Purpose: Somatic mutation status at KRAS, BRAF, and NRAS is associated with prognosis in patients with advanced colorectal cancer (aCRC); however, it remains unclear whether there are intralocus, variant-specific differences in survival and other clinicopathologic parameters.

Experimental Design: We profiled 2,157 aCRCs for somatic mutations in KRAS, BRAF, and NRAS and determined microsatellite instability status. We sought inter- and intralocus correlations between mutations and variant-specific associations with survival and clinicopathology.

Results: KRAS mutations were rarely found together and those in codons 12 and 13 conferred poor prognosis [hazard ratio (HR), 1.44; 95% confidence interval (CI), 1.28–1.61; P = 6.4 × 10−10 and HR, 1.53; 95% CI, 1.26–1.86; P = 1.5 × 10−05, respectively]. For BRAF, more c.1781A>G (p.D594G) CRCs carried RAS mutations [14% (3/21)] compared with c.1799T>A (p.V600E) CRCs [1% (2/178), P = 9.0 × 10−03]. c.1799T>A (p.V600E) was associated with poor prognosis (HR, 2.60; 95% CI, 2.06–3.28; P = 1.0 × 10−15), whereas c.1781A>G (p.D594G) was not (HR, 1.30; 95% CI, 0.73–2.31; P = 0.37); this intralocus difference was significant (P = 0.04). More c.1799T>A (p.V600E) colorectal cancers were found in the right colon [47% (47/100)], compared with c.1781A>G (p.D594G) colorectal cancers [7% (1/15), P = 3.7 × 10−03]. For NRAS, 5% (3/60) of codon 61 mutant colorectal cancers had KRAS mutations compared with 44% (10/23) of codons 12 and 13 mutant colorectal cancers (P = 7.9 × 10−05). Codon 61 mutations conferred poor prognosis (HR, 1.47; 95% CI, 1.09–1.99; P = 0.01), whereas codons 12 and 13 mutations did not (HR, 1.29; 95% CI, 0.64–2.58; P = 0.48).

Conclusions: Our data show considerable intralocus variation in the outcomes of mutations in BRAF and NRAS. These data need to be considered in patient management and personalized cancer therapy. Clin Cancer Res; 23(11); 2742–9. ©2016 AACR.

Translational Relevance

Somatic mutation status at KRAS, BRAF, and NRAS affects prognosis in patients with advanced colorectal cancer (aCRC), and it has been presumed that different variants in the same gene confer similar prognostic outcomes. Here, we studied inter- and intralocus variant cooccurrence and variant-specific differences in survival and clinicopathology by analyzing 2,157 patients with aCRC. We found significant differences between variants in BRAF [c.1781A>G (p.D594G) versus c.1799T>A (p.V600E)] and NRAS (mutant codons 12 and 13 vs. codon 61) both in terms of cooccurrence with KRAS mutations and in their influence on survival. These data need to be considered in patient management and personalized therapy.

The only routinely used prognostic marker for survival after diagnosis of colorectal cancer is clinical stage, which combines depth of tumor invasion, nodal status, and distant metastasis (1). In stage IV disease, Köhne's index based on performance status, white blood cell count, alkaline phosphatase levels, and number of metastatic sites has been proposed (2). Other factors thought to influence survival include lifestyle (3, 4), systemic inflammatory response to the tumor (5), tumor immunologic environment (6), and the germline (7) and somatic (8–11) molecular profiles. By studying patients with advanced colorectal cancer (aCRC) from the Medical Research Council (MRC) COIN trial, we previously showed that the somatic mutation status at KRAS and BRAF, and microsatellite instability (MSI), conferred poor prognosis irrespective of treatment: overall survival (OS, trial enrolment to death) KRAS-mutant 14.4 months (12), BRAF-mutant 8.8 months (12), MSI 9.3 months (13), all wild-type 20.1 months (12). We also showed that neither individual somatic mutations nor mutations grouped by codon or gene affected response to cetuximab (13).

It remains unclear whether there are intralocus, variant-specific differences in survival, and this has been difficult to study for the less frequently mutated loci [such as c.1781A>G (p.D594G) in BRAF] due to the large numbers of samples required to make statistically robust associations. Here, we studied the influence of individual or codon-specific somatic mutations in KRAS, BRAF, and NRAS in 2,157 patients with aCRC from COIN (12) and COIN-B (14).

Patients and samples

We prepared tumor DNA samples from unrelated patients with aCRC from the MRC clinical trials COIN (NCT00182715; ref. 12) and COIN-B (NCT00640081; ref. 14), as described previously (12, 13). All patients had either previous or current histologically confirmed primary adenocarcinomas of the colon or rectum, together with clinical or radiological evidence of advanced and/or metastatic disease, or had histologically/cytologically confirmed metastatic adenocarcinomas, together with clinical and/or radiological evidence of a colorectal primary tumor. COIN patients were randomized 1:1:1 to receive continuous oxaliplatin and fluoropyrimidine chemotherapy, continuous chemotherapy plus cetuximab, or intermittent chemotherapy. COIN-B patients were randomized 1:1 to receive intermittent chemotherapy plus continuous cetuximab or intermittent chemotherapy plus intermittent cetuximab. All patients gave informed consent for their samples to be used for bowel cancer research [approved by REC (04/MRE06/60)].

Somatic analyses

We previously screened for somatic mutations in KRAS (codons 12, 13, and 61), BRAF (codons 594 and 600), and NRAS (codons 12, 13, and 61) using a combination of pyrosequencing and Sequenom (13); for samples analyzed by both technologies (n = 1,612), genotype concordance in KRAS was 99% (8,642/8,719 calls were concordant). MSI status was determined using the markers BAT-25 and BAT-26 (13).

Mutation cooccurrence, survival, and statistical analyses

We sought inter- and intralocus correlations between somatic KRAS, BRAF, and NRAS mutations and MSI status. Data were analyzed using R (http://www.r-project.org). Corrplot was used to create a correlation matrix plot (recode from car was used to recode the data into binary format) and Surv, survfit, survdiff, and coxph from the survival package and ggsurv from the GGally package were used to create and analyze the survival curves. To avoid potential confounding effects from other mutant loci, KRAS mutants (vs. wild-type) were analyzed on a BRAF and NRAS wild-type background; BRAF mutants (vs. wild-type) were analyzed on a KRAS and NRAS (RAS) wild-type and microsatellite stable (MSS) background; NRAS mutants (vs. wild-type) were analyzed on a KRAS and BRAF wild-type background; and MSI (vs. MSS) was analyzed on a RAS and BRAF wild-type background. We found no evidence of heterogeneity in OS between patients when analyzed by trial (COIN vs. COIN-B, P = 0.49), trial arm (P = 0.40; Cochran Q test: P = 1.0, I2 test: P = 0.74), type of chemotherapy received (OxMdG/XELOX; P = 0.60), or cetuximab use (P = 0.41), so we combined these groups for the survival analyses. We used χ2 tests or Fisher exact test to study whether KRAS, BRAF, and NRAS mutations and MSI status were associated with different clinicopathologic findings. We corrected for multiple testing using Bonferroni correction [P < 1.7 × 10−03 (n = 30) for survival tests, P < 1.0 × 10−04 (n = 480) for somatic mutation cross-correlations, P < 1.3 × 10−03 (n = 39) for clinicopathologic analyses of KRAS, BRAF, and NRAS and P < 3.8 × 10−03 (n = 13) for clinicopathologic analyses of MSI].

