Colorectal liver metastases are more often super wild type. toward treatment based on metastatic site genotyping? M. A. Allard1,2,3; R. Saffroy1,2; P. Bouvet de la Maisonneuve1,2; L. Ricca1,2,3,4; N. Bosselut1,2; J. Hamelin1,2; E. Lecorche1,2, M.A. Bejarano1,2; P. Innominato2,5,7; M. Sebagh6; R. Adam3,7; J. F. Morère2,5; A. Lemoine1,2,8*<c>00 33 1 45 59 36 93, antoinette.lemoine@pbr.aphp.fr 1 APHP, Biochemistry and Oncogenetic Department, Paul Brousse Hospital Groupe Hospitalier de Hôpitaux Universitaires Paris-Sud Villejuif, France 2 Inserm UMRS-1004, Université Paris-11 Villejuif, France 3 APHP, Centre Hépato-biliaire, Paul Brousse Hospital Groupe Hospitalier de Hôpitaux Universitaires Paris-Sud Villejuif, France 4 Dipartimento di Scienze Medico-Chirurgiche e Medicina Traslazionale Sapienza Università di Roma Sapienza, Italy 5 APHP , Department of Medical Oncology, Paul Brousse Hospital Groupe Hospitalier de Hôpitaux Universitaires Paris-Sud Villejuif, France 6 APHP Department of Pathology Groupe Hospitalier de Hôpitaux Universitaires Paris-Sud Villejuif, France 7 INSERM UMRS-776 Villejuif, France 1 8 Service de Biochimie Hôpital Paul Brousse 12 av Paul Vaillant Couturier, 94804, Villejuif cedex, France Abstract Recent data showed that metastatic colorectal (mCRC) tumors exhibiting extended RASBRAF mutations were resistant to anti-EGFR monoclonal antibodies, making these drugs suitable for the so-called “super” wild type (WT) patients only. This study aimed to compare the extended RAS-BRAF mutation frequency and characteristics according to location of tumor sampling. All consecutive mCRC specimens (N=1,659) referred to our institution from January 2008 till June 2014 were included in the analysis. Tumor genotyping (first for KRAS exon 2, then for BRAF exon 15 and later for KRAS exons 2,3,4 and NRAS exon 2,3,4) was performed with high resolution melting analysis or allelic discrimination. The factors predicting for the presence of mutation were explored using multivariate binary logistic regression. Overall, the prevalence of KRAS exon 2 was 36.8%, and it was lower in liver metastases (N=138/490; 28.2%) in comparison with primary tumors (N=442/1086; 40.7%), lung metastases (16/32; 50%) or others metastatic sites (15/51; 29.4%; P<0.0001). Similarly, in the 1,428 samples analyzed, BRAF mutations were less often found in liver metastases (N=9/396; 2.3%) as compared to primary tumors (N=79/959; 8.2%), lung metastases (N=2/29; 6.9%) or others metastatic locations (N=2/44; 4.5%; P<0.0002). Overall occurrence of extended RAS mutation was 51.7%. Of the 503 samples tested, the prevalence of extended RAS-BRAF mutations twice as low in liver metastases (N=53/151; 34.2%) as compared to primary tumors (N=191/322; 59.3%, P<0.0001). Univariate analysis identified age ≤ 65 years, male gender and liver localization as predictors of “super” WT status. At multivariate analysis, only liver metastases was retained (RR: 2.85 [95% CI: 1.91 – 4.30]). Colorectal liver metastases are twice as likely to exhibit a “super” WT genotype as compared to other tumor locations, independently from other factors. This molecular feature has the potential to influence therapeutic strategy in mCRC patients. Keywords 2 KRAS, NRAS, BRAF, “Super” wild type, Metastatic colorectal cancer INTRODUCTION Anti-Epidermal Growth Factor Receptor (EGFR) therapies have been shown to increase tumor response rate and survival in metastatic colorectal cancer (mCRC) [1–4]. Several studies have clearly demonstrated that activating mutations in KRAS (exon 2) and BRAF (exon 15), both involved in signal transduction downstream of EGFR, confer resistance to anti-EGFR therapies [2, 5–7]. Recent data [8–10] have shown that tumors carrying mutations in KRAS exon 2,3,4 and in NRAS exon 2,3 and 4 predict for the lack of activity