Combined assessment of the TNM stage and BRAF mutational status at diagnosis in sporadic colorectal cancer patients

INTRODUCTION

Sporadic colorectal cancer (sCRC) ranks the third most common type of cancer worldwide, both in men and women, and the fourth leading cause of cancer-related death [1, 2]. In the last decades, the introduction of targeted therapies has led to great progress in tumor response to treatment and patient survival in many cancer types, including sCRC, but only selected subgroups of patients within a given tumor type may benefit from these novel agents (e.g. anti-EGFR and anti-VEGF therapies) based on better responses to therapy and a significantly improved outcome [3]. At present, the prognostic impact of KRAS mutations, and of mutations in other KRAS-related and unrelated genes such as NRAS and TP53 frequently mutated in sCRC, remains controversial, particularly as regards their impact on patient progression-free survival (PFS) and overall survival (OS) [4–6]. Despite this, the presence of KRAS and NRAS gene mutations, currently precludes administration of anti-EGFR directed therapies in sCRC patients; this translating into a lack of benefit of such novel therapies in this large subgroup of sCRC patients [7]. In turn, preliminary data suggests that the activating BRAF V600E mutation that occurs in a smaller fraction (5-10%) of sCRC patients who lack KRAS mutations, might be associated with a worse patient outcome [8] and significantly shorter OS rates. [9, 10]. Similarly, NRAS mutations occur in only a small fraction (1-6%) of all colorectal tumors [11] and they have been associated with reduced response to monoclonal antibody therapies [12], its potential prognostic value among early stage tumors, still requiring further confirmation in larger series of unselected sCRC patients. Finally, despite TP53 mutations are frequently detected in colorectal tumors (~40% of cases), their specific prognostic impact on the outcome and survival of sCRC patients, still remains controversial.

Since 2013, the European Medicines Agency (EMEA) requires that exons 2-3-4 of both KRAS and NRAS are investigated prior to usage of novel targeted (e.g. anti-EGFR) therapies. From the onward, the number of molecular targets required to be analyzed in sCRC tumors has further expanded and it is expected to increase even more in the near future. Because of this (and other reasons) Sanger sequencing and other traditional sequencing approaches that interrogate the tumor DNA for specific variants of one (or a few) genes are being progressively replaced by next-generation sequencing (NGS). NGS allows simultaneous analysis of multiple gene targets with higher sensitivity, and at lower cost, using reduced amounts of DNA [13]. Thus, an NGS-based approach for simultaneous evaluation of KRAS, NRAS, BRAF, and TP53 gene mutations in hot-spot regions suitable for implementation into routine diagnostics in sCRC patients would be mostly welcome. Other genes that are recurrently mutated in sCRC and that have been associated to tumor response to therapy and/or the outcome of sCRC patients such as the SMAD4 and PIK3CA genes present in ~10-35% and 10%-20% of cases, respectively [14–16], were not considered in the gene panel built for this study.

In the present study, we designed an NGS-based approach for simultaneous identification of hotspot KRAS, NRAS, BRAF and TP53 gene mutations in 87 primary sCRC tumors and evaluated the prognostic impact of the mutations identified in patient PFS and OS. Overall, the BRAF V600E mutation emerged as an independent prognostic factor for both PFS and OS, together with the TNM stage; based on the combined assessment of both parameters at diagnosis a prognostic score was built for stratification of sCRC patients into 3 risk groups with significantly different PFS rates.

RESULTS

Patient characteristics

Overall, 87 patients diagnosed with sCRC at the HUS (51 males and 36 females; median age of 72 years, ranging from 38 to 91 years) were studied. Median follow-up at the moment of closing the study was 19 months (range: 8 to 36 months). According to the histological type, 80 cases corresponded to adenocarcinomas, 4 cases to signet ring cell carcinomas, 2 cases to mucinous adenocarcinomas and one case to a neuroendocrine tumor. According to tumor grade, 24 cases were classified as well-differentiated tumors, 49 as moderately- and 7 as poorly-differentiated carcinomas. In all cases, histopathological grade was systematically confirmed in a second independent evaluation by another experienced pathologist. The most relevant clinical and laboratory data for each individual sCRC patient studied are summarized in Table 1.

Eighty-five patients underwent complete tumor resection (R0), while the remaining two cases showed positive tumor tissue margins to be affected (one patient showed microscopical and one macroscopical involvement). Adjuvant treatment was administered to 34/87 cases (39%) including: Xelox in 17/34 patients (50%), Capecitabine in 12 cases (35 %), Tomox in another 2 (6%), Tomudex in 1 case (3%), Utefox in another patient (3%) and cisplatin etoposide in the remaining case (3%).

NGS quality control

The custom NGS panel here designed, successfully amplified the 53 targeted amplicons, (covering the corresponding 5.3kb) with a mean depth of 2400x reads in all tumor samples analyzed (Supplementary Table 1). The reference sample, which included several mutations interrogated by our custom panel –e.g. KRAS (G12D and G13D), NRAS (Q61K) and BRAF (V600E)- was also successfully analyzed, the variant allele frequencies (VAF) detected by NGS being very similar to the expected VAF as estimated by the DNA input (Table 2). According to these results, our panel allows detection of variants at low VAF as confirmed by the detection of KRAS G12D variant (6% VAF).