We screened for somatic KRAS, BRAF, and NRAS mutations and for MSI status in aCRCs from 2,157 patients from the clinical trials COIN and COIN-B. In total, we detected 14 KRAS mutations [c.34G>A (p.G12S), c.34G>C (p.G12R), c.34G>T (p.G12C), c.35G>A (p.G12D), c.35G>C (p.G12A), c.35G>T (p.G12V), c.37G>A (p.G13S), c.37G>C (p.G13R), c.37G>T (p.G13C), c.38G>A (p.G13D), c.38G>T (p.G13V), c.182A>G (p.Q61R), c.182A>T (p.Q61L), and c.183A>C (p.Q61H)] in 40% (858/2,157) aCRCs, 2 BRAF mutations [c.1781A>G (p.D594G) and c.1799T>A (p.V600E)] in 9% (199/2,097), 9 NRAS mutations [c.34G>T (p.G12C), c.35G>A (p.G12D), c.35G>T (p.G12V), c.37G>C (p.G13R), c.38G>A (p.G13D), c.181C>A (p.Q61K), c.182A>G (p.Q61R), c.182A>T (p.Q61L), and c.183A>C (p.Q61H)] in 4% (83/2,092), and MSI in 4% (66/1,567). Over 99% (2,152/2,157) of aCRCs harboring KRAS, BRAF, and NRAS mutations carried only a single variant allele at their respective loci (five colorectal cancers carried two KRAS mutations; however, due to their rarity, these were likely to reflect mixed tumor populations).

Inter- and intragenic mutation correlations

All mutations in KRAS, regardless of whether analyzed individually or by codon, showed similar effects in terms of mutual exclusivity (Supplementary Fig. S1). Codon 12 (4/627 mutant colorectal cancers), 13 (4/161), and 61 (2/35) mutations were rarely found together.

Only specific mutations in BRAF [c.1799T>A (p.V600E)] and NRAS (codon 61 mutations) shared this characteristic. Only 1% (2/178) of BRAF c.1799T>A (p.V600E) colorectal cancers had RAS mutations compared with 47% (894/1,908) of BRAF wild-type colorectal cancers (P < 2.2 × 10−16, P < 1.1 × 10−13 after correction for multiple testing). In contrast, more BRAF c.1781A>G (p.D594G) mutations cooccurred with RAS mutations [14% (3/21)] as compared with c.1799T>A (p.V600E; P = 9.0 × 10−03), albeit less commonly than found in BRAF wild-type colorectal cancers (P = 3.0 × 10−03). We noted one case of KRAS c.37G>A (p.G13S), which cooccurred with BRAF c.1799T>A [p.V600E; P = 2.5 × 10−03 as compared with other KRAS mutations (1/812 cooccurred)]. For NRAS, only 5% (3/60) of codon 61 mutant colorectal cancers had KRAS mutations compared with 43% (10/23) of codons 12 and 13 mutant colorectal cancers (P = 7.9 × 10−05, P = 0.04 after correction), the latter being at a similar level to that found in wild-type colorectal cancers [40% (808/2,018), P = 0.98].

We also observed differences in the relationship between BRAF mutations and MSI status. BRAF c.1799T>A (p.V600E) was strongly associated with MSI [11% (20/178) c.1799T>A (p.V600E) colorectal cancers had MSI compared with 2% (46/1,908) wild-type colorectal cancers, P = 5.3 × 10−10, P = 2.5 × 10−07 after correction], whereas BRAF c.1781A>G (p.D594G) and MSI did not cooccur (0/21).

Survival analyses

Five KRAS mutations [c.34G>A (p.G12S), c.35G>A (p.G12D), c.35G>C (p.G12A), c.35G>T (p.G12V), and c.38G>A (p.G13D)] individually showed significantly poorer prognosis with a median reduction in survival of 213, 111, 65, 160, and 165 days, respectively; four of these remained significant after correction for multiple testing (Table 1). When grouped by codons, both codon 12 and 13 mutations conferred poor prognosis [hazard ratio (HR), 1.44; 95% confidence interval (CI), 1.28–1.61; P = 6.4 × 10−10, P = 1.9 × 10−08 after correction, and HR, 1.53; 95% CI, 1.26–1.86; P = 1.5 × 10−05, P = 4.5 × 10−04 after correction, respectively], whereas codon 61 mutations did not (HR, 1.23; 95% CI, 0.84–1.81; P = 0.28; Table 1); these intralocus differences were not significant.

Table 1.

Prognostic outcomes of individual mutations, or mutations grouped by codon or gene on OS