of anti-EFGR therapies. This has led to the restriction to the use of anti EGFR therapies only in patients with “super” wild type (WT) metastatic colorectal cancers. In most cases in daily practice, tumor RAS mutational status is assessed in samples issued from primary tumor at the time of diagnostic colonoscopy. The rationale for the use antiEGFR monoclonal antibodies in mCRC patients relies on the good concordance of mutational status between primary tumor and metastatic sites, shown in several studies [11–13]. However, significant differences in the incidence of KRAS exon 2 mutations among tumor locations have been reported [14]. The aim of the present study was to compare the prevalence of extended RAS-BRAF mutation according to tumor location in a prospective cohort of consecutive mCRC samples and to identify factors associated with “super” WT status. MATERIALS AND METHODS Study population The department of molecular biology at the Paul Brousse Hospital is a platform, labellized by the French Health Cancer Institute (INCa, Institut National du Cancer) to routinely perform molecular diagnosis of mutations observed in solid malignancies that can be targeted by tyrosine kinase inhibitors (TKIs). CRC genotyping is only indicated in metastatic patients. Thus, mCRC samples are daily genotyped for a list of mutations, in accordance with French guidelines [15]. Routine tumor genotyping for KRAS exon 2 started in January 2008; later, 3 BRAF exon 15 genotyping was added in June 2008. Finally, genotyping was extended to KRAS exon 3, 4 and NRAS exon 2, 3 and 4 since September 2013. Thus, at July 2014, our study population consists of 1659 tumors for KRAS exon 2, 1428 for BRAF exon 15, and 503 for extended RAS-BRAF genotyping. All results and basic clinical data have been collected since January 2008 in a prospectively maintained-database. RS, NB and JH are specifically responsible for entering data on a daily basis. Clinical variables include age, sex, origin of sample organ, percentage of tumor cell, date of analysis and mutational status. DNA samples Hematoxylin-eosin-stained slides from tumors are assessed for the ratio percentage of tumor cells/sample area (non tumoral tissue, stroma of the tumor) by pathologists of our department. Three sections of 10 µm thickness were obtained from the paraffin-embedded tissue containing at least 10% of tumoral cells. DNA extraction was performed using the QIAmp DNA Mini kit (Qiagen, Courtaboeuf, France) according to the manufacturer instructions. Somatic gene mutations were detected both by the High Resolution Melting analysis, as described [16–18]. and allelic discrimination with the LightCycler 480 system (Roche). Statistical analysis Chi square test or Fisher exact test were used to compare categorical data, as appropriate. Test T was used to compare means. All variables associated with “super” WT status (P < 0.15) were entered into a multiple regression logistic model. Multivariate analysis was performed according to a backward stepwise procedure. The threshold for statistical significance was set for P<0.05. Calculations were performed with R 3.0.0. RESULTS Study population and overall mutation frequency (Table 1) Clinical and demographic features of the study population genotyped for KRAS exon 2 (N=1659) are detailed in Table 1, together with those of the more recent subgroup genotyped for extended RAS-BRAF (N=503). The two groups were similar in terms of age, sex ratio, locations of tumor sampling and KRAS exon 2 or BRAF exon 15 mutation frequencies. 