Frequency and type of KRAS, NRAS, BRAF and TP53 mutations detected

Overall, mutations were detected in 46/87 cases analyzed (53%). The TP53 gene was the most frequently mutated gene (29/87 tumors; 33%), followed by the KRAS (mutated in 24/87 tumors; 28%) and BRAF (6/87 tumors; 7%) genes, while the NRAS gene was mutated in only one case (1%). The specific VAF for each individual mutation identified is listed in Supplementary Table 2. Regarding the individual hotspot mutations identified, KRAS G12D (8% of cases) and G12V (6%) corresponded to those showing the highest frequencies, followed by BRAF V600E (7%) and the TP53 R282W (3%) and R175H (3%) variants (Table 3). Thirty-two cases (37%) showed one single mutation (the most common being KRAS), while 14 cases (16%) had two genes affected. Coexisting TP53 and BRAF mutations were present in four cases (5%) whereas TP53 and KRAS were simultaneously mutated in 10 (11%) patients (Figure 1). As expected, no single case carried (simultaneously) KRAS and BRAF mutations (Figure 1).

Association between the mutational profile of sCRC tumors and other features of the disease

Once sCRC patients with and without mutations in the EGFR-pathway were compared, similar disease features were observed for the two patient groups, for most parameters analyzed including therapy (Table 4). Despite this, tumors with KRAS mutations were more frequently localized in the right colon (p=0.05), corresponded to well-differentiated tumors (p=0.04) and displayed no lymphovascular invasion (p=0.05). In contrast BRAF mutated sCRC tumors were significantly associated with a poorly- or moderately- differentiated histopathology (p=0.02), peritoneal carcinomatosis (p=0.006) and microsatellite instability (p=0.007). In turn, patients who displayed TP53 mutations showed a greater prevalence in males (p=0.05). No further statistically significant associations were found between the mutational status of KRAS, NRAS, BRAF and TP53 and patient age, the histological subtype of the tumor, tumor stage, presence of lymph node involvement and metastases, tumor size, CEA serum levels and the presence of perineural invasion. For these later variables that did not achieve a statistically significant association, a post-hoc power estimation was performed which showed limited power (<0.8) to detect potential associations due to small sample size (Supplementary Table 3).

From all mutations identified only the BRAF mutational status showed a statistically significant association with the microsatellite status with 2/48 microsatellite stable (MSS) patients (4%) and 3/6 microsatellite instable (MSI) cases (50%) being BRAF-mutated (p=0.007).

Prognostic impact of the KRAS, NRAS, BRAF and TP53 mutational status

In total, 74/87 patients were included in survival analyses. The other 13 patients were excluded because they had sCRC tumors other than adenocarcinoma (n=7), died within the first 30 days after surgery (n=4) and/or did not undergo complete resection of the tumor (n=2). From the prognostic point of view, the BRAF mutation was the only mutation that had an adverse impact on PFS (median 2-year PFS of 60% vs. 79% for wild type BRAF tumors; p=0.05) (Table 5) and OS (median 2-year OS of 80% vs. 94% for wild type BRAF; p=0.001). Other clinical, biological and histopahologic characteristics of the disease that displayed a significant adverse influence on PFS in the univariate analysis included: female gender (2-year PFS rates of 63% vs. 90% in males; p=0.03), advanced TNM stage (2-year PFS rates of 0% for stage IV vs. 75% for stage III and 91% for stages 0-II ; p<0.001), and the tumor grade of differentiation (2-year PFS rates of 95% for well-differentiated vs. 77% for moderately- and 33% for poorly-differentiated tumors; p=0.03), the presence of lymphovascular involvement at diagnosis (2-year PFS rates of 54% vs. 83% for cases who showed no lymphovascular invasion; p=0.004), and perineural invasion (2-year PFS rates of 62% vs. 83% for cases who had no perineural invasion; p=0.03) (Table 5). Multivariate analysis of prognostic factors showed that the TNM stage at diagnosis, together with the BRAF mutational status were the only independent variables predicting for PFS -HR=2.77, 95%CI of 1.55-4.96 (p=0.001) and HR=4.9, 95%CI of 1.04-23.75 (p=0.05) respectively- (Table 5).

Based on the above results, a prognostic score was established and applied to each patient, which was based on the TNM stage and the presence vs. absence of BRAF mutations. Thus, patients with an early TNM stage (stage 0, I or II) were considered low-risk patients independently of their BRAF mutational status (score 0), whereas cases with TNM stage III and wild-type BRAF were classified into the intermediate-risk group (score 1) and patients with TNM stage III and BRAF mutation and those with TNM stage IV (with or without BRAF mutation), were considered to be high-risk (score 2). As could be expected, significantly different PFS rates were observed for cases with score 0 vs. score 1 vs. score 2: 2-year PFS of 91% vs. 77% vs. 0% respectively (p<0.0001) (Figure 2). Once the BRAF mutational status was excluded from the multivariate analyses, only TNM stage maintained statistically significance as an independent factor. Thus, no other plausible combinations were considered. Regarding OS, the TNM stage at diagnosis (p=0.02), the size of the tumor (p=0.005), and the microsatellite instability status (p=0.01) together with the BRAF mutational status (p=0.001), were the only individual parameters that showed an impact on patient outcome, the latter being the only variable with an independent impact on patient OS -p=0.022; HR of 4.4 (95% CI of 0.7-28)- (Table 5). However, post-hoc power estimation revealed limited power to detect additional statistically significant differences on survival analyses due to limited sample sizes (Supplementary Table 4).

Validation of the prognostic score in an independent series of sCRC patients

In order to confirm the prognostic impact of the proposed score described above, we investigated its prognostic impact in an independent series of sCRC patients from the public GEO database (n=533). These patients were classified according to TNM stage and BRAF mutational status at diagnosis into the three previously mentioned subgroups: score 0 (n=281), score 1 (n=130) and score 2 (n=80). PFS rates were very similar for cases with score 0 vs. score 1 vs. score 2: 2-year PFS of 90% vs. 75% vs. 48% respectively (p<0.0001) (Supplementary Figure 1). These results support previous findings in our dataset.