Gene/eventMutation/codonNo of eventsaHR95% CIsP
KRAS c.34G>A (p.G12S) 35 1.78 1.27–2.50 9.2 × 10−04 (0.03) 
 c.34G>C (p.G12R) 10 0.95 0.51–1.78 0.88 
 c.34G>T (p.G12C) 52 1.21 0.91–1.60 0.18 
 c.35G>A (p.G12D) 187 1.48 1.26–1.74 1.6 × 10−06 (4.8 × 10−05) 
 c.35G>C (p.G12A) 41 1.43 1.04–1.96 0.03 
 c.35G>T (p.G12V) 161 1.48 1.25–1.76 7.5 × 10−06 (2.3 × 10−04) 
 c.37G>T (p.G13C) 6 1.36 0.61–3.03 0.46 
 c.38G>A (p.G13D) 116 1.53 1.26–1.87 2.2 × 10−05 (6.6 × 10−04) 
 c.38G>T (p.G13V) 1 — — — 
 c.182A>G (p.Q61R) 6 1.41 0.63–3.15 0.41 
 c.182A>T (p.Q61L) 6 1.27 0.57–2.84 0.56 
 c.183A>C (p.Q61H) 15 1.17 0.70–1.95 0.56 
 Codon 12 486 1.44 1.28–1.61 6.4 × 10−10 (1.9 × 10−08) 
 Codon 13 123 1.53 1.26–1.86 1.5 × 10−05 (4.5 × 10−04) 
 Codon 61 27 1.23 0.84–1.81 0.28 
 Any KRAS mutation 632 1.45 1.30–1.61 1.9 × 10−11 (5.7 × 10−10) 
BRAF c.1781A>G (p.D594G) 12 1.30 0.73–2.31 0.37 
 c.1799T>A (p.V600E) 87 2.60 2.06–3.28 1.0 × 10−15 (3.0 × 10−14) 
 Any BRAF mutation 99 2.31 1.85–2.87 7.8 × 10−14 (2.3 × 10−12) 
NRAS c.34G>T (p.G12C) 5 1.42 0.59–3.43 0.43 
 c.35G>A (p.G12D) 2 — — — 
 c.35G>T (p.G12V) 1 — — — 
 c.181C>A (p.Q61K) 21 1.43 0.96–2.21 0.11 
 c.182A>G (p.Q61R) 13 1.58 0.91–2.73 0.11 
 c.182A>T (p.Q61L) 11 1.51 0.83–2.73 0.18 
 Codons 12 and 13 8 1.29 0.64–2.58 0.48 
 Codon 61 45 1.47 1.09–1.99 0.01 
 Any NRAS mutation 53 1.44 1.09–1.90 0.01 
MSI MSI 23 1.86 1.22–2.83 4.0 × 10−03 
 MSS 476 1.00 Ref. Ref. 
Gene/eventMutation/codonNo of eventsaHR95% CIsP
KRAS c.34G>A (p.G12S) 35 1.78 1.27–2.50 9.2 × 10−04 (0.03) 
 c.34G>C (p.G12R) 10 0.95 0.51–1.78 0.88 
 c.34G>T (p.G12C) 52 1.21 0.91–1.60 0.18 
 c.35G>A (p.G12D) 187 1.48 1.26–1.74 1.6 × 10−06 (4.8 × 10−05) 
 c.35G>C (p.G12A) 41 1.43 1.04–1.96 0.03 
 c.35G>T (p.G12V) 161 1.48 1.25–1.76 7.5 × 10−06 (2.3 × 10−04) 
 c.37G>T (p.G13C) 6 1.36 0.61–3.03 0.46 
 c.38G>A (p.G13D) 116 1.53 1.26–1.87 2.2 × 10−05 (6.6 × 10−04) 
 c.38G>T (p.G13V) 1 — — — 
 c.182A>G (p.Q61R) 6 1.41 0.63–3.15 0.41 
 c.182A>T (p.Q61L) 6 1.27 0.57–2.84 0.56 
 c.183A>C (p.Q61H) 15 1.17 0.70–1.95 0.56 
 Codon 12 486 1.44 1.28–1.61 6.4 × 10−10 (1.9 × 10−08) 
 Codon 13 123 1.53 1.26–1.86 1.5 × 10−05 (4.5 × 10−04) 
 Codon 61 27 1.23 0.84–1.81 0.28 
 Any KRAS mutation 632 1.45 1.30–1.61 1.9 × 10−11 (5.7 × 10−10) 
BRAF c.1781A>G (p.D594G) 12 1.30 0.73–2.31 0.37 
 c.1799T>A (p.V600E) 87 2.60 2.06–3.28 1.0 × 10−15 (3.0 × 10−14) 
 Any BRAF mutation 99 2.31 1.85–2.87 7.8 × 10−14 (2.3 × 10−12) 
NRAS c.34G>T (p.G12C) 5 1.42 0.59–3.43 0.43 
 c.35G>A (p.G12D) 2 — — — 
 c.35G>T (p.G12V) 1 — — — 
 c.181C>A (p.Q61K) 21 1.43 0.96–2.21 0.11 
 c.182A>G (p.Q61R) 13 1.58 0.91–2.73 0.11 
 c.182A>T (p.Q61L) 11 1.51 0.83–2.73 0.18 
 Codons 12 and 13 8 1.29 0.64–2.58 0.48 
 Codon 61 45 1.47 1.09–1.99 0.01 
 Any NRAS mutation 53 1.44 1.09–1.90 0.01 
MSI MSI 23 1.86 1.22–2.83 4.0 × 10−03 
 MSS 476 1.00 Ref. Ref. 

NOTE: KRAS mutants (vs. wild-type) were analyzed on a BRAF and NRAS wild-type background; BRAF mutants (vs. wild-type) were analyzed on a RAS wild-type and MSS background; NRAS mutants (vs. wild-type) were analyzed on a KRAS and BRAF wild-type background; and MSI (vs. MSS) was analyzed on a RAS and BRAF wild-type background.

Number of events, HR, CIs, and P values are shown (except for cases where number of events ≤2).

aMutations not listed when number of events = 0.

P values that remained significant after correction for multiple testing are shown in parentheses.

View Large

c.1799T>A (p.V600E) in BRAF was strongly associated with poor prognosis (HR, 2.60; 95% CI, 2.06–3.28; P = 1.0 × 10−15, P = 3.0 × 10−14 after correction, median reduction in survival 320 days; Fig. 1), whereas c.1781A>G (p.D594G) was not (HR, 1.30; 95% CI, 0.73–2.31; P = 0.37); this intralocus difference was significant (P = 0.04; Table 1).

Figure 1.
Figure 1. Kaplan–Meier plots showing the prognostic outcome of c.1781A>G (p.D594G) and c.1799T>A (p.V600E) in BRAF (A), and codons 12 and 13 and codon 61 mutations in NRAS (B).
View largeDownload slide

Kaplan–Meier plots showing the prognostic outcome of c.1781A>G (p.D594G) and c.1799T>A (p.V600E) in BRAF (A), and codons 12 and 13 and codon 61 mutations in NRAS (B).

Figure 1.
Figure 1. Kaplan–Meier plots showing the prognostic outcome of c.1781A>G (p.D594G) and c.1799T>A (p.V600E) in BRAF (A), and codons 12 and 13 and codon 61 mutations in NRAS (B).
View largeDownload slide

Kaplan–Meier plots showing the prognostic outcome of c.1781A>G (p.D594G) and c.1799T>A (p.V600E) in BRAF (A), and codons 12 and 13 and codon 61 mutations in NRAS (B).

Close modal

Although individual NRAS mutations showed no differences in survival, when grouped by codon, codon 61 mutations conferred poor prognosis (HR, 1.47; 95% CI, 1.09–1.99; P = 0.01, median reduction in survival 131 days; Fig. 1), whereas codons 12 and 13 mutations did not (HR, 1.29; 95% CI, 0.64–2.58; P = 0.48); however, this intralocus difference was not significant (P = 0.73).

Patients with MSI colorectal cancers had worse prognosis compared with those with stable tumors (HR, 1.86; 95% CI, 1.22–2.83; P = 4.0 × 10−03), in agreement with our previous study (13).

For all analyses described herein, there were no significant differences measured using heterogeneity tests when the analyses were performed using date of diagnosis to death instead of OS (Supplementary Table S1) or when split by cetuximab use (Supplementary Table S2).

Clinicopathologic analyses

KRAS.

More KRAS-mutant colorectal cancers were found in the right colon [58% (182/314)] and cecum [70% (62/88)] as compared with the left colon [38% (123/326), P = 4.6 × 10−07 and 8.4 × 10−08, respectively], and more were associated with metastases in the lung [50% (358/715)] as compared with the liver only [37% (156/418), P = 4.2 × 10−05; Table 2].

Table 2.