4 Overall incidence of KRAS exon 2 mutation was 36.8% , and that of BRAF exon 15 mutation was 6.4% Similar rates were observed in the subgroup with extended RAS-BRAF genotyping: 42.1% for KRAS exons 2,3,4 mutations and 6.2% for BRAF exon 15 mutations. NRAS exon 2,3, 4 mutations were found in 18 (3.6%) samples of the 503 analyzed. Codon 12 was more frequently involved (76.8%) than codon 13 in KRAS exon 2 mutations. KRAS exon 2 and BRAF mutations by origin of tumor sample (Table 2) KRAS exon 2 mutations frequency was higher in samples originating from primary intestinal lesions (40.7%) or lung localizations (50.0%) in comparison to liver (28.2%) or others metastatic sites (29.4%; P < 0.0001). Similarly, BRAF exon 15 mutations were observed more often in primary tumors or lung metastases (8.2% and 6.9%, respectively) as compared to liver or others metastatic sites (2.3% and 4.5%, respectively; P=0.0002). In KRAS exon 2 mutations, no significant difference according to organ of origin was observed for the codon involved (P=0.65). Mutation frequency by origin of tumor sample in the group genotyped for extended RASBRAF (Table 3) The prevalence of extended RAS-BRAF mutations was 59.3% in primary tumors, 55.6% in lung lesions, 56.5% in other metastatic sites and 34.2% in liver metastases (P<0.0001). In this subgroup, as for the whole study population, KRAS and BRAF mutations were both significantly less frequent in liver metastases (28.9% and 2.8%) as compared with primary tumors (47.5% and 8.1%; P=0.0002 and P= 0.04, respectively) The incidence of NRAS mutations was similar according to sampled organ, but they were only found in 18 patients. Factors associated with super WT status: univariate and multivariate analysis (Table 4) Clinical variables tested as potential predictors of super WT status included age, gender, type surgical specimen (biopsy vs surgery) and sample location organ. At univariate analysis, super WT status was more frequently observed in males, ≤ 65 years and in liver metastases. Multivariate analysis showed that liver localization (RR: 2.85 [1.91 – 4.30]) and male sex (RR 1.83 [1.26 – 2.67]) were independent predictors of a super WT status. DISCUSSION 5 In this large monocentric cohort of consecutive samples from metastatic colorectal cancer, we report that the prevalence of KRAS exon 2 and BRAF exon 15 mutations, as well extended RAS/RAF mutation genotype, was significantly lower (about 2-fold) in liver metastases than in primary tumors or other metastatic sites. The overall frequency of KRAS exon 2 mutations observed in our cohort was in the range of the frequency usually described, albeit their prevalence is subject to variations [19, 20]. Indeed, in three multicentric phase III trials, the prevalence of KRAS exon 2 mutation ranged from 30 to 45% [2, 3, 6]. Although there is no clear explanation to account for these differences, it is likely that variability in demographic features, techniques of mutation detection used, and the tumor location analyzed may be involved. Several studies have investigated mutational concordance for RAS between primary tumor and metastatic sites [11, 12, 21–24]. Discordance rates range from 0% to 25%, depending on the number of patients, the metastatic sites sampled and stage of the disease. Globally, it appears that there is a good level of concordance between primary tumor and metastases [13]. A lower frequency of KRAS exon 2 mutations in liver lesions (32.3%) as compared to other sites (lung: 62.0%; brain: 56.5%) has been reported by Tie et al [14]. Similarly, Vauthey et al [25] observed a KRAS (codon 12 and 13) mutation frequency of 16.1%, even lower than the present data. These and our data, clearly show significant differences in KRAS mutations incidence according to the tumor location. Mutations in BRAF exon 15 have been associated with microsatellite instability and mismatch repair deficient status in non-metastatic patients [26] and have been shown to predict poor survival outcome [4]. The mutational rate for BRAF in primary tumor in our study population is similar to that