DISCUSSION

General consensus exists about the need to simultaneously assess multiple molecular markers for more accurate classification and personalized treatment of sCRC patients. In this study, we designed an NGS panel to investigate the frequency of KRAS, NRAS, BRAF and TP53 mutations and applied it to a series of 87 sCRC patients, to determine the potential impact of the mutations detected in patient outcome, and their association with the clinical and pathological features of the disease. Overall, our results showed a good performance of the NGS panel (and method) used and the feasibility of its use in routine diagnostics. Overall, the frequency and type of the distinct mutations identified was generally concordant with previous data from the literature [17–19]. In addition, we observed that from the 4 genes investigated; only the BRAF V600E mutation alone or together with the TNM stage at diagnosis had an independent prognostic impact both for patient OS and PFS respectively. Although other histopathological (e.g. tumor grade, perineural and lymphovascular invasion), clinical (e.g. age) and genetic (e.g the microsatellite status) features of the disease also showed to be of potential prognostic value when individually considered, they lost their significance in the multivariate analysis.

The frequency and significance of KRAS mutations have been thoroughly investigated in sCRC patients, particularly in the context of anti-EGFR therapies [17, 20, 21]. Despite this, the prognostic impact of KRAS mutations in sCRC patients remains controversial. Thus, while some studies have found that exon 2 (codon 13) mutations predict for a worse prognosis including higher recurrence and shorter survival rates,[22] others highlighted the potential value of KRAS mutations in exons 3 and 4 (codons 61, 146 and 147) rather than in exon 2, as a predictive marker for PFS [23, 24] and OS [24, 25], and Palomba et al, did not find a significant prognostic impact for mutated KRAS in a large series of 1,284 CRC patients [26]. These apparently controversial results on the potential prognostic significance of KRAS mutations in sCRC, are probably due to the great genetic heterogeneity of this type of tumors, the effect of additional genetic markers and potentially also, differences in the series of patients analyzed [27]. In this regard, several studies have shown discrepancies in the KRAS mutational status of different tumoral samples from the same patient, and among distinct techniques. For example, Li et al, found a discordance rate of 19% in the frequency of KRAS mutations once studied by quantitative real-time PCR vs. MassARRAY(®) techniques, between primary tumors and their paired metastatic lesions in Chinese patients with sCRC [28]. In the present study, we found KRAS mutations in around one third of the patients, particularly among well differentiated tumors, localized in the right colon, in the absence of lymphovascular invasion, in line with previous observations [29–32]. However, in our series, KRAS mutated sCRC patients showed similar PFS and OS rates to wild-type KRAS cases.

Similarly, while some previous studies have reported an association among NRAS-mutated patients and a shorter OS [33], we could not confirm these findings in our series, due to the fact that only 1/87 patients investigated here showed NRAS mutation, in line with previous studies [11]. Therefore, further studies in larger series of sCRC patients are required to establish the prognostic impact of NRAS mutations.

As expected, the overall frequency of BRAF mutations detected in our series was lower than that of KRAS [23, 26]. Interestingly, patients who carried BRAF mutations more frequently had poor-prognosis disease features such as poor-differentiated tumors [34], presence of peritoneal carcinomatosis [31, 35] and microsatellite instability, as also reported by others [32, 36]. In addition and in line with previous studies, half of the patients in our series with BRAF mutation were found in stage III [37]. Despite this, the BRAF mutation showed an independent adverse prognostic impact on both PFS and OS, as also pointed out previously by others [4, 23, 38, 39]. Further prospective studies in larger series of patients are required to confirm these observations.

TP53 gene mutations were detected in one third of our cases, a frequency that is in the lower range of that described previously by others [40, 41]. In line with previous observations, TP53 mutations were associated with female gender [42], but showed no impact on patient PFS and OS. In this regard, previous studies on the potential association between TP53 mutations and the prognosis of sCRC have yielded inconsistent results [43, 44] This might be due to insufficient statistical power (in our and also other series) to detect modest survival differences between wild-type and mutated TP53 patients, the need for longer follow-up and the potential influence of adjuvant therapy. On top of this, it should be noted that the non-mutated TP53 allele of the gene could be functional and counteract the mutated phenotype; thus, further studies are required to investigate the potential impact of TP53 genetic alterations involving the other TP53 gene -e.g. del(17p)- [45].

As far as the slightly lower mutation frequency detected in our series, particularly KRAS, and since the level of sensitivity achieved is good, we consider it could be attributed to intratumoral heterogeneity (ITH). Tumor biopsy represents a limited fraction of the tumoral clones, simply due to spatial ITH [46].

As expected, several clinical and biological variables showed a significant association with the outcome of sCRC patients. Thus, female gender, advanced disease stage at diagnosis, poorly-differentiated tumors and the presence of lymphovascular/perineural invasion at diagnostic surgery, together with the BRAF mutational status, were all associated with an adverse impact on PFS in the univariate analysis, in line with previous studies [45, 47–49]. In contrast, only the TNM stage at diagnosis, the tumor size, microsatellite instability and the BRAF status showed a prognostic impact on OS (in the univariate analysis) among our cases. Multivariate analysis of prognostic factors showed that the best combination of independent variables for predicting PFS in sCRC patients were the presence of BRAF mutations and an advanced TNM stage at diagnosis, the former also retaining its independent prognostic value for OS. Based on these results, we built an original scoring system that allowed stratification of the sCRC patients analysed into three different risk groups with significantly different PFS rates at 2 years. Validation of this score in an independent dataset further strengthens this evidence. Of note, in this model, the BRAF mutational status specifically contributed to the sub-stratification of TNM stage III patients into intermediate vs. high-risk cases. Despite the fact that an association has been reported between the mutational status of genes of the EGFR signaling pathway and the prognosis of sCRC patients [38, 50–53], to the best of our knowledge, this is the first time that a scoring system based on combined assessment of the TNM stage at diagnosis and the BRAF mutational status is proposed, for the identification of sCRC patients undergoing complete tumor resection who are still at high risk of recurrence of the disease in the first 2-years after diagnosis (score 2). If the prognostic value of this new risk stratification model is confirmed in prospective series of sCRC patients it might contribute to pave the way for trials evaluating BRAF-targeted therapies in this specific subgroup of sCRC cases.

