Molecular biomarkers and histological parameters impact on survival and response to first- line systemic therapy of metastatic colorectal cancer patients

Colorectal cancer (CRC) is one of the most common cancer and one of the leading causes of cancer death in the world. It is one of the most common cancers in Slovenia and, according to the Cancer Registry of Slovenia, 1353 new patients were diagnosed with CRC in 2015.1 Metastatic disease is still incurable, with 5% five-year overall survival (OS) without treatment. With the introduction of new chemotherapy, using oxaliplatin and irinotecan in the current management of metastatic disease, in combination with biologicals, targeting epidermal growth factor- mediated growth regulatory pathway and the vascular endothelial growth factor-mediated angiogenesis pathway we can prolong the progression-free survival (PFS) and OS of these patients.2, 3, 4, 5, 6, 7, 8, 9, 10 In selected patients with appropriate combination of therapy and surgery we can achieve approximately a 50% five-year OS.

The development of CRC is a multistep process, which accumulates different gene mutations, chromosomal abnormalities and epigenetic changes.11 The mutations within KRAS proto-oncogene, predominately within codons 12 and 13, activate RAS/RAF signalling and are thought to occur early in carcinogenesis of CRC. The KRAS status is the first molecular marker to predict the response to anti-EGFR monoclonal antibodies cetuximab and panitumumab in metastatic CRC (mCRC) patients, and it needs to be determined before deciding in favour of treatment with anti-EGFR antibodies. As the KRAS mutations occur early in CRC formation, there is a high concordance between the KRAS mutations of primary tumour and metastases, which was confirmed in previous studies.11 In the retrospective study, de Roock et al. raised the possibility that the patients with the KRAS mutation in codon 13 might have benefited from anti-EGFR antibodies treatment.12, 13 The mutations in KRAS gene were found in approximately 30 to 40% of mCRC patients, reported in previous literature, but, only 40 to 60% of these patients with wtKRAS will respond to anti-EGFR antibodies treatment. Therefore, other molecular markers downstream of EGFR in the RAS/RAF/MAPK pathway and other effector pathways were found to be involved to predict the response to specific systemic therapy.

The BRAF gene encodes a serine/threonine protein kinase of the RAS/RAF/MEK/ERK kinase pathway and it is also involved in CRC carcinogenesis.11, 14, 15 The most common mutation of the BRAF gene is V600E which is found in approximately 5 to 9% of mCRC.14, 15 The same was reported in our previous study carried on Slovenian patients with CRC where the BRAF V600E mutation was found in 5.1% of patients.16 Previous retrospective studies suggested that mutated BRAF was a marker of resistance to anti-EGFR therapy and that the patients with mutated BRAF had significantly shorter PFS and OS than the patients with wtBRAF tumours.14, 15 The mutations in the KRAS and BRAF genes have been reported to be mutually exclusive. In the retrospective analysis by Fariña- Sarasqueta et al., it was also shown that the BRAF V600E mutation was an independent prognostic factor for the OS of patients with CRC in stages II and III, while the KRAS mutations did not have any effect on the OS of these patients.17 They concluded that the prognostic role of the KRAS mutations in an adjuvant setting had to be determined. In published clinical studies the BRAF V600E mutation in mCLC is conferred to a poor prognosis regardless of treatment, but these patients may have some benefit from the treatment with cetuximab in combination with chemotherapy as the first-line therapy, except for the patients in whom the disease has progressed after the first-line therapy.15, 16, 17 The status of mutations in the RAS gene is a new molecular predictive factor for response to treatment with EGFR inhibitors in mCRC. These mutations in the RAS gene, in the codons 12, 13, 61 and 146, according to the literature data, are about 15%, and they are determined from autumn 2013 at our Institute of Oncology.3, 4, 9

The aim of this prospective study was to analyse overall response rate (ORR), time to progression (TTP) and OS of the patients with mCRC treated with first-line systemic therapy in respect of histological parameters of primary tumour KRAS and BRAF status. We additionally retrospectively analysed other mutations in RAS genes and their impact on OS and TTP.