observed in others large cohorts, in which the BRAF mutation rate is about 8% [27]. In the present study, the rate of BRAF mutation in liver metastases was 2.3%, significantly lower than that in primary tumor. This finding is in line with previous studies that reported a rate of BRAF mutation ranging from 1% to 3.2% in liver lesions [14, 25]. These results suggest that BRAF mutated primary tumor may have a lower propensity to metastasize in the liver. Alongside these confirmatory observations in a much larger sample size, our main finding is that liver metastases are more often super WT as compared to others tumor locations and particularly primary tumor. This difference was true when considering KRAS exon 2 only, 6 and become greater when considering the up to date genotyping recommendations (ie, extended RAS-BRAF mutations). This result is strengthened by multivariate analysis that enables to exclude potential confounding factors. A few hypotheses may be postulated in order to explain these differences according to tumor location. First, discordance alone is unlikely to account for the observed differences in mutation frequencies. Studies dealing with the issue of genotype concordance, consider the genotype of a tumor location at a given time of the history of the disease. Even if this cannot be demonstrated here, it is likely that genotype may evolve during the evolution of the disease over the time, because of treatment or intrinsic tumor evolution [28]. Another factor involved could be the intratumoral heterogeneity. This feature has been previously described in CRC [29–31]. Baldus et al. [31] compared tumor centers and invasion front of 100 primary CRC tumors and found a rate of discrepancy of 8% for KRAS exon 2 and 1% for BRAF. Giaretti et al[30] found a 33% of KRAS heterogeneity defined by the presence of mutated and WT clone within the same tumor. Liver metastases are known to exhibit spontaneous fibrosis or necrosis [32] and may be subject to o a higher heterogeneity, as compared to other locations. The mechanisms underlying metastatic process are only very partially understood [33]. On one hand, concordance between primary tumor and metastases for oncogene support the concept that KRAS mutation is an event occurring early in the tumorigenesis and may confer specific metastatic properties in accordance with the theory proposed by Bernards and Weinberg [34]. On the other hand, discordance argues in favor of clonal selection process as already advocated [35] and established in other malignancy [36] Lung metastases exhibited the highest frequency of KRAS exon 2 and extended RAS-BRAF mutations compared to liver metastases. Moreover, clinical data have clearly show that RAS mutated patents have a high propensity to develop lung metastases [14, 25]. Our results are in line with these observations which suggest the importance of the tumor-host tissue interaction [37], a concept evocated a century ago as the “seed and soil” theory and highlight the relevancy of lung surveillance in RAS mutated patients [14, 38]. The lower frequency of RAS mutation in liver metastases in comparison with primary tumors suggests that liver metastases may arise from a WT clone in patients harboring a mutated primary tumor. This uncommon situation may be advanced to explain unexpected disease control with anti EGFR therapies in mutated patients (0 to 11% of patients according to these recent clinical series) [39–41]. 