In summary, here we confirm the adverse prognosis of BRAF mutations in sCRC, and point out their utility, together with the TNM stage, for the identification, already at diagnosis, of a subgroup of sCRC patients (TNM stage IV plus TNM stage III and BRAF mutated patients) who, despite following complete resection of the tumor, still retain a high-risk of recurrence during the first two years after diagnosis. Additional prospective studies are required to confirm the utility of the proposed predictive model in larger series of homogeneously treated sCRC patients.

MATERIALS AND METHODS

Patients and samples

Freshly-frozen sCRC tissues were obtained from primary tumors of 87 Caucasian patients diagnosed with sCRC, and classified according to the World Health Organization (WHO) criteria [54], after each patient gave his/her informed consent to participate in this study. All patients underwent surgical resection of primary tumor tissues at the Department of Surgery of the University Hospital of Salamanca (HUS; Salamanca, Spain) before they had received any treatment.

Histopathological diagnosis was established by an experienced pathologist, that ensured the selection of sections representative of the tumor tissue with >70% tumor cell infiltration for further genetic analyses. DNA was extracted and isolated from freshly-frozen primary sCRC tumor tissues, using a Maxwell® 16 System for Genomic DNA Extraction (Promega, Mannheim, Germany) and quantified using a Qubit dsDNA BR assay (Life Technologies, Carlsbad CA). A positive control sample -Quantitative Multiplex DNA reference standard (Horizon Discovery, Cambridge, UK)- was analyzed in parallel to each set of samples, for validation of the custom panel designed, and evaluation of the specific sensitivity of each variant call.

The study was approved by the local ethics committee of the HUS.

Custom amplicon panel design

A custom amplicon panel for NGS analysis of the hotspot regions of KRAS/NRAS (exons 2, 3 and 4), and BRAF (exon 15), was designed; additionally, the whole TP53 coding regions were also included in the panel for a total of 53 amplicons and 5,300bp, with an estimated coverage per sample of 7,000x.

Preparation of DNA libraries

DNA libraries were prepared with the Truseq Custom Amplicon Panel (Illumina, San Diego, CA), according to the manufacturer´s protocol. Briefly, 50-250 ng of gDNA in 10 μl water was hybridized with a pool of oligonucleotides. Then, the unbound oligonucleotides were removed, and extension-ligation of the bound oligonucleotides was followed by PCR amplification. PCR products were cleaned and their quality checked using the 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA). A minimum size of the PCR product of ~275 bp was required. Then, the DNA libraries were then quantified using the Qubit dsDNA HS Assay Kit and the Qubit 2.0 fluorometer (Life Technologies). Subsequently, each DNA library was diluted to a concentration of 4 nM and pooled with the other libraries in aliquots of equal volumes. The amplicon DNA libraries were paired-end sequenced using a MiSeq (Illumina) instrument.

Analysis of DNA sequences

The sequence data generated were pre-processed with the MiSeq Reporter (MiSeq integrated software, Illumina), which uses a Burrows-Wheeler Aligner (BWA)[55] and the Genome Analysis Tool Kit (GATK) [56] for variant calling of single-nucleotide polymorphisms (SNPs) and short insertions and deletions (InDels). The identified variants were exported in the VCF data file format for further analysis using the sequencer-accompanied software (Variant Studio, Illumina) and the Integrative Genomics Viewer (IGV) software (Broad Institute, Cambridge, MA). We employed the Illumina Variant studio population frequency filters (based on the www.1000genomes.org database) to exclude variants with an overall minor allele frequency greater than 1.0%, considered common SNPs, taking as reference the Caucasian population. The following criteria were used to define and report a variant: minimum coverage of 100x, minimum variant frequency of 5%, confirmed by visual inspection using the Integrative Genomics Viewer 16 software. Detailed visual inspection with the IGV software was performed to confirm the presence and the read depth of the amplicons. Variant allele frequency (VAF) was established based on the number of reads called for the altered allele and the total number of reads called at that position by the Variant Studio Software (Illumina).

External validation of the proposed score

A prognostic score based on the TNM stage and BRAF mutational status is established. External validation of the proposed score was performed in a previously reported group of sCRC patients from which TNM stage and BRAF mutational status at diagnosis and follow-up data were publicly available at the GEO database (accession number GSE39582) [57]. Patients included in this external validation group showed mutated BRAF in 48/533 (9%) and were classified according to the TNM staging system as follows: 38 patients had stage I (7.1%), 248 had stage II (46.5%), 186 had stage III (34.9%) and 61 had stage IV (11.4%).

Statistical analyses

For all continuous variables, median and mean values and their standard deviation (SD) and range were calculated using the SPSS software package (SPSS 22.0 Inc, Chicago, IL); for dichotomic variables, frequencies were reported. In order to evaluate the statistical significance of differences observed between groups of mutated vs. non mutated patients, the Student’s T and the Mann-Whitney U tests were used for continuous variables, depending on whether they displayed or not a normal distribution, respectively. For qualitative variables, the X² test was applied (cross-tab; SPSS).

