic regression models identified the following factors as statistically significantly associated with having wild-type KRAS status: age of 70 years or older, current/former smokers, 30 or more pack-years of smoking, family history of colorectal cancer, and tumors classified as either high grade, left-sided, or dMMR (Figure 1A). BRAF V600E mutations were statistically significantly associated with advanced age, female sex, white race, four or more positive nodes, high grade, stage T4, proximal tumors, current/former smoking, 30 or more pack-years of smoking, conditions of the esophagus and colon, and dMMR (Figure 1B).
Figure 1.
Forest plots of univariate logistic model associations with KRAS mutation status (A) and BRAF mutation status (B). P values are for two-sided Wald χ2 test. CI = confidence interval; CRC = colorectal cancer; dMMR = deficient mismatch repair; GI = gastrointestinal; LCL = lower confidence limit; OR = odds ratio; UCL = upper confidence limit.
In the analysis using multivariable logistic regression models, we reviewed models that included tumor and clinicopathologic characteristics separate from the epidemiologic factors (Table 3). As shown univariably, tumors with BRAF V600E mutations were statistically significantly associated with dMMR status, proximal (right) tumor location, high histologic grade, and four or more positive lymph nodes for tumor factors. Epidemiologic factors associated with BRAF V600E mutations included age of 70 years or older, white race, female sex, and current/former smoking history. Patients with mutated KRAS tumors were statistically significantly less likely to have a first-degree relative with colorectal cancer, be a current/former smoker, or have tumors that are left-sided, have a high histologic grade, or have dMMR status.
Table 3.
Multivariable logistic regression model associations between patient, patient questionnaire risk, and tumor and KRAS or BRAF mutations*
| Factor | Mutant KRAS | Mutant BRAF | ||
|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | |
| Patient and PQ risk characteristics | ||||
| Smoker: former/current (referent: never) | 0.80 (0.67 to 0.96) | .01‡ | 1.64 (1.26 to 2.14) | <.001 |
| Family history of CRC (referent: no history of CRC | 0.59 (0.45 to 0.78) | <.001 | — | — |
| Aged ≥70 y | — | — | 3.33 (2.50 to 4.42) | <.001 |
| Non-white (referent: white) | — | — | 0.33 (0.20 to 0.57) | <.001 |
| Male sex (referent: female sex) | — | — | 0.41 (0.31 to 0.53) | <.001 |
| Tumor characteristics† | ||||
| Right side, proximal (referent: left side, distal) | 2.05 (1.70 to 2.47) | <.001 | 4.01 (2.81 to 5.71) | <.001 |
| High grade (referent: low grade) | 0.73 (0.59 to 0.92) | .007‡ | 1.71 (1.28 to 2.28) | <.001 |
| dMMR (referent: pMMR) | 0.21 (0.15 to 0.31) | <.001 | 7.75 (5.66 to 10.61) | <.001 |
| ≥4 lymph nodes (referent: 1–3 lymph nodes) | — | — | 1.73 (1.30 to 2.29) | <.001 |
* CI = confidence interval; CRC = colorectal cancer; dMMR = deficient mismatch repair; OR = odds ratio; pMMR = proficient mismatch repair; PQ = patient questionnaire.
† Mutant KRAS and mutant BRAF models exclude each other because of the near mutual exclusivity of mutations.
‡ P values would lose statistical significance in the multivariable analysis if adjustment for tests of association of 16 individual factors requiring P < .003 were to be considered statistically significant. All statistical tests were two-sided.
Discussion
The primary outcome (disease-free survival) of patients enrolled in clinical trial N0147 (17) and the prognostic impact of molecular features (KRAS, BRAF, MMR) in the patients’ resected colon tumors (21) have been previously reported. This article summarizes epidemiologic (ie, cigarette smoking) and clinicopathologic features (ie, MMR and tumor grade) associated with the KRAS and BRAF V600E mutation status of tumors in enrolled subjects.
