Colorectal cancer (CRC) is the most common gastrointestinal cancer and one of the leading causes of cancer-related deaths worldwide. Various biomarkers have been identified for chemotherapy in advanced CRC. Particularly, KRAS/BRAF mutation status and microsatellite instability (MSI) status are known to be effective as predictive biomarkers. One of the important signalling pathways in CRC, activation of the RAS-RAF-MAPK pathway, which consists of KRAS/BRAF, is known (1). The pathway lies downstream from the epidermal growth factor receptor (EGFR), a transmembrane protein receptor, and contributes to cell proliferation, survival, growth, apoptosis resistance, invasion and migration (2,3). EGFR is overexpressed in most CRCs and antibodies against it inhibit stimulation of several intracellular signalling pathways, such as RAS-RAF-MAPK pathways (4). However, previous studies have shown that KRAS-mutant CRC is resistant to EGFR antibodies (5,6). KRAS mutation occurs in approximately 40% of CRC cases (6). Therefore, analysis of KRAS mutations is important for the selection of anti-EGFR therapy, and it is necessary before treatment in advanced CRC. In addition, CRC with wild-type KRAS is not always sensitive to EGFR antibodies and BRAF-mutant CRC has a poor prognosis (7). It is suggested that the efficiency of EGFR antibodies is further restricted to CRC, with both KRAS/BRAF wild-types.

BRAF, a member of the RAF family of serin/threonine kinases, is directly downstream from KRAS. BRAF mutations lead to constitutive activation of a MAPK pathway. KRAS/BRAF mutations are considered to be mutually exclusive. BRAF mutations are present in approximately 6% advanced CRC cases (5,7–9). Patients with BRAF-mutant advanced CRC are more likely to be older, of the female gender, have right-sided primary tumours and show an unusual pattern of metastatic spread, including frequent peritoneal and distant lymph node involvement. BRAF-mutant advanced CRC has proven to be a poor prognosis (5,7,9). The BRAF inhibitor vemurafenib as well as dabrafenib, have resulted in significantly prolonged progression-free survival and overall survival in patients with BRAF-mutated advanced melanoma (10,11). However, in contrast to BRAF-mutant melanoma, BRAF-mutant advanced CRC has shown a lack of sensitivity to BRAF inhibitor monotherapy in previous clinical trials (12). Nevertheless, FOLFOXIRI + bevacizumab and BRAF inhibitor + MEK or EGFR inhibitors, might be a reasonable therapy for BRAF-mutant advanced CRC (13–16). BRAF is a good biomarker, not only for a poor prognosis but also for the selection of molecular-targeted therapy.

MSI is a genetic change caused by a deficiency in the mismatch repair (MMR) system. The MMR system detects and repairs the mismatches that occur during DNA replication. It has been reported that approximately 15% of CRC cases show MSI in western countries, and approximately 6% of CRC cases in Asian countries (9,17). Recently, advanced CRC with MSI-H have been shown to have a high response rate to programmed death-1 (PD-1) inhibitor therapy, namely an immune checkpoint inhibitor (18). MSI status may be a helpful biomarker for immune therapy.

Based on the above, evaluating KRAS/BRAF mutation status and MSI status may be important to choose the regimen and predict the prognosis for advanced CRC. However, acquiring the various mutations during the CRC progression causes cancer-cell heterogeneity. The prevalence of intratumoural genetic heterogeneity was investigated in the cases of resistance to cancer therapy in previous studies, and the resistance to therapy may be explained by the presence of intratumoural heterogeneity (19). Evaluation of whether KRAS/BRAF mutation status and MSI status could change during the progression of metastatic disease might be useful to decide appropriate treatment for advanced CRC. KRAS mutation is recognized as an early event in colorectal carcinogenesis (20,21). Therefore, concordance of KRAS mutation status between primary CRC and corresponding metastases should be expected, and previous studies demonstrated high concordance rate (22–24). Nevertheless, some other studies reported discordance of KRAS mutation status between primary CRC and corresponding metastases. Therefore, there is still conflict about its concordance. Besides, concordance of BRAF mutation status and MSI status between primary CRC and corresponding metastases, is still unclear because of the small number of advanced CRC cases with BRAF mutation or MSI-H. In the present study, we assessed the concordance of KRAS/BRAF mutation status and MSI status in primary CRC and corresponding metastases.