7 Another possible clinical explanation could involve a selection bias: samples from organs other than primary lesion involved, most of the times, a surgical resection of the metastases. Since very rarely surgery of metastatic lesions is indicated in case of disease progression, the higher rate of super WT tumors in the liver could be partially due to the fact that this molecular subgroup is more likely to respond to chemotherapy including anti-EGFR targeted drugs. Hence, samples obtained from resected metastases could be the result of less extensive disease dissemination related to intrinsic less aggressive biological features or higher sensitivity to preoperative chemotherapy. However, lesions obtained from sites other than the liver seem to display a similar incidence of RAS-BRAF mutations as compared to primary tumor, suggesting that liver metastases could have a different molecular pattern. The lower frequency of extended RAS-BRAF mutations in liver metastases compared to primary tumor therefore supports the proposal of considering genotyping of liver metastases in patients with mutated primary tumor. This attitude may identify about 10 to 15 % of the mCRC patients with a mutated tumor but with WT liver metastases. The superiority of anti EGFR therapy over the alternative treatment such as anti VEGF therapy in WT patients remains a matter of debate [42, 43], which precludes any recommendations. However, if anti EGFR therapy becomes the standard in the future, liver metastases genotyping in primary tumor-mutated patients may be worth to consider. Conflict of interest The authors have no conflicts of interest. References 1. Cunningham D, Humblet Y, Siena S, et al. (2004) Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. The New England journal of medicine 351:337–45. doi: 10.1056/NEJMoa033025 2. Douillard J-Y, Siena S, Cassidy J, et al. (2010) Randomized, phase III trial of panitumumab with infusional fluorouracil, leucovorin, and oxaliplatin (FOLFOX4) versus FOLFOX4 alone as first-line treatment in patients with previously untreated metastatic colorectal cancer: the PRIME study. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 28:4697–705. doi: 10.1200/JCO.2009.27.4860 3. Van Cutsem E, Peeters M, Siena S, et al. (2007) Open-label phase III trial of panitumumab plus best supportive care compared with best supportive care alone in patients with chemotherapy-refractory metastatic colorectal cancer. Journal of clinical oncology : 8 official journal of the American Society of Clinical Oncology 25:1658–64. doi: 10.1200/JCO.2006.08.1620 4. Douillard J-Y, Oliner KS, Siena S, et al. (2013) Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. The New England journal of medicine 369:1023– 34. doi: 10.1056/NEJMoa1305275 5. Lièvre A, Bachet J-B, Le Corre D, et al. (2006) KRAS mutation status is predictive of response to cetuximab therapy in colorectal cancer. Cancer research 66:3992–5. doi: 10.1158/0008-5472.CAN-06-0191 6. Peeters M, Price TJ, Cervantes A, et al. (2010) Randomized phase III study of panitumumab with fluorouracil, leucovorin, and irinotecan (FOLFIRI) compared with FOLFIRI alone as second-line treatment in patients with metastatic colorectal cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 28:4706–13. doi: 10.1200/JCO.2009.27.6055 7. Mao C, Liao R-Y, Qiu L-X, et al. (2011) BRAF V600E mutation and resistance to antiEGFR monoclonal antibodies in patients with metastatic colorectal cancer: a metaanalysis. Molecular biology reports 38:2219–23. doi: 10.1007/s11033-010-0351-4 8. Douillard J-Y, Oliner KS, Siena S, et al. (2013) Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. The New England journal of medicine 369:1023– 34. doi: 10.1056/NEJMoa1305275 9. De Roock W, Claes B, Bernasconi D, et al. (2010) Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapyrefractory metastatic colorectal cancer: a retrospective consortium analysis. The lancet oncology 11:753–62. doi: 10.1016/S1470-2045(10)70130-3 10. Sorich MJ, Wiese MD, Rowland A, et al. (2014) Annals of Oncology Advance Access published August 12 , 2014. 1–27. 