Cigarette smoking history is a known risk factor for developing colon cancer (22). Studies have demonstrated that carcinogens present in tobacco smoke can induce cancer-related base substitutions such as G:C → A:T transitions in RAS oncogenes (23,24). However, our study found that colon cancers from patients with a history of current or former smoking were less likely to harbor a KRAS mutation. Similarly several population-based studies found cigarette smoking to be associated with colorectal cancers with wild-type copies of the KRAS gene (25–27). In contrast with KRAS mutations, our study reveals an association between current or former smoking history with the presence of BRAF V600E mutations in tumors. Previous studies in colorectal cancer have described this association of cigarette smoking with BRAF V600E mutations, in addition to finding these tumors to be CpG island methylator phenotype (CIMP)–high (28,29). Although the exact mechanism remains unknown, preclinical studies have shown that tobacco exposure can stimulate DNA methyltransferase activity that is associated with CIMP (30).
CIMP-high is detected in approximately 20% to 30% of colon cancers (31,32) and leads to transcriptional silencing of certain genes, including MLH1, that results in dMMR sporadically (33,34). Our study assessed the MMR status of tumors and found that 13% of the study cohort showed dMMR status, which is consistent with other studies (35,36). Most (90%) colon cancers with dMMR status result from the inactivation of MLH1 in a background of CIMP; the remaining cases are due to germline mutations in the MMR genes that produce Lynch syndrome (37,38). A strong association between BRAF V600E mutations and dMMR status was found in our study, and although this study had the largest number of dMMR cases reported, the estimated odds ratios and 95% confidence intervals reflect the relatively modest size of the sample. This analysis lacked results of the CIMP status of the tumors to associate it with MMR, BRAF V600E, and smoking history results; however, an expanded molecular analysis of this cohort is ongoing as additional support for this work is secured. Nevertheless, previous data indicate that the CIMP-high subgroup, which exhibits a very high frequency of cancer-specific DNA hypermethylation, is strongly associated with epigenetic inactivation of MLH1 and the BRAF V600E mutation (39). Therefore, the BRAF V600E mutation can serve as a surrogate marker for the CIMP-high group showing sporadic dMMR status. The CIMP-low subgroup has been shown to be enriched for KRAS mutations 39). Although the temporal order of molecular events associating smoking history with dMMR and CIMP-high colon cancers remains incompletely defined, BRAF mutant tumors appear to be strongly associated with these epigenetically mediated phenotypes. Further studies providing a molecular explanation for these associations are needed.
Consistent with other reports (28), patients with BRAF V600E-mutated tumors in our study were more likely to be aged 70 years or older, women, and of non-Hispanic white ethnicity. BRAF V600E-mutated tumors were also more likely to have four or more positive nodes and T4 stage of disease, which suggest a more aggressive biology, as evidenced in the literature (13). Both BRAF V600E- and KRAS-mutated tumors were more likely to be located in the proximal colon. To our knowledge, this is the first study to suggest that patients with KRAS-mutated tumors were less likely to have a family history of colorectal cancer. This seems contrary to previous studies that have suggested equal frequencies of KRAS mutations in sporadic and hereditary colon tumors, such as those seen in hereditary nonpolyposis colorectal cancer patients (40,41). dMMR status was less likely to be associated with KRAS mutations. However, if a tumor did have a KRAS mutation and showed dMMR, the mutation was more likely to be a KRAS Gly13Asp mutation (9% vs 4%). This reinforces the findings by Oliveira et al. (41) who demonstrated that sporadic dMMR tumors due to hypermethylation of the MLH1 promoter have a lower frequency of KRAS mutations compared with dMMR tumors with germline mutations or somatic mutations unrelated to MLH1. Among the latter groups, there was a higher frequency of Gly13Asp mutations compared with other KRAS mutations (41). It was also noticed that among KRAS-mutant tumors, there was a positive association between KRAS Gly13Asp mutations and high-grade histology (28% vs 18%).
Recent studies evaluating the pathogenesis of colon cancers identified a “serrated pathway of carcinogenesis” in which colon tumors that arise from precursor serrated polyps are typically characterized by a proximal colonic location, dMMR status, CIMP-high status, BRAF mutation-positive status, and an absence of KRAS mutations (42). This described phenotype matches many of the patients in our study whose tumors possess BRAF V600E mutations and supports the existence of this serrated pathway. Furthermore, because cigarette smoking has been associated with colon tumors bearing this phenotype, it has been suggested that cigarette smoking may affect colon cancer risk through the serrated pathway of carcinogenesis (29).