Clinicopathology according to KRAS mutation status

CharacteristicsFrequency of KRAS mutationsaCodons 12 and 13 (n = 760)Codon 61(n = 33)Wild-type(n = 1,002)P (codons 12 and 13 vs. wild-type)P (codon 61 vs.wild-type)P (codons 12 and 13 vs. codon 61)
Sex Female 289/593 (49) 275 (36) 14 (42) 304 (30) 0.01 0.20 0.59 
 Male 503/1,201 (42) 485 (64) 19 (58) 698 (70) 0.01 0.20 0.59 
Age Mean NA 63 61 63 NA NA NA 
Primary tumor site Right colon 182/314 (58) 173 (23) 9 (27) 132 (13) 1.9 × 10−07 [7.4 × 10−06] 0.04 0.70 
 Cecum 62/88 (70) 61 (8) 1 (3) 26 (3) 3.4 × 10−07 [1.3 × 10−05] 0.59 0.51 
 Transverse colon 14/35 (40) 14 (2) 0 (0) 21 (2) 0.84 1.0 1.0 
 Left colon 123/326 (38) 117 (15) 6 (18) 203 (20) 0.01 0.94 0.85 
 Sigmoid colon 44/159 (28) 41 (5) 3 (9) 115 (11) 1.3 × 10−05 [5.1 × 10−04] 1.0 0.42 
 Rectosigmoid junction 108/269 (40) 105 (14) 3 (9) 161 (16) 0.22 0.34 0.61 
 Rectum 251/577 (44) 241 (32) 10 (30) 326 (33) 0.75 0.94 1.0 
Site of metastasesb Liver only 156/418 (37) 152 (20) 4 (12) 262 (26) 3.1 × 10−03 0.07 0.37 
 Liver 598/1,356 (44) 577 (76) 22 (67) 758 (76) 0.94 0.33 0.32 
 Nodal 359/832 (43) 345 (45) 15 (45) 473 (47) 0.48 0.98 1.0 
 Lung 358/715 (50) 342 (45) 16 (48) 357 (36) 8.4 × 10−05 [3.3 × 10−03] 0.18 0.83 
 Peritoneum 126/259 (49) 117 (15) 9 (27) 133 (13) 0.23 0.04 0.11 
CharacteristicsFrequency of KRAS mutationsaCodons 12 and 13 (n = 760)Codon 61(n = 33)Wild-type(n = 1,002)P (codons 12 and 13 vs. wild-type)P (codon 61 vs.wild-type)P (codons 12 and 13 vs. codon 61)
Sex Female 289/593 (49) 275 (36) 14 (42) 304 (30) 0.01 0.20 0.59 
 Male 503/1,201 (42) 485 (64) 19 (58) 698 (70) 0.01 0.20 0.59 
Age Mean NA 63 61 63 NA NA NA 
Primary tumor site Right colon 182/314 (58) 173 (23) 9 (27) 132 (13) 1.9 × 10−07 [7.4 × 10−06] 0.04 0.70 
 Cecum 62/88 (70) 61 (8) 1 (3) 26 (3) 3.4 × 10−07 [1.3 × 10−05] 0.59 0.51 
 Transverse colon 14/35 (40) 14 (2) 0 (0) 21 (2) 0.84 1.0 1.0 
 Left colon 123/326 (38) 117 (15) 6 (18) 203 (20) 0.01 0.94 0.85 
 Sigmoid colon 44/159 (28) 41 (5) 3 (9) 115 (11) 1.3 × 10−05 [5.1 × 10−04] 1.0 0.42 
 Rectosigmoid junction 108/269 (40) 105 (14) 3 (9) 161 (16) 0.22 0.34 0.61 
 Rectum 251/577 (44) 241 (32) 10 (30) 326 (33) 0.75 0.94 1.0 
Site of metastasesb Liver only 156/418 (37) 152 (20) 4 (12) 262 (26) 3.1 × 10−03 0.07 0.37 
 Liver 598/1,356 (44) 577 (76) 22 (67) 758 (76) 0.94 0.33 0.32 
 Nodal 359/832 (43) 345 (45) 15 (45) 473 (47) 0.48 0.98 1.0 
 Lung 358/715 (50) 342 (45) 16 (48) 357 (36) 8.4 × 10−05 [3.3 × 10−03] 0.18 0.83 
 Peritoneum 126/259 (49) 117 (15) 9 (27) 133 (13) 0.23 0.04 0.11 

NOTE: Mutations were analyzed on an NRAS and BRAF wild-type background.

Abbreviation: NA, not applicable.

aThere was a significant difference between KRAS-mutant colorectal cancers in the location of the primary tumor (P = 6.4 × 10−14) and in the sites of metastases (P = 4.6 × 10−04) as compared with wild-type colorectal cancers.

bSome patients had multiple metastases, so percentages do not add up to 100%.

Percentages are shown in regular parentheses.

P values that remained significant after correction for multiple testing are shown in square parentheses.

Discrepancies in column totals are due to patients with multiple mutations or due to missing data.

View Large

In terms of codon-specific mutations, more KRAS codon 12 and 13 mutant colorectal cancers were found in the right colon [23% (173/760) vs. 13% (132/1,002), P = 1.9 × 10−07] and cecum [8% (61/760) vs. 3% (26/1,002), P = 3.4 × 10−07], less in the left colon [15% (117/760) vs. 20% (203/1,002), P = 0.01] and sigmoid colon [5% (41/760) vs. 11% (115/1,002), P = 1.3 × 10−05], and more were associated with metastases in the lung [45% (342/760) vs. 36% (357/1,002), P = 8.4 × 10−05] and less in liver only [20% (152/760) vs. 26% (262/1,002), P = 3.1 × 10−03], as compared with wild-type colorectal cancers; the correlations for right colon, cecum, sigmoid colon, and lung remained significant after correction for multiple testing (Table 2). More KRAS codon 61 mutant patients had colorectal cancers in the right colon [27% (9/33) vs. 13% (132/1,002), P = 0.04] and more had peritoneal metastases [27% (9/33) vs. 13% (133/1,002), P = 0.04] as compared with wild-type patients. However, there were no significant differences in clinicopathology between KRAS codons 12 and 13 versus codon 61 mutant patients.

BRAF.

More BRAF-mutant colorectal cancers were found in the right colon [38% (48/128)] as compared with the left colon [12% (18/146), P = 2.4 × 10−05], and more were associated with metastases in the peritoneum [23% (25/107)] as compared with liver only [10% (22/214), P = 3.1 × 10−03; Table 3].

Table 3.