For survival analyses, patients i) who had sCRC tumors other than adenocarcinoma (n=7), ii) those who dies within the first 30 days after surgery (n=4), and/or iii) did not undergo R0 resection (n=2), were excluded from the study. PFS and OS curves were plotted for the remaining 74 patients according to the method of Kaplan and Meier, and the one-sided log-rank test (one-sided) was used to establish the statistical significance of differences observed between survival curves (survival; SPSS). Multivariate analysis of prognostic factors for PFS and OS was performed using the Cox stepwise regression (forward selection) model (regression, SPSS). For multivariate analysis, only those variables showing a statistically significant association with PFS or OS in the univariate analysis, were considered. Due to the limited sample size, post-hoc power estimation was performed for all association and survival analyses and a power estimation of <0.8 was considered to be inadequate. Statistical significance was set at p values < 0.05.

CONFLICTS OF INTEREST

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

FUNDING

This work has been partially supported by grants from the Instituto de Salud Carlos III (ISCIII; Ministerio de Sanidad y Consumo, Madrid, Spain) (PI12/02053-FIS), Gerencia Regional de Salud de Castilla y León, Valladolid, Spain (GRS1302/A/16), Consejería de Sanidad (Junta de Castilla y León, Valladolid, Spain) (BIO/SA46/14, BIO/SA02/13), RTICC from the ISCIII (RD12/0020/0035-FEDER, RD12/0036/0048-FEDER) and CIBERONC (CB16/12/00400 and CB16/12/00233), Fundación Memoria de Don Samuel Solórzano Barruso, (Salamanca, Spain) and Fundación Eugenio Rodríguez Pascual, (Madrid, Spain). JM Sayagués and ME Sarasquete are supported by grants (CES11/004 and CP13/00080) from the ISCIII, Ministerio de Ciencia e Innovación, Madrid, Spain.

REFERENCES

1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer J Clin. 2015; 65:5–29. https://doi.org/10.3322/caac.21254.

2. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer. 2015; 136:E359–86. https://doi.org/10.1002/ijc.29210.

3. Peeters M, Douillard JY, Van Cutsem E, Siena S, Zhang K, Williams R, Wiezorek J. Mutant KRAS codon 12 and 13 alleles in patients with metastatic colorectal cancer: assessment as prognostic and predictive biomarkers of response to panitumumab. J Clin Oncol. 2013; 31:759–65. https://doi.org/10.1200/JCO.2012.45.1492.

4. Modest DP, Ricard I, Heinemann V, Hegewisch-Becker S, Schmiegel W, Porschen R, Stintzing S, Graeven U, Arnold D, von Weikersthal LF, Giessen-Jung C, Stahler A, Schmoll HJ, et al. Outcome according to KRAS-, NRAS- and BRAF-mutation as well as KRAS mutation variants: pooled analysis of five randomized trials in metastatic colorectal cancer by the AIO colorectal cancer study group. Ann Oncol. 2016; 27:1746–53. https://doi.org/10.1093/annonc/mdw261.

5. Shen Y, Han X, Wang J, Wang S, Yang H, Lu SH, Shi Y. Prognostic impact of mutation profiling in patients with stage II and III colon cancer. Sci Rep. 2016; 6:24310. https://doi.org/10.1038/srep24310.

6. Lo Nigro C, Ricci V, Vivenza D, Granetto C, Fabozzi T, Miraglio E, Merlano MC. Prognostic and predictive biomarkers in metastatic colorectal cancer anti-EGFR therapy. World J Gastroenterol. 2016; 22:6944–54. https://doi.org/10.3748/wjg.v22.i30.6944.

7. De Roock W, De Vriendt V, Normanno N, Ciardiello F, Tejpar S. KRAS, BRAF, PIK3CA, and PTEN mutations: implications for targeted therapies in metastatic colorectal cancer. Lancet Oncol. 2011; 12:594–603. https://doi.org/10.1016/S1470-2045(10)70209-6.

8. Safaee Ardekani G, Jafarnejad SM, Tan L, Saeedi A, Li G. The prognostic value of BRAF mutation in colorectal cancer and melanoma: a systematic review and meta-analysis. PLoS One. 2012; 7:e47054. https://doi.org/10.1371/journal.pone.0047054.

9. Tol J, Nagtegaal ID, Punt CJ. BRAF mutation in metastatic colorectal cancer. N Engl J Med. 2009; 361:98–99. https://doi.org/10.1056/NEJMc0904160.

10. Tran B, Kopetz S, Tie J, Gibbs P, Jiang ZQ, Lieu CH, Agarwal A, Maru DM, Sieber O, Desai J. Impact of BRAF mutation and microsatellite instability on the pattern of metastatic spread and prognosis in metastatic colorectal cancer. Cancer. 2011; 117:4623–32. https://doi.org/10.1002/cncr.26086.

11. Irahara N, Baba Y, Nosho K, Shima K, Yan L, Dias-Santagata D, Iafrate AJ, Fuchs CS, Haigis KM, Ogino S. NRAS mutations are rare in colorectal cancer. Diagn Mol Pathol. 2010; 19:157–63. https://doi.org/10.1097/PDM.0b013e3181c93fd1.

12. De Roock W, Claes B, Bernasconi D, De Schutter J, Biesmans B, Fountzilas G, Kalogeras KT, Kotoula V, Papamichael D, Laurent-Puig P, Penault-Llorca F, Rougier P, Vincenzi B, 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; 11:753–62. https://doi.org/10.1016/S1470-2045(10)70130-3.

13. Malapelle U, Vigliar E, Sgariglia R, Bellevicine C, Colarossi L, Vitale D, Pallante P, Troncone G. Ion Torrent next-generation sequencing for routine identification of clinically relevant mutations in colorectal cancer patients. J Clin Pathol. 2015; 68:64–68. https://doi.org/10.1136/jclinpath-2014-202691.