There are several limitations to our study. First, participants recollected the answers to several questions from memory when filling out the PQ form, hence possibly introducing reporting errors while classifying several patient risk characteristics such as smoking and alcohol history, physical activity levels, and family history of colorectal cancers. Our study lacked data on tumor CIMP status and cause of dMMR status (germline vs sporadic), which could have strengthened the associations and provided additional insights for some of the observations noted in our study. Finally, our population was mainly composed of non-Hispanic whites, and it is unknown whether the associations detected in our study would be similar for other ethnic groups, which should be further investigated in more diverse subject populations. Nevertheless, this is the largest prospective study that assessed both KRAS and BRAF V600E mutation status simultaneously to determine their independent associations with host and tumor characteristics. To our knowledge, this is also the first study to specifically look at the KRAS Gly13Asp mutation among KRAS-mutant tumors and describe its association with dMMR and high-grade histology colon tumors.
In conclusion, our study suggests that specific epidemiologic and clinicopathologic characteristics are associated with KRAS and BRAF V600E mutations. Specifically, age of 70 years or older, smoking, high-grade histology, and dMMR status were associated with a lower incidence of mutant KRAS tumors but a higher incidence of BRAF V600E-mutated tumors. Both mutations tend to be right-sided and nearly mutually exclusive, but BRAF V600E-mutated tumors are more common in females, non-Hispanic white patients, those having four or more positive lymph nodes, stage T4 disease, and those with a history of gastrointestinal conditions. Finally, within KRAS-mutant tumors, those with a KRAS Gly13Asp mutation tend to be associated with dMMR status and high-grade histology. Further studies are warranted to define the mechanism and temporal order of events brought about by the aforementioned epidemiologic and clinicopathologic characteristics that may explain their association with these specific mutations.
Funding
This trial was conducted as a collaborative trial of the North Central Cancer Treatment Group [NCCTG Legacy (Alliance)] and Mayo Clinic and was supported in part by Public Health Service grants CA-25224, CA-37404, CA-35103, CA-63844, CA-35113, CA-35272, CA-114740, CA-32102, CA-14028, CA49957, CA21115, CA31946, CA12027, CA37377 from the National Cancer Institute, Department of Health and Human Services. Bristol-Myers Squibb, ImClone, sanofi-aventis, and Pfizer provided unrestricted support to the NCCTG for conduct of this trial.
R. R. McWilliams, D. J. Sargent, R. M. Goldberg, and S. R. Alberts were responsible for study concept and design. R. R. McWilliams, M. R. Mahoney, and G. D. Nelson were responsible for analysis and interpretation of data. R. R. McWilliams, F. A. Sinicrope, W. I. Gonsalves, M. R. Mahoney, and G. D. Nelson were responsible for drafting of the manuscript. R. R. McWilliams, W. I. Gonsalves, D. J. Sargent, M. R. Mahoney, G. D. Nelson, S. R. Alberts, F. A. Sinicrope, P. J. Limburg, A. Grothey, J. M. Hubbard, R. M. Goldberg, E. Chan, J. L. Berenberg, and S. N. Thibodeau were responsible for critical revision of the manuscript for important intellectual content. M. R. Mahoney and G. D. Nelson were responsible for statistical analysis.
The authors declared on conflict of interest. The content is solely the responsibility of the authors and does not necessarily represent the views of the National Cancer Institute or the National Institutes of Health.
This work was previously presented at the Gastrointestinal Cancers Symposium of the American Society of Clinical Oncology, January 19–21, 2012, San Francisco, California.