Clinicopathology according to BRAF mutation status

CharacteristicsFrequency of BRAF mutationsac.1781A>G (p.D594G; n = 15)c.1799T>A (p.V600E; n = 100)Wild-type (n = 693)P (c.1781A>G [p.D594G] vs. wild-type)p (c.1799T>A [p.V600E] vs. wild-type)p (c.1781A>G [p.D594G] vs. c.1799T>A [p.V600E))
Sex Female 55/249 (22) 7 (47) 48 (48) 194 (28) 0.20 8.0 × 10−05 [3.1 × 10−03] 1.0 
 Male 60/559 (11) 8 (53) 52 (52) 499 (72) 0.20 8.0 × 10−05 [3.1 × 10−03] 1.0 
Age Mean NA 67 63 63 NA NA NA 
Primary tumor site Right colon 48/128 (38) 1 (7) 47 (47) 80 (12) 1.0 <2.2 × 10−16 [<8.6 × 10−15] 3.7 × 10−03 
 Cecum 4/24 (17) 0 (0) 4 (4) 20 (3) 1.0 0.53 1.0 
 Transverse colon 4/20 (20) 0 (0) 4 (4) 16 (2) 1.0 0.30 1.0 
 Left colon 18/146 (12) 1 (7) 17 (17) 128 (18) 0.34 0.83 0.46 
 Sigmoid colon 4/77 (5) 2 (13) 2 (2) 73 (11) 0.67 3.2 × 10−03 0.08 
 Rectosigmoid junction 15/141 (11) 2 (13) 13 (13) 126 (18) 1.0 0.26 1.0 
 Rectum 20/254 (8) 9 (60) 11 (11) 234 (34) 0.07 7.1 × 10−06 [2.8 × 10−04] 1.7 × 10−05 [6.6 × 10−04] 
Site of metastasesb Liver only 22/214 (10) 3 (20) 19 (19) 192 (28) 0.77 0.09 1.0 
 Liver 83/619 (13) 13 (87) 70 (70) 536 (77) 0.54 0.14 0.23 
 Nodal 53/368 (14) 7 (47) 46 (46) 315 (45) 1.0 1.0 1.0 
 Lung 35/272 (13) 6 (40) 29 (29) 237 (34) 0.85 0.36 0.57 
 Peritoneum 25/107 (23) 1 (7) 24 (24) 82 (12) 1.0 1.5 × 10−03 0.19 
CharacteristicsFrequency of BRAF mutationsac.1781A>G (p.D594G; n = 15)c.1799T>A (p.V600E; n = 100)Wild-type (n = 693)P (c.1781A>G [p.D594G] vs. wild-type)p (c.1799T>A [p.V600E] vs. wild-type)p (c.1781A>G [p.D594G] vs. c.1799T>A [p.V600E))
Sex Female 55/249 (22) 7 (47) 48 (48) 194 (28) 0.20 8.0 × 10−05 [3.1 × 10−03] 1.0 
 Male 60/559 (11) 8 (53) 52 (52) 499 (72) 0.20 8.0 × 10−05 [3.1 × 10−03] 1.0 
Age Mean NA 67 63 63 NA NA NA 
Primary tumor site Right colon 48/128 (38) 1 (7) 47 (47) 80 (12) 1.0 <2.2 × 10−16 [<8.6 × 10−15] 3.7 × 10−03 
 Cecum 4/24 (17) 0 (0) 4 (4) 20 (3) 1.0 0.53 1.0 
 Transverse colon 4/20 (20) 0 (0) 4 (4) 16 (2) 1.0 0.30 1.0 
 Left colon 18/146 (12) 1 (7) 17 (17) 128 (18) 0.34 0.83 0.46 
 Sigmoid colon 4/77 (5) 2 (13) 2 (2) 73 (11) 0.67 3.2 × 10−03 0.08 
 Rectosigmoid junction 15/141 (11) 2 (13) 13 (13) 126 (18) 1.0 0.26 1.0 
 Rectum 20/254 (8) 9 (60) 11 (11) 234 (34) 0.07 7.1 × 10−06 [2.8 × 10−04] 1.7 × 10−05 [6.6 × 10−04] 
Site of metastasesb Liver only 22/214 (10) 3 (20) 19 (19) 192 (28) 0.77 0.09 1.0 
 Liver 83/619 (13) 13 (87) 70 (70) 536 (77) 0.54 0.14 0.23 
 Nodal 53/368 (14) 7 (47) 46 (46) 315 (45) 1.0 1.0 1.0 
 Lung 35/272 (13) 6 (40) 29 (29) 237 (34) 0.85 0.36 0.57 
 Peritoneum 25/107 (23) 1 (7) 24 (24) 82 (12) 1.0 1.5 × 10−03 0.19 

NOTE: Mutations analyzed on a RAS wild-type and MSS background.

Abbreviation: NA, not applicable.

aThere was a significant difference between BRAF-mutant colorectal cancers in the location of the primary tumor (P = 1.2 × 10−13) and in the sites of metastases (P = 0.03) as compared with wild-type colorectal cancers.

bSome patients had multiple metastases so percentages do not add up to 100%.

Percentages are shown in regular parentheses.

P values that remained significant after correction for multiple testing are shown in square parentheses.

Discrepancies in column totals are due to patients with multiple mutations or due to missing data.

View Large

In terms of individual mutations, BRAF c.1781A>G (p.D594G) colorectal cancers had similar clinicopathology to wild-type colorectal cancers (Table 3). In contrast, more BRAF c.1799T>A (p.V600E) colorectal cancers were found in the right colon [47% (47/100) vs. 12% (80/693), P < 2.2 × 10−16], and less in the rectum [11% (11/100) vs. 34% (234/693), P = 7.1 × 10−06] and sigmoid colon [2% (2/100) vs. 11% (73/693), P = 3.2 × 10−03], and more were associated with peritoneal metastases [24% (24/100) vs. 12% (82/693), P = 1.5 × 10−03] as compared with wild-type colorectal cancers; the correlations for right colon and rectum remained significant after correction for multiple testing (Table 3).

In terms of intralocus differences, there was a significant difference between c.1781A>G (p.D594G) and c.1799T>A (p.V600E) colorectal cancers in the location of the primary tumor (P = 9.3 × 10−05, P = 3.6 × 10−03 after correction), due to fewer c.1781A>G (p.D594G) colorectal cancers in the right colon [7% (1/15) vs. 47% (47/100), P = 3.7 × 10−03] and more in the rectum [60% (9/15) vs. 11% (11/100), P = 1.7 × 10−05, P = 6.6 × 10−04 after correction; Table 3]. There was no significant difference between the sites of metastases associated with these mutations.

NRAS.

There was no difference between the frequency of NRAS-mutant and wild-type colorectal cancers in the site of the primary tumor (Table 4). However, more NRAS-mutant colorectal cancers were associated with metastases in the lung [11% (43/400)] as compared with the liver only [4% (10/272), P = 1.4 × 10−03].

Table 4.