11. Santini D, Loupakis F, Vincenzi B, et al. (2008) High concordance of KRAS status between primary colorectal tumors and related metastatic sites: implications for clinical practice. The oncologist 13:1270–5. doi: 10.1634/theoncologist.2008-0181 12. Knijn N, Mekenkamp LJM, Klomp M, et al. (2011) KRAS mutation analysis: a comparison between primary tumours and matched liver metastases in 305 colorectal cancer patients. British journal of cancer 104:1020–6. doi: 10.1038/bjc.2011.26 13. Han C-B, Li F, Ma J-T, Zou H-W (2012) Concordant KRAS mutations in primary and metastatic colorectal cancer tissue specimens: a meta-analysis and systematic review. Cancer investigation 30:741–7. doi: 10.3109/07357907.2012.732159 14. Tie J, Lipton L, Desai J, et al. (2011) KRAS mutation is associated with lung metastasis in patients with curatively resected colorectal cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 17:1122–30. doi: 10.1158/1078-0432.CCR-10-1720 9 15. Institut National du Cancer (2010) Bonnes pratiques pour la recherche à visée théranostique demutations somatiques dans les tumeurs solides. wwwe-cancerfr 16. Mancini I, Pinzani P, Pupilli C, et al. (2012) A high-resolution melting protocol for rapid and accurate differential diagnosis of thyroid nodules. The Journal of molecular diagnostics : JMD 14:501–9. doi: 10.1016/j.jmoldx.2012.03.003 17. Simi L, Pratesi N, Vignoli M, et al. (2008) High-resolution melting analysis for rapid detection of KRAS, BRAF, and PIK3CA gene mutations in colorectal cancer. American journal of clinical pathology 130:247–53. doi: 10.1309/LWDY1AXHXUULNVHQ 18. Willmore-Payne C, Holden JA, Hirschowitz S, Layfield LJ (2006) BRAF and c-kit gene copy number in mutation-positive malignant melanoma. Human pathology 37:520–7. doi: 10.1016/j.humpath.2006.01.003 19. Brink M, de Goeij AFPM, Weijenberg MP, et al. (2003) K-ras oncogene mutations in sporadic colorectal cancer in The Netherlands Cohort Study. Carcinogenesis 24:703–10. 20. Samowitz WS, Curtin K, Schaffer D, et al. (2000) Relationship of Ki- ras Mutations in Colon Cancers to Tumor Location , Stage , and Survival : A Population-based Study Relationship of Ki- ras Mutations in Colon Cancers to Tumor Location ,. 1193–1197. 21. Cejas P, López-Gómez M, Aguayo C, et al. (2009) KRAS mutations in primary colorectal cancer tumors and related metastases: a potential role in prediction of lung metastasis. PloS one 4:e8199. doi: 10.1371/journal.pone.0008199 22. Park JH, Han S-W, Oh D-Y, et al. (2011) Analysis of KRAS, BRAF, PTEN, IGF1R, EGFR intron 1 CA status in both primary tumors and paired metastases in determining benefit from cetuximab therapy in colon cancer. Cancer chemotherapy and pharmacology 68:1045–55. doi: 10.1007/s00280-011-1586-z 23. Losi L, Benhattar J, Costa J (1992) Stability of K-ras mutations throughout the natural history of human colorectal cancer. European journal of cancer (Oxford, England : 1990) 28A:1115–20. 24. Weber J-C, Meyer N, Pencreach E, et al. (2007) Allelotyping analyses of synchronous primary and metastasis CIN colon cancers identified different subtypes. International journal of cancer Journal international du cancer 120:524–32. doi: 10.1002/ijc.22343 25. Vauthey J-N, Zimmitti G, Kopetz SE, et al. (2013) RAS mutation status predicts survival and patterns of recurrence in patients undergoing hepatectomy for colorectal liver metastases. Annals of surgery 258:619–26; discussion 626–7. doi: 10.1097/SLA.0b013e3182a5025a 26. Gavin PG, Colangelo LH, Fumagalli D, et al. (2012) Mutation profiling and microsatellite instability in stage II and III colon cancer: an assessment of their prognostic and oxaliplatin predictive value. Clinical cancer research : an official journal of the American Association for Cancer Research 18:6531–41. doi: 10.1158/1078-0432.CCR-12-0605 10 27. Venderbosch S, Nagtegaal ID, Maughan TS, et al. (2014) Mismatch repair status and