14. Mehrvarz Sarshekeh A, Advani S, Overman MJ, Manyam G, Kee BK, Fogelman DR, Dasari A, Raghav K, Vilar E, Manuel S, Shureiqi I, Wolff RA, Patel KP, et al. Association of SMAD4 mutation with patient demographics, tumor characteristics, and clinical outcomes in colorectal cancer. PLoS One. 2017; 12:e0173345. https://doi.org/10.1371/journal.pone.0173345.

15. Sartore-Bianchi A, Martini M, Molinari F, Veronese S, Nichelatti M, Artale S, Di Nicolantonio F, Saletti P, De Dosso S, Mazzucchelli L, Frattini M, Siena S, Bardelli A. PIK3CA mutations in colorectal cancer are associated with clinical resistance to EGFR-targeted monoclonal antibodies. Cancer Res. 2009; 69:1851–57. https://doi.org/10.1158/0008-5472.CAN-08-2466.

16. Prenen H, De Schutter J, Jacobs B, De Roock W, Biesmans B, Claes B, Lambrechts D, Van Cutsem E, Tejpar S. PIK3CA mutations are not a major determinant of resistance to the epidermal growth factor receptor inhibitor cetuximab in metastatic colorectal cancer. Clin Cancer Res. 2009; 15:3184–88. https://doi.org/10.1158/1078-0432.CCR-08-2961.

17. Yu J, Wu WK, Li X, He J, Li XX, Ng SS, Yu C, Gao Z, Yang J, Li M, Wang Q, Liang Q, Pan Y, et al. Novel recurrently mutated genes and a prognostic mutation signature in colorectal cancer. Gut. 2015; 64:636–45. https://doi.org/10.1136/gutjnl-2013-306620.

18. Deschoolmeester V, Boeckx C, Baay M, Weyler J, Wuyts W, Van Marck E, Peeters M, Lardon F, Vermorken JB. KRAS mutation detection and prognostic potential in sporadic colorectal cancer using high-resolution melting analysis. Br J Cancer. 2010; 103:1627–36. https://doi.org/10.1038/sj.bjc.6605959.

19. Zlobec I, Kovac M, Erzberger P, Molinari F, Bihl MP, Rufle A, Foerster A, Frattini M, Terracciano L, Heinimann K, Lugli A. Combined analysis of specific KRAS mutation, BRAF and microsatellite instability identifies prognostic subgroups of sporadic and hereditary colorectal cancer. Int J Cancer. 2010; 127:2569–75. https://doi.org/10.1002/ijc.25265.

20. Dallol A, Buhmeida A, Al-Ahwal MS, Al-Maghrabi J, Bajouh O, Al-Khayyat S, Alam R, Abusanad A, Turki R, Elaimi A, Alhadrami HA, Abuzenadah M, Banni H, et al. Clinical significance of frequent somatic mutations detected by high-throughput targeted sequencing in archived colorectal cancer samples. J Transl Med. 2016; 14:118. https://doi.org/10.1186/s12967-016-0878-9.

21. Al-Shamsi HO, Jones J, Fahmawi Y, Dahbour I, Tabash A, Abdel-Wahab R, Abousamra AO, Shaw KR, Xiao L, Hassan MM, Kipp BR, Kopetz S, Soliman AS, et al. Molecular spectrum of KRAS, NRAS, BRAF, PIK3CA, TP53, and APC somatic gene mutations in Arab patients with colorectal cancer: determination of frequency and distribution pattern. J Gastrointest Oncol. 2016; 7:882–902. https://doi.org/10.21037/jgo.2016.11.02.

22. Aldiab A, Al Khayal KA, Al Obaid OA, Alsheikh A, Alsaleh K, Shahid M, Alkharji H. Clinicopathological Features and Predictive Factors for Colorectal Cancer Outcome in the Kingdom of Saudi Arabia. Oncology. 2017; 92:75–86. https://doi.org/10.1159/000450857.

23. Therkildsen C, Bergmann TK, Henrichsen-Schnack T, Ladelund S, Nilbert M. The predictive value of KRAS, NRAS, BRAF, PIK3CA and PTEN for anti-EGFR treatment in metastatic colorectal cancer: A systematic review and meta-analysis. Acta Oncol. 2014; 53:852–64. https://doi.org/10.3109/0284186X.2014.895036.

24. Loupakis F, Pollina L, Stasi I, Ruzzo A, Scartozzi M, Santini D, Masi G, Graziano F, Cremolini C, Rulli E, Canestrari E, Funel N, Schiavon G, et al. PTEN expression and KRAS mutations on primary tumors and metastases in the prediction of benefit from cetuximab plus irinotecan for patients with metastatic colorectal cancer. J Clin Oncol. 2009; 27:2622–29. https://doi.org/10.1200/JCO.2008.20.2796.

25. Douillard JY, Oliner KS, Siena S, Tabernero J, Burkes R, Barugel M, Humblet Y, Bodoky G, Cunningham D, Jassem J, Rivera F, Kocákova I, Ruff P, et al. Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. N Engl J Med. 2013; 369:1023–34. https://doi.org/10.1056/NEJMoa1305275.

26. Palomba G, Doneddu V, Cossu A, Paliogiannis P, Manca A, Casula M, Colombino M, Lanzillo A, Defraia E, Pazzola A, Sanna G, Putzu C, Ortu S, et al. Prognostic impact of KRAS, NRAS, BRAF, and PIK3CA mutations in primary colorectal carcinomas: a population-based study. J Transl Med. 2016; 14:292. https://doi.org/10.1186/s12967-016-1053-z.

27. Foltran L, De Maglio G, Pella N, Ermacora P, Aprile G, Masiero E, Giovannoni M, Iaiza E, Cardellino GG, Lutrino SE, Mazzer M, Giangreco M, Pisa FE, et al. Prognostic role of KRAS, NRAS, BRAF and PIK3CA mutations in advanced colorectal cancer. Future Oncol. 2015; 11:629–40. https://doi.org/10.2217/fon.14.279.