High concordance of KRAS/BRAF mutation status and MSI status was observed between primary CRC and corresponding metastases in the present study. These high concordance rates were not different between synchronous and metachronous metastases. These results are in agreement with the notion that KRAS/BRAF mutations occur early in CRC carcinogenesis (20,28). Lymph node metastases showed a slightly lower concordance rate than other metastatic sites. Mao et al demonstrated that lymph node metastases indicated a lower concordance rate with KRAS mutation status (29) and this support our results. However, concordance rate of KRAS/BRAF mutation and MSI-H was >90% in lymph node suspected metastases macroscopically, metastatic lymph node will be useful for mutation analysis after confirmation of enough tumour cells and content microscopically. With regard to the other metastatic sites, a high concordance was observed between primary CRC and corresponding liver metastases. Knijin et al demonstrated a high concordance, i.e. 96.4%, in 305 liver metastases (30). In addition, this study showed high concordance of KRAS/BRAF mutation between primary tumour and peritoneal metastases. No other studies have systematically compared the concordance of KRAS/BRAF mutation status in primary CRC with corresponding peritoneal metastases.

Regarding MSI status, a high concordance rate was also shown between primary CRC and corresponding metastases. This result suggested that cancer cells do not change their MSI status during progression. MSI-H CRC consists of three types, which harbours a germline mutation in the MMR gene (e.g. Lynch syndrome), acquires epigenetic change in the MMR gene (e.g. MLH1 promoter hypermethylation) and uncertified germline mutation without MLH1 promoter hypermethylation (e.g. Lynch-like syndrome). Our results indicated perfect concordance of MSI status was observed in two types, i.e. Lynch syndrome and MLH1 promotor hypermethylation (3 Lynch syndrome cases and 10 MLH1 promoter hypermethylation cases) between primary CRC and corresponding metastases (Table II). This is the first study of concordance rate of MSI status between primary and metastatic CRC using Bethesda markers. Recently, Haraldsdottir et al reported perfect concordance of MMR deficiency evaluated by immunohistochemistry (IHC) between primary CRC and corresponding metastases (31).

In this study, 23 cases showed discordance of mutation status between primary CRC and corresponding metastases. Several reasons are conceivable. First, it could be speculated that discrepancies may depend on the molecular heterogeneity in primary CRCs. For instance, intratumoural heterogeneity for a KRAS point mutation was observed within 20–60% CRC cases in previous studies (32,33). In contrast to KRAS mutation status, BRAF mutation status did not show heterogeneity in previous studies (34,35). Mao et al have reported a higher concordance rate of BRAF mutation status (93.6%) between primary and lymph node metastases (29). Our results, which showed a perfect concordance rate in BRAF-mutant cases, is in agreement with these studies.

Second, discordant results could be explained by the acquisition of the mutation during the disease progression. However, KRAS mutation occur in early stage of carcinogenesis (20,28), it may be rare that CRC acquired KRAS mutation after metastasis (21,36).

Third, selecting improper samples containing a high number of normal or necrotic cells, could create discordance between primary CRC and corresponding metastases. In this study, samples that were suspected to contain enough cancer cells macroscopically by surgeons were used for mutation testing. Consequently, these samples might not include enough cancer cells especially in the lymph nodes. High concordance rate might be shown in previous studies that used laser microdissection-collected cancer cells from lymph node metastases (37,38).

In conclusion, although attention should be paid to selecting and sampling tissue, high concordance rate of KRAS/BRAF mutation status and MSI status was observed between primary CRC and corresponding metastases, regardless of metastatic sites and synchronous/metachronous types. Therefore, to choose the appropriate regimen for therapy, either primary or metastatic CRC can be used for testing KRAS/BRAF mutation status and MSI status.