Clinicopathology according to NRAS mutation status

CharacteristicsFrequency of NRAS mutationsaCodons 12 and 13 (n = 11)Codon 61 (n = 57)Wild-type (n = 1,002)P (codons 12 and 13 vs. wild-type)P (codon 61 vs. wild-type)P (codons 12 and 13 vs. codon 61)
Sex Female 20/324 (6) 2 (18) 18 (32) 304 (30) 0.52 0.96 0.49 
 Male 48/746 (6) 9 (82) 39 (68) 698 (70) 0.52 0.96 0.49 
Age Mean NA 59 62 63 NA NA NA 
Primary tumor site Right colon 5/137 (4) 2 (18) 3 (5) 132 (13) 0.65 0.10 0.18 
 Cecum 4/30 (13) 0 (0) 4 (7) 26 (3) 1.0 0.07 1.0 
 Transverse colon 3/24 (13) 0 (0) 3 (5) 21 (2) 1.0 0.13 1.0 
 Left colon 12/215 (6) 1 (9) 11 (19) 203 (20) 0.70 1.0 0.67 
 Sigmoid colon 11/126 (9) 2 (18) 9 (16) 115 (11) 0.37 0.44 1.0 
 Rectosigmoid junction 10/171 (6) 0 (0) 10 (18) 161 (16) 0.23 0.91 0.20 
 Rectum 19/345 (6) 6 (55) 13 (23) 326 (33) 0.22 0.17 0.08 
Site of metastasesb Liver only 10/272 (4) 3 (27) 7 (12) 262 (26) 1.0 0.03 0.35 
 Liver 52/810 (6) 8 (73) 44 (77) 758 (76) 0.74 0.92 0.71 
 Nodal 35/508 (7) 3 (27) 32 (56) 473 (47) 0.23 0.24 0.11 
 Lung 43/400 (11) 4 (36) 39 (68) 357 (36) 1.0 1.3 × 10−06 [5.1 × 10−05] 0.08 
 Peritoneum 5/138 (4) 1 (9) 4 (7) 133 (13) 1.0 0.22 1.0 
CharacteristicsFrequency of NRAS mutationsaCodons 12 and 13 (n = 11)Codon 61 (n = 57)Wild-type (n = 1,002)P (codons 12 and 13 vs. wild-type)P (codon 61 vs. wild-type)P (codons 12 and 13 vs. codon 61)
Sex Female 20/324 (6) 2 (18) 18 (32) 304 (30) 0.52 0.96 0.49 
 Male 48/746 (6) 9 (82) 39 (68) 698 (70) 0.52 0.96 0.49 
Age Mean NA 59 62 63 NA NA NA 
Primary tumor site Right colon 5/137 (4) 2 (18) 3 (5) 132 (13) 0.65 0.10 0.18 
 Cecum 4/30 (13) 0 (0) 4 (7) 26 (3) 1.0 0.07 1.0 
 Transverse colon 3/24 (13) 0 (0) 3 (5) 21 (2) 1.0 0.13 1.0 
 Left colon 12/215 (6) 1 (9) 11 (19) 203 (20) 0.70 1.0 0.67 
 Sigmoid colon 11/126 (9) 2 (18) 9 (16) 115 (11) 0.37 0.44 1.0 
 Rectosigmoid junction 10/171 (6) 0 (0) 10 (18) 161 (16) 0.23 0.91 0.20 
 Rectum 19/345 (6) 6 (55) 13 (23) 326 (33) 0.22 0.17 0.08 
Site of metastasesb Liver only 10/272 (4) 3 (27) 7 (12) 262 (26) 1.0 0.03 0.35 
 Liver 52/810 (6) 8 (73) 44 (77) 758 (76) 0.74 0.92 0.71 
 Nodal 35/508 (7) 3 (27) 32 (56) 473 (47) 0.23 0.24 0.11 
 Lung 43/400 (11) 4 (36) 39 (68) 357 (36) 1.0 1.3 × 10−06 [5.1 × 10−05] 0.08 
 Peritoneum 5/138 (4) 1 (9) 4 (7) 133 (13) 1.0 0.22 1.0 

NOTE: Mutations analyzed on a KRAS and BRAF wild-type background.

Abbreviation: NA, not applicable.

aThere was a significant difference between NRAS-mutant colorectal cancers in the sites of metastases (P = 2.5 × 10−03) as compared with wild-type colorectal cancers.

bSome patients had multiple metastases so percentages do not add up to 100%.

Percentages are shown in regular parentheses.

P values that remained significant after correction for multiple testing are shown in square parentheses.

Discrepancies in column totals are due to patients with multiple mutations or due to missing data.

View Large

In terms of individual codons, codon 12 and 13 mutant colorectal cancers showed similar clinicopathology to wild-type colorectal cancers (Table 4). Codon 61 mutant colorectal cancers had similar primary tumor distributions but significantly fewer liver only [12% (7/57) vs. 26% (262/1,002), P = 0.03] and more lung metastases [68% (39/57) vs. 36% (357/1,002), P = 1.3 × 10−06, P = 5.1 × 10−05 after correction] as compared with wild-type colorectal cancers (Table 4). There were no significant differences in clinicopathology between codons 12 and 13 vs. codon 61 mutant colorectal cancers.

MSI.

More MSI colorectal cancers were found in the right colon [41% (12/29) vs. 12% (80/693), P = 9.2 × 10−06, P = 1.2 × 10−04 after correction] and less in the rectosigmoid junction [3% (1/29) vs. 18% (126/693), P = 0.04], and less were associated with liver metastases [48% (14/29) vs. 77% (536/693), P = 7.3 × 10−04, P = 9.5 × 10−03 after correction] as compared with MSS colorectal cancers (Table 5).

Table 5.

Clinicopathology according to MSI status

CharacteristicsFrequency of MSIaMSI (n = 29)MSS (n = 693)P (MSI vs. MSS)
Sex Female 11/205 (5) 11 (38) 194 (28) 0.34 
 Male 18/517 (3) 18 (62) 499 (72) 0.34 
Age Mean NA 58 63 NA 
Primary tumor site Right colon 12/92 (13) 12 (41) 80 (12) 9.2 × 10−06 [1.2 × 10−04] 
 Cecum 1/21 (5) 1 (3) 20 (3) 0.58 
 Transverse colon 1/17 (6) 1 (3) 16 (2) 0.51 
 Left colon 7/135 (5) 7 (24) 128 (18) 0.60 
 Sigmoid colon 1/76 (1) 1 (3) 73 (11) 0.35 
 Rectosigmoid junction 1/127 (1) 1 (3) 126 (18) 0.04 
 Rectum 6/240 (3) 6 (21) 234 (34) 0.21 
Site of metastasesb Liver only 3/195 (2) 3 (10) 192 (28) 0.05 
 Liver 14/550 (3) 14 (48) 536 (77) 7.3 × 10−04 [9.5 × 10−03] 
 Nodal 17/332 (5) 17 (59) 315 (45) 0.23 
 Lung 6/243 (2) 6 (21) 237 (34) 0.19 
 Peritoneum 7/89 (8) 7 (24) 82 (12) 0.09 
CharacteristicsFrequency of MSIaMSI (n = 29)MSS (n = 693)P (MSI vs. MSS)
Sex Female 11/205 (5) 11 (38) 194 (28) 0.34 
 Male 18/517 (3) 18 (62) 499 (72) 0.34 
Age Mean NA 58 63 NA 
Primary tumor site Right colon 12/92 (13) 12 (41) 80 (12) 9.2 × 10−06 [1.2 × 10−04] 
 Cecum 1/21 (5) 1 (3) 20 (3) 0.58 
 Transverse colon 1/17 (6) 1 (3) 16 (2) 0.51 
 Left colon 7/135 (5) 7 (24) 128 (18) 0.60 
 Sigmoid colon 1/76 (1) 1 (3) 73 (11) 0.35 
 Rectosigmoid junction 1/127 (1) 1 (3) 126 (18) 0.04 
 Rectum 6/240 (3) 6 (21) 234 (34) 0.21 
Site of metastasesb Liver only 3/195 (2) 3 (10) 192 (28) 0.05 
 Liver 14/550 (3) 14 (48) 536 (77) 7.3 × 10−04 [9.5 × 10−03] 
 Nodal 17/332 (5) 17 (59) 315 (45) 0.23 
 Lung 6/243 (2) 6 (21) 237 (34) 0.19 
 Peritoneum 7/89 (8) 7 (24) 82 (12) 0.09 

NOTE: MSI status was analyzed on an RAS and BRAF wild-type background.