28. Li ZZ, Bai L, Wang F, Zhang ZC, Wang F, Zeng ZL, Zeng JB, Zhang DS, Wang FH, Wang ZQ, Li YH, Shao JY, Xu RH. Comparison of KRAS mutation status between primary tumor and metastasis in Chinese colorectal cancer patients. Med Oncol. 2016; 33:71. https://doi.org/10.1007/s12032-016-0787-z.

29. Kawabata Y, Tomita N, Monden T, Ohue M, Ohnishi T, Sasaki M, Sekimoto M, Sakita I, Tamaki Y, Takahashi J, Yagyu T, Mishima H, Kikkawa N, Monden M. Molecular characteristics of poorly differentiated adenocarcinoma and signet-ring-cell carcinoma of colorectum. Int J Cancer. 1999; 84:33–38. https://doi.org/10.1002/(SICI)1097-0215(19990219)84:13.0.CO;2-Z.

30. Summers MG, Smith CG, Maughan TS, Kaplan R, Escott-Price V, Cheadle JP. BRAF and NRAS Locus-Specific Variants Have Different Outcomes on Survival to Colorectal Cancer. Clin Cancer Res. 2017; 23:2742–49. https://doi.org/10.1158/1078-0432.CCR-16-1541.

31. Li ZZ, Wang F, Zhang ZC, Wang F, Zhao Q, Zhang DS, Wang FH, Wang ZQ, Luo HY, He MM, Wang DS, Jin Y, Ren C, et al. Mutation profiling in chinese patients with metastatic colorectal cancer and its correlation with clinicopathological features and anti-EGFR treatment response. Oncotarget. 2016; 7:28356–68. https://doi.org/10.18632/oncotarget.8541.

32. Eklöf V, Wikberg ML, Edin S, Dahlin AM, Jonsson BA, Öberg Å, Rutegård J, Palmqvist R. The prognostic role of KRAS, BRAF, PIK3CA and PTEN in colorectal cancer. Br J Cancer. 2013; 108:2153–63. https://doi.org/10.1038/bjc.2013.212.

33. Schirripa M, Cremolini C, Loupakis F, Morvillo M, Bergamo F, Zoratto F, Salvatore L, Antoniotti C, Marmorino F, Sensi E, Lupi C, Fontanini G, De Gregorio V, et al. Role of NRAS mutations as prognostic and predictive markers in metastatic colorectal cancer. Int J Cancer. 2015; 136:83–90. https://doi.org/10.1002/ijc.28955.

34. Olevian DC, Nikiforova MN, Chiosea S, Sun W, Bahary N, Kuan SF, Pai RK. Colorectal poorly differentiated neuroendocrine carcinomas frequently exhibit BRAF mutations and are associated with poor overall survival. Hum Pathol. 2016; 49:124–34. https://doi.org/10.1016/j.humpath.2015.11.004.

35. Yaeger R, Cercek A, Chou JF, Sylvester BE, Kemeny NE, Hechtman JF, Ladanyi M, Rosen N, Weiser MR, Capanu M, Solit DB, D’Angelica MI, Vakiani E, Saltz LB. BRAF mutation predicts for poor outcomes after metastasectomy in patients with metastatic colorectal cancer. Cancer. 2014; 120:2316–24. https://doi.org/10.1002/cncr.28729.

36. Nakaji Y, Oki E, Nakanishi R, Ando K, Sugiyama M, Nakashima Y, Yamashita N, Saeki H, Oda Y, Maehara Y. Prognostic value of BRAF V600E mutation and microsatellite instability in Japanese patients with sporadic colorectal cancer. J Cancer Res Clin Oncol. 2017; 143:151–60. https://doi.org/10.1007/s00432-016-2275-4.

37. Ogino S, Shima K, Meyerhardt JA, McCleary NJ, Ng K, Hollis D, Saltz LB, Mayer RJ, Schaefer P, Whittom R, Hantel A, Benson AB 3rd, Spiegelman D, et al. Predictive and prognostic roles of BRAF mutation in stage III colon cancer: results from intergroup trial CALGB 89803. Clin Cancer Res. 2012; 18:890–900. https://doi.org/10.1158/1078-0432.CCR-11-2246.

38. Seligmann JF, Fisher D, Smith CG, Richman SD, Elliott F, Brown S, Adams R, Maughan T, Quirke P, Cheadle J, Seymour M, Middleton G. Investigating the poor outcomes of BRAF-mutant advanced colorectal cancer: analysis from 2530 patients in randomised clinical trials. Ann Oncol. 2017; 28:562–68. https://doi.org/10.1093/annonc/mdw645.

39. Rosty C, Young JP, Walsh MD, Clendenning M, Walters RJ, Pearson S, Pavluk E, Nagler B, Pakenas D, Jass JR, Jenkins MA, Win AK, Southey MC, et al. Colorectal carcinomas with KRAS mutation are associated with distinctive morphological and molecular features. Mod Pathol. 2013; 26:825–34. https://doi.org/10.1038/modpathol.2012.240.

40. Chang SC, Lin PC, Lin JK, Lin CH, Yang SH, Liang WY, Chen WS, Jiang JK. Mutation Spectra of Common Cancer-Associated Genes in Different Phenotypes of Colorectal Carcinoma Without Distant Metastasis. Ann Surg Oncol. 2016; 23:849–55. https://doi.org/10.1245/s10434-015-4899-z.

41. Calistri D, Rengucci C, Seymour I, Leonardi E, Truini M, Malacarne D, Castagnola P, Giaretti W. KRAS, p53 and BRAF gene mutations and aneuploidy in sporadic colorectal cancer progression. Cell Oncol. 2006; 28:161–66.