References
- 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin. 2013;63(1):11–30 [DOI] [PubMed] [Google Scholar]
- 2. Vogelstein B, Fearon ER, Hamilton SR, et al. Genetic alterations during colorectal-tumor development. N Engl J Med. 1988;319(9):525–532 [DOI] [PubMed] [Google Scholar]
- 3. Fang JY, Richardson BC. The MAPK signalling pathways and colorectal cancer. Lancet Oncol. 2005;6(5):322–327 [DOI] [PubMed] [Google Scholar]
- 4. Schubbert S, Shannon K, Bollag G. Hyperactive Ras in developmental disorders and cancer. Nat Rev Cancer. 2007;7(4):295–308 [DOI] [PubMed] [Google Scholar]
- 5. Wan PTC, Garnett MJ, Roe SM, et al. Mechanism of activation of the RAF-ERK signaling pathway by oncogenic mutations of B-RAF. Cell. 2004;116(6):855–867 [DOI] [PubMed] [Google Scholar]
- 6. Neumann J, Zeindl-Eberhart E, Kirchner T, et al. Frequency and type of KRAS mutations in routine diagnostic analysis of metastatic colorectal cancer. Pathol Res Pract. 2009;205(12):858–862 [DOI] [PubMed] [Google Scholar]
- 7. Amado RG, Wolf M, Peeters M, et al. Wild-type KRAS is required for panitumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol. 2008;26(10):1626–1634 [DOI] [PubMed] [Google Scholar]
- 8. Van Cutsem E, Kohne CH, Hitre E, et al. Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer. New Engl J Med. 2009;360(14):1408–1417 [DOI] [PubMed] [Google Scholar]
- 9. De Roock W, Jonker DJ, Di Nicolantonio F, et al. Association of KRAS p.G13D mutation with outcome in patients with chemotherapy-refractory metastatic colorectal cancer treated with cetuximab. JAMA. 2010;304(16):1812–1820 [DOI] [PubMed] [Google Scholar]
- 10. Gajate P, Sastre J, Bando I, et al. Influence of KRAS p.G13D mutation in patients with metastatic colorectal cancer treated with cetuximab. Clin Colorect Cancer. 2012;11(4):291–296 [DOI] [PubMed] [Google Scholar]
- 11. Tejpar S, Celik I, Schlichting M, et al. Association of KRAS G13D tumor mutations with outcome in patients with metastatic colorectal cancer treated with first-line chemotherapy with or without cetuximab. J Clin Oncol. 2012;30(29):3570–3577 [DOI] [PubMed] [Google Scholar]
- 12. Jass JR. Classification of colorectal cancer based on correlation of clinical, morphological and molecular features. Histopathology. 2007;50(1):113–130 [DOI] [PubMed] [Google Scholar]
- 13. Yokota T, Ura T, Shibata N, et al. BRAF mutation is a powerful prognostic factor in advanced and recurrent colorectal cancer. Br J Cancer. 2011;104(5):856–862 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Bokemeyer C, Van Cutsem E, Rougier P, et al. Addition of cetuximab to chemotherapy as first-line treatment for KRAS wild-type metastatic colorectal cancer: pooled analysis of the CRYSTAL and OPUS randomised clinical trials. Eur J Cancer. 2012;48(10):1466–1475 [DOI] [PubMed] [Google Scholar]
- 15. De Roock W, Claes B, Bernasconi D, et al. Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol. 2010;11(8):753–762 [DOI] [PubMed] [Google Scholar]
- 16. Laurent-Puig P, Cayre A, Manceau G, et al. Analysis of PTEN, BRAF, and EGFR status in determining benefit from cetuximab therapy in wild-type KRAS metastatic colon cancer. J Clin Oncol. 2009;27(35):5924–5930 [DOI] [PubMed] [Google Scholar]
- 17. Alberts SR, Sargent DJ, Nair S, et al. Effect of oxaliplatin, fluorouracil, and leucovorin with or without cetuximab on survival among patients with resected stage III colon cancer: a randomized trial. JAMA. 2012;307(13):1383–1393 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. McKay J, Murray G, McLeod H. Clinician’s guide to immunohistochemistry. UroOncology. 2000;1(1):11–16 [Google Scholar]
- 19. Altman D. Practical Statistics for Medical Research. London: Chapman & Hall; 1991 [Google Scholar]
- 20. Fleiss J. Statistical Methods for Rates and Proportions. 2nd ed. New York: John Wiley & Sons; 1981 [Google Scholar]
- 21. Sinicrope FA, Mahoney MR, Smyrk TC, et al. Prognostic impact of deficient DNA mismatch repair in patients with stage III colon cancer from a randomized trial of FOLFOX-based adjuvant chemotherapy. J Clin Oncol. 2013;31(29):3664–3672 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Botteri E, Iodice S, Bagnardi V, et al. Smoking and colorectal cancer: a meta-analysis. JAMA. 2008;300(23):2765–2778 [DOI] [PubMed] [Google Scholar]