Abbreviation: NA, not applicable.

aThere was a significant difference between MSI colorectal cancers in the location of the primary tumor (P = 2.5 × 10−04) and in the sites of metastases (P = 0.02) as compared with MSS colorectal cancers.

bSome patients had multiple metastases so percentages do not add up to 100%.

Percentages are shown in regular parentheses.

P values that remained significant after correction for multiple testing are shown in square parentheses.

Discrepancies in column totals are due to patients with multiple mutations or due to missing data.

View Large

Variants in BRAF and NRAS have been presumed to confer similar oncogenic and prognostic outcomes; however, here we demonstrate clear intralocus differences. For BRAF, c.1799T>A (p.V600E) was almost mutually exclusive of RAS mutations and was associated with poor prognosis. In contrast, c.1781A>G (p.D594G) was more often associated with RAS mutations and had no apparent influence on survival. However, c.1781A>G (p.D594G) is unlikely to be benign and more likely to be hypomorphic, as it had significantly fewer cooccurrences with RAS mutations as compared with BRAF wild-type colorectal cancers. Interestingly, our data are consistent with a recent report showing that patients with codon 594 or 596 mutated tumors had longer OS compared with those with c.1799T>A (p.V600E) colorectal cancers (15). There are clear biological differences between these mutant codons to support our observed pathologic differences; p.V600E increased ERK and NFκB signaling and the transformation of NIH3T3 cells, whereas p.D594V failed to activate ERK (16) and did not affect NFκB signaling nor NIH3T3 transforming activity (17).

Others have reported that NRAS-mutant patients have shorter OS as compared with wild-type patients (HR, 1.91; 95% CI, 1.39–3.86; P = 1.3 × 10−03; ref. 18). Here, we noted a more complex relationship; NRAS codon 61 mutations, which were rarely associated with KRAS mutations, conferred a poor prognosis, but codons 12 and 13 mutations, which cooccurred with KRAS mutations at similar frequencies to wild-type colorectal cancers, had little influence on survival. Together, our data suggest that NRAS codons 12 and 13 mutations may have a minor role in colorectal tumorigenesis. Interestingly, using mouse models, others have shown that endogenous levels of Nras p.Q61R, but not Nras p.G12D, were able to efficiently drive in vivo melanomagenesis (19), supporting their differing biological effects.

We have also shown that different mutant loci are associated with differences in the clinicopathology of the primary tumors and/or their sites of metastases. For example, in agreement with two recent reports (20, 21), we observed more KRAS-mutant colorectal cancers in the cecum (70%) and, to a lesser extent, in the right colon (58%), as compared with the left colon (38%). It has been suggested that different somatic profiles are associated with different clinicopathology, by influencing the tumor's biological behavior (22). Here, we focused on intralocus differences and found a significant difference between c.1781A>G (p.D594G) and c.1799T>A (p.V600E) in BRAF in the location of the primary tumor providing additional support for these variants having different biological effects.

In conclusion, our study shows considerable intralocus variations in survival, particularly in the outcomes of mutations in BRAF and NRAS. These data need to be considered in patient management.

No potential conflicts of interest were disclosed.

Conception and design: M.G. Summers, T.S. Maughan, V. Escott-Price, J.P. Cheadle

Development of methodology: M.G. Summers, J.P. Cheadle

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.G. Summers, C.G. Smith, T.S. Maughan, R. Kaplan, J.P. Cheadle

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.G. Summers, C.G. Smith, T.S. Maughan, V. Escott-Price, J.P. Cheadle

Writing, review, and/or revision of the manuscript: M.G. Summers, C.G. Smith, T.S. Maughan, R. Kaplan, V. Escott-Price, J.P. Cheadle

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.G. Summers

Study supervision: T.S. Maughan, J.P. Cheadle

Other (responsible for data analyses and drafting of paper): M.G. Summers

Other (CI of associated study): T.S. Maughan

Other (co-directed this study): V. Escott-Price, J.P. Cheadle

We thank the patients and their families who participated and gave their consent for this research and the investigators and pathologists throughout the United Kingdom who submitted samples for assessment.

This work was supported by the Wales Gene Park, Cancer Research Wales, and the National Institute for Social Care and Health Research Cancer Genetics Biomedical Research Unit. The COIN and COIN-B trials were funded by Cancer Research UK and an educational grant from Merck-Serono. COIN and COIN-B were coordinated by the Medical Research Council Clinical Trials Unit and conducted with the support of the National Institute of Health Research Cancer Research Network.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Walther
 
A
,
Johnstone
 
E
,
Swanton
 
C
,
Midgley
 
R
,
Tomlinson
 
I
,
Kerr
 
D
. 
Genetic prognostic and predictive markers in colorectal cancer
.
Nat Rev Cancer
 
2009
;
9
:
489
–
99
.
Google Scholar
Crossref
Search ADS
PubMed
 
2.
Köhne
 
CH
,
Cunningham
 
D
,
Di Costanzo
 
F
,
Glimelius
 
B
,
Blijham
 
G
,
Aranda
 
E
, et al 
Clinical determinants of survival in patients with 5-fluorouracil-based treatment for metastatic colorectal cancer: results of a multivariate analysis of 3825 patients
.
Ann Oncol
 
2002
;
13
:
308
–
17
.
Google Scholar
Crossref
Search ADS
PubMed
 
3.
Haydon
 
AM
,
Macinnis
 
RJ
,
English
 
DR
,
Giles
 
GG
. 
Effect of physical activity and body size on survival after diagnosis with colorectal cancer
.
Gut
 
2006
;
55
:
62
–
7
.
Google Scholar
Crossref
Search ADS
PubMed
 
4.
Reeves
 
GK
,
Pirie
 
K
,
Beral
 
V
,
Green
 
J
,
Spencer
 
E
,
Bull
 
D
. 
Cancer incidence and mortality in relation to body mass index in the Million Women Study: cohort study
.
BMJ
 
2007
;
335
:
1134
.
Google Scholar
Crossref
Search ADS
PubMed
 
5.
Leitch
 
EF
,
Chakrabarti
 
M
,
Crozier
 
JE
,
McKee
 
RF
,
Anderson
 
JH
,
Horgan
 
PG
, et al 
Comparison of the prognostic value of selected markers of the systematic inflammatory response in patients with colorectal cancer
.
Br J Cancer
 
2007
;
97
:
1266
–
70
.
Google Scholar
Crossref
Search ADS
PubMed
 
6.
Galon
 
J
,
Costes
 
A
,
Sanchez-Cabo
 
F
,
Kirilovsky
 
A
,
Mlecnik
 
B
,
Lagorce-Pages
 
C
, et al 
Type, density and location of immune cells with human colorectal tumors predict clinical outcome
.
Science
 