42. Cai ZX, Tang XD, Gao HL, Tang C, Nandakumar V, Jones L, Ye H, Lou F, Zhang D, Sun H, Dong H, Zhang G, Liu Z, et al. APC, FBXW7, KRAS, PIK3CA, and TP53 gene mutation in human Colorectal Cancer Tumor Frequenly Detected by Next-Generation DNA Sequencing. J Mol Genet Med. 2014; 8. https://doi.org/10.4172/1747-0862.1000145.

43. Munro AJ, Lain S, Lane DP. P53 abnormalities and outcomes in colorectal cancer: a systematic review. Br J Cancer. 2005; 92:434–44. https://doi.org/10.1038/sj.bjc.6602358.

44. Tang R, Wang JY, Fan CW, Tsao KC, Chen HH, Wu CM, Chen JS, Changchien CR, Hsieh LL. p53 is an independent pre-treatment markers for long-term survival in stage II and III colorectal cancers: an analysis of interaction between genetic markers and fluorouracil-based adjuvant therapy. Cancer Lett. 2004; 210:101–09. https://doi.org/10.1016/j.canlet.2004.01.006.

45. González-González M, Muñoz-Bellvis L, Mackintosh C, Fontanillo C, Gutiérrez ML, Abad MM, Bengoechea O, Teodosio C, Fonseca E, Fuentes M, De Las Rivas J, Orfao A, Sayagués JM. Prognostic Impact of del(17p) and del(22q) as assessed by interphase FISH in sporadic colorectal carcinomas. PLoS One. 2012; 7:e42683. https://doi.org/10.1371/journal.pone.0042683.

46. Árnadóttir SS, Jeppesen M, Lamy P, Bramsen JB, Nordentoft I, Knudsen M, Vang S, Madsen MR, Thastrup O, Thastrup J, L Andersen C. Characterization of genetic intratumor heterogeneity in colorectal cancer and matching patient-derived spheroid cultures. Mol Oncol. 2018; 12:132–47. https://doi.org/10.1002/1878-0261.12156.

47. Zhou Y, Wang H, Gong H, Cao M, Zhang G, Wang Y. Clinical significance of perineural invasion in stages II and III colorectal cancer. Pathol Res Pract. 2015; 211:839–44. https://doi.org/10.1016/j.prp.2015.09.001.

48. Das V, Kalita J, Pal M. Predictive and prognostic biomarkers in colorectal cancer: A systematic review of recent advances and challenges. Biomed Pharmacother. 2017; 87:8–19. https://doi.org/10.1016/j.biopha.2016.12.064.

49. Dienstmann R, Mason MJ, Sinicrope FA, Phipps AI, Tejpar S, Nesbakken A, Danielsen SA, Sveen A, Buchanan DD, Clendenning M, Rosty C, Bot B, Alberts SR, et al. Prediction of overall survival in stage II and III colon cancer beyond TNM system: a retrospective, pooled biomarker study. Ann Oncol. 2017; 28:1023–31. https://doi.org/10.1093/annonc/mdx052.

50. Ålgars A, Sundström J, Lintunen M, Jokilehto T, Kytölä S, Kaare M, Vainionpää R, Orpana A, Österlund P, Ristimäki A, Carpen O, Ristamäki R. EGFR gene copy number predicts response to anti-EGFR treatment in RAS wild type and RAS/BRAF/PIK3CA wild type metastatic colorectal cancer. Int J Cancer. 2017; 140:922–29. https://doi.org/10.1002/ijc.30507.

51. Sugai T, Eizuka M, Takahashi Y, Fukagawa T, Habano W, Yamamoto E, Akasaka R, Otuska K, Matsumoto T, Suzuki H. Molecular subtypes of colorectal cancers determined by PCR-based analysis. Cancer Sci. 2017; 108:427–34. https://doi.org/10.1111/cas.13164.

52. Nakayama I, Shinozaki E, Matsushima T, Wakatsuki T, Ogura M, Ichimura T, Ozaka M, Takahari D, Suenaga M, Chin K, Mizunuma N, Yamaguchi K. Retrospective study of RAS/PIK3CA/BRAF tumor mutations as predictors of response to first-line chemotherapy with bevacizumab in metastatic colorectal cancer patients. BMC Cancer. 2017; 17:38. https://doi.org/10.1186/s12885-016-2994-6.

53. Barras D, Missiaglia E, Wirapati P, Sieber OM, Jorissen RN, Love C, Molloy PL, Jones IT, McLaughlin S, Gibbs P, Guinney J, Simon IM, Roth AD, et al. BRAF V600E Mutant Colorectal Cancer Subtypes Based on Gene Expression. Clin Cancer Res. 2017; 23:104–15. https://doi.org/10.1158/1078-0432.CCR-16-0140.

54. Hamilton SR. ALA. World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of the Digestive System. IARC Press:Lyon. 2000; 2000:1–25.

55. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics. 2009; 25:1754–60. https://doi.org/10.1093/bioinformatics/btp324.

56. McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, Garimella K, Altshuler D, Gabriel S, Daly M, DePristo MA. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010; 20:1297–303. https://doi.org/10.1101/gr.107524.110.

57. Marisa L, de Reyniès A, Duval A, Selves J, Gaub MP, Vescovo L, Etienne-Grimaldi MC, Schiappa R, Guenot D, Ayadi M, Kirzin S, Chazal M, Fléjou JF, et al. Gene expression classification of colon cancer into molecular subtypes: characterization, validation, and prognostic value. PLoS Med. 2013; 10:e1001453. https://doi.org/10.1371/journal.pmed.1001453.