- 23. Harris CC. Chemical and physical carcinogenesis: advances and perspectives for the 1990s. Cancer Res. 1991;51(18 Suppl):5023s–5044s [PubMed] [Google Scholar]
- 24. Porta M, Crous-Bou M, Wark PA, et al. Cigarette smoking and K-ras mutations in pancreas, lung and colorectal adenocarcinomas: etiopathogenic similarities, differences and paradoxes. Mutat Res. 2009;682(2–3):83–93 [DOI] [PubMed] [Google Scholar]
- 25. Samadder NJ, Vierkant RA, Tillmans LS, et al. Cigarette smoking and colorectal cancer risk by KRAS mutation status among older women. Am J Gastroenterol. 2012;107(5):782–789 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Slattery ML, Anderson K, Curtin K, et al. Lifestyle factors and Ki-ras mutations in colon cancer tumors. Mutat Res. 2001;483(1–2):73–81 [DOI] [PubMed] [Google Scholar]
- 27. Weijenberg MP, Aardening PWM, de Kok TM, et al. Cigarette smoking and KRAS oncogene mutations in sporadic colorectal cancer: results from the Netherlands Cohort Study. Mutat Res. 2008;652(1):54–64 [DOI] [PubMed] [Google Scholar]
- 28. Samowitz WS, Albertsen H, Sweeney C, et al. Association of smoking, CpG island methylator phenotype, and V600E BRAF mutations in colon cancer. J Natl Cancer Inst. 2006;98(23):1731–1738 [DOI] [PubMed] [Google Scholar]
- 29. Limsui D, Vierkant RA, Tillmans LS, et al. Cigarette smoking and colorectal cancer risk by molecularly defined subtypes. J Natl Cancer Inst. 2010;102(14):1012–1022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Hammons GJ, Yan Y, Lopatina NG, et al. Increased expression of hepatic DNA methyltransferase in smokers. Cell Biol Toxicol. 1999;15(6):389–394 [DOI] [PubMed] [Google Scholar]
- 31. Samowitz WS, Albertsen H, Herrick J, et al. Evaluation of a large, population-based sample supports a CpG island methylator phenotype in colon cancer. Gastroenterology. 2005;129(3):837–845 [DOI] [PubMed] [Google Scholar]
- 32. Toyota M, Ahuja N, Ohe-Toyota M, et al. CpG island methylator phenotype in colorectal cancer. Proc Natl Acad Sci U S A. 1999;96(15):8681–8686 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Bird AP. CpG-rich islands and the function of DNA methylation. Nature. 1986;321(6067):209–213 [DOI] [PubMed] [Google Scholar]
- 34. Jones PA, Laird PW. Cancer epigenetics comes of age. Nat Genet. 1999;21(2):163–167 [DOI] [PubMed] [Google Scholar]
- 35. Aaltonen LA, Salovaara R, Kristo P, et al. Incidence of hereditary nonpolyposis colorectal cancer and the feasibility of molecular screening for the disease. N Engl J Med. 1998;338(21):1481–1487 [DOI] [PubMed] [Google Scholar]
- 36. Peltomaki P. Deficient DNA mismatch repair: a common etiologic factor for colon cancer. Hum Mol Genet. 2001;10(7):735–740 [DOI] [PubMed] [Google Scholar]
- 37. Cunningham JM, Christensen ER, Tester DJ, et al. Hypermethylation of the hMLH1 promoter in colon cancer with microsatellite instability. Cancer Res. 1998;58(15):3455–3460 [PubMed] [Google Scholar]
- 38. Cunningham JM, Kim CY, Christensen ER, et al. The frequency of hereditary defective mismatch repair in a prospective series of unselected colorectal carcinomas. Am J Hum Genet. 2001;69(4):780–790 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Hinoue T, Weisenberger DJ, Lange CP, et al. Genome-scale analysis of aberrant DNA methylation in colorectal cancer. Genome Res. 2012;22(2):271–282 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Aaltonen LA, Peltomaki P, Leach FS, et al. Clues to the pathogenesis of familial colorectal cancer. Science. 1993;260(5109):812–816 [DOI] [PubMed] [Google Scholar]
- 41. Oliveira C, Westra JL, Arango D, et al. Distinct patterns of KRAS mutations in colorectal carcinomas according to germline mismatch repair defects and hMLH1 methylation status. Hum Mol Genet. 2004;13(19):2303–2311 [DOI] [PubMed] [Google Scholar]
- 42. Leggett B, Whitehall V. Role of the serrated pathway in colorectal cancer pathogenesis. Gastroenterology. 2010;138(6):2088–2100 [DOI] [PubMed] [Google Scholar]