BRAF mutation status in metastatic colorectal cancer patients: a pooled analysis of the CAIRO, CAIRO2, COIN and FOCUS studies. Clinical Cancer Research. doi: 10.1158/1078-0432.CCR-14-0332 28. Misale S, Yaeger R, Hobor S, et al. (2012) Emergence of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal cancer. Nature 486:532–6. doi: 10.1038/nature11156 29. Losi L, Baisse B, Bouzourene H, Benhattar J (2005) Evolution of intratumoral genetic heterogeneity during colorectal cancer progression. Carcinogenesis 26:916–22. doi: 10.1093/carcin/bgi044 30. Giaretti W, Monaco R, Pujic N, et al. (1996) Intratumor heterogeneity of K-ras2 mutations in colorectal adenocarcinomas: association with degree of DNA aneuploidy. The American journal of pathology 149:237–45. 31. Baldus SE, Schaefer K-L, Engers R, et al. (2010) Prevalence and heterogeneity of KRAS, BRAF, and PIK3CA mutations in primary colorectal adenocarcinomas and their corresponding metastases. Clinical cancer research : an official journal of the American Association for Cancer Research 16:790–9. doi: 10.1158/1078-0432.CCR-09-2446 32. Poultsides G a, Bao F, Servais EL, et al. (2012) Pathologic response to preoperative chemotherapy in colorectal liver metastases: fibrosis, not necrosis, predicts outcome. Annals of surgical oncology 19:2797–804. doi: 10.1245/s10434-012-2335-1 33. Talmadge JE, Fidler IJ (2010) AACR centennial series: the biology of cancer metastasis: historical perspective. Cancer research 70:5649–69. doi: 10.1158/0008-5472.CAN-101040 34. Bernards R, Weinberg R a (2002) A progression puzzle. Nature 418:823. doi: 10.1038/418823a 35. Fidler IJ, Kripke ML (1977) Metastasis results from preexisting variant cells within a malignant tumor. Science (New York, NY) 197:893–5. 36. Wu X, Northcott P a, Dubuc A, et al. (2012) Clonal selection drives genetic divergence of metastatic medulloblastoma. Nature 482:529–33. doi: 10.1038/nature10825 37. Langley RR, Fidler IJ (2007) Tumor cell-organ microenvironment interactions in the pathogenesis of cancer metastasis. Endocrine reviews 28:297–321. doi: 10.1210/er.20060027 38. Kim M-J, Lee HS, Kim JH, et al. (2012) Different metastatic pattern according to the KRAS mutational status and site-specific discordance of KRAS status in patients with colorectal cancer. BMC cancer 12:347. doi: 10.1186/1471-2407-12-347 39. Lièvre A, Bachet J-B, Boige V, et al. (2008) KRAS mutations as an independent prognostic factor in patients with advanced colorectal cancer treated with cetuximab. 11 Journal of clinical oncology : official journal of the American Society of Clinical Oncology 26:374–9. doi: 10.1200/JCO.2007.12.5906 40. Di Fiore F, Blanchard F, Charbonnier F, et al. (2007) Clinical relevance of KRAS mutation detection in metastatic colorectal cancer treated by Cetuximab plus chemotherapy. British journal of cancer 96:1166–9. doi: 10.1038/sj.bjc.6603685 41. Khambata-Ford S, Garrett CR, Meropol NJ, et al. (2007) Expression of epiregulin and amphiregulin and K-ras mutation status predict disease control in metastatic colorectal cancer patients treated with cetuximab. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 25:3230–7. doi: 10.1200/JCO.2006.10.5437 42. Heinemann V, von Weikersthal LF, Decker T, et al. (2014) FOLFIRI plus cetuximab versus FOLFIRI plus bevacizumab as first-line treatment for patients with metastatic colorectal cancer (FIRE-3): a randomised, open-label, phase 3 trial. The lancet oncology 2045:1–11. doi: 10.1016/S1470-2045(14)70330-4 43. Sclafani F, Cunningham D (2014) Cetuximab or bevacizumab in metastatic colorectal cancer? The lancet oncology 2045:2–3. doi: 10.1016/S1470-2045(14)70360-2 Table 1. Overview of the study population Variables Mean age, yrs (SD,range) Gender M/F Site of tumor sampling Primary tumor Liver Lung Others*/** Origin of sampling Surgical