2006
;
313
:
1960
–
4
.
Google Scholar
Crossref
Search ADS
PubMed
 
7.
Smith
 
CG
,
Fisher
 
D
,
Harris
 
R
,
Maughan
 
TS
,
Phipps
 
AI
,
Richman
 
SD
, et al 
Analyses of 7,635 patients with colorectal cancer using independent training and validation cohorts show that rs9929218 in CDH1 is a prognostic marker of survival
.
Clin Can Res
 
2015
;
21
:
3453
–
61
.
Google Scholar
Crossref
Search ADS
 
8.
Popat
 
S
,
Hubner
 
R
,
Houlston
 
RS
. 
Systematic review of microsatellite instability and colorectal cancer prognosis
.
J Clin Oncol
 
2005
;
23
:
609
–
18
.
Google Scholar
Crossref
Search ADS
PubMed
 
9.
Walther
 
A
,
Houlston
 
R
,
Tomlinson
 
I
. 
Association between chromosomal instability and prognosis in colorectal cancer: a meta-analysis
.
Gut
 
2008
;
57
:
941
–
50
.
Google Scholar
Crossref
Search ADS
PubMed
 
10.
Lochhead
 
P
,
Kuchiba
 
A
,
Imamura
 
Y
,
Liao
 
X
,
Yamauchi
 
M
,
Nishihara
 
R
, et al 
Microsatellite instability and BRAF mutation testing in colorectal cancer prognostication
.
J Natl Can Inst
 
2013
;
105
:
1151
–
6
.
Google Scholar
Crossref
Search ADS
 
11.
Eklöf
 
V
,
Wikberg
 
ML
,
Edin
 
S
,
Dahlin
 
AM
,
Johnsson
 
BA
,
Oberg
 
A
, et al 
The prognostic role of KRAS, BRAF, PIK3CA and PTEN in colorectal cancer
.
Br J Cancer
 
2013
;
108
:
2153
–
63
.
Google Scholar
Crossref
Search ADS
PubMed
 
12.
Maughan
 
TS
,
Adams
 
RA
,
Smith
 
CG
,
Meade
 
AM
,
Seymour
 
MT
,
Wilson
 
RH
, et al 
Addition of cetuximab to oxaliplatin-based first-line combination chemotherapy for treatment of advanced colorectal cancer: results of the randomised phase 3 MRC COIN trial
.
The Lancet
 
2011
;
377
:
2103
–
14
.
Google Scholar
Crossref
Search ADS
 
13.
Smith
 
CG
,
Fisher
 
D
,
Claes
 
B
,
Maughan
 
TS
,
Idziaszczyk
 
S
,
Peuteman
 
G
, et al 
Somatic profiling of the epidermal growth factor receptor pathway in tumours from patients with advanced colorectal cancer treated with chemotherapy ± cetuximab
.
Clin Can Res
 
2013
;
19
:
4104
–
13
.
Google Scholar
Crossref
Search ADS
 
14.
Wasan
 
H
,
Meade
 
AM
,
Adams
 
R
,
Wilson
 
R
,
Pugh
 
C
,
Fisher
 
D
, et al 
Intermittent chemotherapy plus either intermittent or continuous cetuximab for first-line treatment of patients with KRAS wild-type advanced colorectal cancer (COIN-B): a randomised phase 2 trial
.
Lancet Oncol
 
2014
;
15
:
631
–
9
.
Google Scholar
Crossref
Search ADS
PubMed
 
15.
Cremolini
 
C
,
Di Bartolomeo
 
M
,
Amatu
 
A
,
Antoniotti
 
C
,
Moretto
 
R
,
Berenato
 
R
, et al 
BRAF codons 594 and 596 mutations identify a new molecular subtype of metastatic colorectal cancer at favorable prognosis
.
Ann Oncol
 
2015
;
26
:
2092
–
7
.
Google Scholar
Crossref
Search ADS
PubMed
 
16.
Wan
 
PT
,
Garnett
 
MJ
,
Roe
 
SM
,
Lee
 
S
,
Niculescu-Duvaz
 
D
,
Good
 
VM
, et al 
Mechanism of activation of the RAF-ERK signaling pathway by oncogenic mutations of B-RAF
.
Cell
 
2004
;
116
:
855
–
67
.
Google Scholar
Crossref
Search ADS
PubMed
 
17.
Ikenoue
 
T
,
Hikiba
 
Y
,
Kanai
 
F
,
Tanaka
 
Y
,
Imamura
 
J
,
Imamura
 
T
, et al 
Functional analysis of mutations within the kinase activation segment of B-Raf in human colorectal tumors
.
Can Res
 
2003
;
63
:
8132
–
7
.
Google Scholar
 
18.
Schirripa
 
M
,
Cremolini
 
C
,
Loupakis
 
F
,
Morvillo
 
M
,
Bergamo
 
F
,
Zoratto
 
F
, et al 
Role of NRAS mutations as prognostic and predictive markers in metastatic colorectal cancer
.
Int J Cancer
 
2015
;
136
:
83
–
90
.
Google Scholar
Crossref
Search ADS
PubMed
 
19.
Burd
 
CE
,
Liu
 
W
,
Huynh
 
MV
,
Waqas
 
MA
,
Gillahan
 
JE
,
Clark
 
KS
, et al 
Mutation-specific RAS oncogenicity explains NRAS codon 61 selection in melanoma
.
Can Discov
 
2014
;
4
:
1418
–
29
.
Google Scholar
Crossref
Search ADS
 
20.
Yamauchi
 
M
,
Morikawa
 
T
,
Kuchiba
 
A
,
Imamura
 
Y
,
Qian
 
ZR
,
Nishihara
 
R
, et al 
Assessment of colorectal cancer molecular features along bowel subsites challenges the conception of distinct dichotomy of proximal versus distal colorectum
.
Gut
 
2012
;
61
:
847
–
54
.
Google Scholar
Crossref
Search ADS
PubMed
 
21.
Rosty
 
C
,
Young
 
JP
,
Walsh
 
MD
,
Clendenning
 
M
,
Walters
 
RJ
,
Pearson
 
S
, et al 
Colorectal carcinomas with KRAS mutation are associated with distinctive morphological and molecular features
.
Mod Pathol
 
2013
;
26
:
825
–
34
.
Google Scholar
Crossref
Search ADS
PubMed
 
22.
Tran
 
B
,
Kopetz
 
S
,
Tie
 
J
,
Gibbs
 
P
,
Jiang
 
ZQ
,
Lieu
 
CH
, et al 
Impact of BRAF mutation and microsatellite instability on the pattern of metastatic spread and prognosis in metastatic colorectal cancer
.
Cancer
 
2011
;
117
:
4623
–
32
.
Google Scholar
Crossref
Search ADS
PubMed
 
©2016 American Association for Cancer Research.
2016
American Association for Cancer Research.

Supplementary data

Supplementary Figure and Tables

Figure - Correlation plot showing inter- and intra-genic relationships between somatic KRAS, BRAF and NRAS mutations and MSI status in aCRC. Tables - Heterogeneity Tests comparing OS to diagnosis to death (Table S1) and the effect of cetuximab on OS (Table S2).

- docx file