specimen Biopsy Mutations KRAS exon 2 KRAS codon 12 KRAS codon 13 BRAF exon 15 extended RAS-RAF mutated status KRAS exon 3 KRAS exon 4 KRAS exon 2 only N = 1,156 No. (%) 65.9 (13) 637 (55.1)/519 (44.9) Extended RAS N = 503 No. (%) 65.3 (14) 289 (57.5)/214 (42.5) Total P 0.41 0.99 N = 1,659 No. (%) 65.7 (13) 926 (55.8)/733 (44.2) 764 (66.1) 341 (29.5) 23 (2.0) 28 (2.4) 322 (64) 149 (29.6) 9 (1.8) 23* (4.6) 0.13 1086 (65.5) 490 (29.5) 32 (1.9) 51** (3.1) 860 (76.6) 263 (23.4) 408 (81.1) 95 (18.9) 0.05 1268 (76.4) 358 (21.6) 425 (36.7) 320 (27.7) 105 (9.0) 60 (6.4) 186 (37) 149 (29.6) 37 (7.6) 32 (6.2) 0.97 0.45 0.29 0.99 611 (36.8) 469 (28.3) 142 (8.6) 92/1428 (6.4) 12 260 (51.7) _ 6 (1.2) 20 (4) _ _ NRAS 18 (3.6) _ NRAS exon 2 6 (1.2) _ NRAS exon 3 12 (2.4) _ NRAS exon 4 0 (0) _ * Peritoneum (N =15); Lymph node (N=6); Brain (N=1); Adrenal gland (N = 1) **Others: Peritoneal metastases (N=24 ); Lymph node (N=14); skin metastases (N=2); pancreatic metastasis (N=1); brain metastasis (N= 1); adrenal gland metastasis (N=1 ); bone metastases (N=2 ); ovarian metastases (N=6) Table 2. KRAS exon 2 and BRAF mutations by origin of tumor sampling Origin of tumor sampling P Primary Liver Lung Others* No. (%) No. (%) No. (%) No. (%) KRAS exon 2 (N=1,659) WT 644 (59.3) 352 (71.8) 16 (50) 36 (70.6) <0.0001 Mutant 442 (40.7) 138 (28.2) 16 (50) 15 (29.4) Codon 12 335 (30.8) 111 (22.7) 12 (37.5) 11 (21.6) 0.003 Codon 13 107 (9.9) 27 (5.5) 4 (12.5) 4 (7.8) 0.01 BRAF (N = 1,428) WT 880 (91.8) 387 (97.7) 27 (93.1) 42 (95.5) 0.0002 Mutant 79 (8.2) 9 (2.3) 2 (6.9) 2 (4.5) *Others: Peritoneal metastases (N=24 ); Lymph node (N=14); ovarian metastases (N=6); skin metastases (N=2); pancreatic metastasis (N=1); brain metastasis (N= 1); adrenal gland metastasis (N=1 ); bone metastases (N=2 ) Table 3. Mutations by origin of sampling in the cohort genotyped for extended RAS-BRAF Origin of tumor sampling Primary Liver Lung No. No. (%) No. (%) (%) Extended RAS-BRAF Super WT Mutant KRAS WT Mutant KRAS exon 2 Codon 12 Others* P No. (%) 131 (40.7) 191 (59.3) 98 (65.8) 53 (34.2) 4 (44.4) 10 (43.5) <0.0001 5 (55.6) 13 (56.5) 169 (52.5) 153 (47.5) 138 (42.9) 112 (34.8) 106 (71.1) 43 (28.9) 35 (23.5) 29 (19.5) 5 (55.6) 4 (44.4) 4 (44.4) 2 (22.2) 13 11 (47.8) 0.0009 12 (52.2) 9 (39.1) 0.0004 6 (26.1) 0.005 Codon 13 KRAS exon 3 KRAS exon 4 BRAF WT Mutant NRAS WT Mutant NRAS exon 2 NRAS exon 3 NRAS exon 4 26 (8.1) 4 (1.2) 11 (3.4) 6 (4.0) 3 (2.0) 5 (3.4) 2 (22.2) 3 (13.0) 0 (0) 0 (0) 0 (0) 3 (13) 296 (91.9) 26 (8.1) 145 (97.2) 9 (100) 22 (95.7) 0.12 4 (2.8) 0 (0) 1 (4.3) 309 (96) 13 (4.0) 4 (1.2) 9 (2.8) 0 (0) 145 (97.3) 4 (2.7) 2 (1.3) 2 (1.3) 0 (0) 8 (88.9) 1 (11.1) 0 (0) 0 (0) 0 (0) 23 (100) 0 (0) 0 (0) 0 (0) 0 (0) 0.05 0.80 0.15 0.36 0.99 0.26 >0.99 Table 4. Factors associated with “super” WT status: univariate and multivariate analysis Variables ≤ 65 yrs Age > 65 yrs Female Gender Male Surgical spe. Biopsy Sampling Liver location No Yes Univariate analysis Super Mutant WT 82 (33.7) 72 (27.7) 161 188 (66.3) (72.3) 127 87 (35.8) (48.8) 156 133 (64.2) (51.2) 200 208 (80) (82.3) 43 (17.7) 52 (20) 145 209 (59.7) (80.4) 98 (40.3) 51 (19.6) Multivariate analysis P RR 95%CI P 0.14 0.71 0.48 - 1.06 0.09 0.003 1.83 1.26 - 2.67 0.001 0.51 <0.0001 2.85 1.91 - 4.31 <0.0001 Supplementary table: Factors associated with BRAF mutation: univariate and multivariate analysis Age ≤ 65 yrs > 65 yrs Female Gender Male Univariate analysis for BRAF mutation WT Mutant P 642 (48) 31 (33) 0.009 653 (52) 61 (67) 65 575 (43) <0.0001 (70.7) 27 761 (57) (29.3) 14 Multivariate analysis RR 95%CI 1.84 1.15 - 3.00 P 0.01 0.29 0.17 - 0.46 <0.0001 Surgical spe. Sampling Biopsy Tumor location Primary tumor & Other loc. Liver 987 (75.2) 326 (24.8) 69 (76.7) 21 (23.3) 949 (71) 83 (90) 387 (29) 9 (10) 15 0.75 <0.0001 0.32 0;15 - 0.62 0.001