To our knowledge a prospective clinical study in patients with CRC determining the prognostic and predictive significance of the histological characteristics of primary tumours with mutations in the KRAS and BRAF gene regarding response to treatment has not yet been published. Our study was the first of that kind. The study was conducted from November 2010 to December 2012 and154 patients were included, which was in line with the research plan. The percentage of patients with T4 tumour, G3 tumour, with more than 3 affected regional lymph nodes (N2), vascular, perinevral invasion and lymphangiosis were comparable with previously published data.2, 18 Parameters as perforation of the primary tumour during surgery, presence of tumour deposits in subserosa, mesentery, peritonealised pericolic or perirectal mesenteric tissues and iresectability of the primary tumour, were excluded, because of the insufficient number

According to our results, poorly differentiated tumour cells and regionally advanced disease with a greater number of affected regional lymph nodes indicate a worse prognosis and a shorter survival of these patients. Other histological parameters of the primary tumour, such as the size of T4 tumour, carcinomatous lymphangiosis, vascular and perinevral invasion, which speak of a greater biological aggressiveness of the disease and are important for decision on adjuvant systemic treatment in stages II and III, have not proved to be predictive factors of a worse outcome of the disease, although the difference in size of the primary T4 tumour was marginal (p = 0.08). We hypothesize that the cause of this could be the insufficient number of patients with such histological parameters as well as the short observation period (median observation period of 15 months), since the median OS in these patients has not yet been reached. Another possible cause may be that these histological parameters are not significantly relevant for the further course of the disease. According to the data from the literature, locally advanced primary tumour, with the infiltration of the free surface of serosa, and the incidence of a large number of regional lymph nodes, vascular invasion and carcinomatous lymphangiosis are independent prognostic factors for worse outcome of the operable disease.2, 19 In the case of disseminating disease this parameters also predicts aggressive course although their importance in the spread of disease is not fully defined.

The percentage of mutations in the codon 12 and 13 of the KRAS gene was found in 42% of patients, which is in accordance with the literature data in which 30 to 40% mutations are reported.20, 21, 22, 23 According to the data from the retrospective analyses of the published clinical studies the mutations in codon 12 are more frequent than mutations in codon 13, which is also confirmed by our study. Namely, 73% of the patients had mutations in codon 12, and 27% of the patients had mutations in codon 13. All the mutations analyses in our study were performed on the tumour tissue of the primary tumour. Given the proven high degree of compliance of the mutations between tumour tissue of the primary tumour and metastases, no further analysis of the status of mutations in the KRAS gene in operable metastases or their biopsy has been done.24 In our already published retrospective analysis, which included 176 patients, KRAS mutations in the genes were present in 30% of the patients and more often in codon 12 than in codon 13 (84% in codon 12 vs. 16% in codon 13).24

The median OS of patients with the non-mutated KRAS gene was shorter in our study than in patients with mutant KRAS genes, but, the difference was not statistically significant (p = 0.47), suggesting that the mutation status in the KRAS gene predicts response to the treatment with EGFR inhibitors, but is not a prognostic factor for survival of the patients. In some patients with a wt-KRAS gene the presence of the V600E mutation in the BRAF gene most likely can be the cause of a shorter survival, since the presence of the V600E mutation in the BRAF gene, according to retrospective reports of previous clinical studies, indicates a worse outcome of metastatic disease.25, 26, 27, 28 The mutation of the V600E in the BRAF gene was present only in patients with a non-mutated KRAS gene, which was expected in accordance with previous reports of mutually exclusive exclusion of the simultaneous presence of the both mutations.24 The median OS of patients with mutations in codon 12 was longer than in patients with mutations in codon 13 but was not significant (p = 0.41). According to the data from the literature, patients with mutations in codon 13 have a shorter survival than patients with mutations in codon 12. Patients with mutations in codon 13 respond to treatment with EGFR inhibitors, otherwise in a lower percentage as patients without mutations in the KRAS gene. The retrospective analysis of the efficacy data of these patients with mutations both in codon 12 and codon 13 treated with another EGFR panitumumab inhibitor did not confirm this.28, 29, 30 In standard treatment with regard to systemic treatment recommendations, patients with mutant KRAS gene do not receive a therapy with EGFR inhibitors.15 In our study, we also analysed the efficacy of systemic therapy in patients with mutant KRAS gene with respect to the presence of mutations in codons 12 and 13, none of whom received EGFR therapy. In our analysis, however, we found that patients with mutations in codon 13 respond to a higher percentage to systemic therapy than patients with mutations in codon 12 (47% vs. 23%). The definitive conclusions of this analysis are difficult to give, given that the patients with mutations in codon 13 are in lower number than patients with mutations in codon 12 (27% vs. 73%). These mutations are less frequent in accordance with the already published data from our retrospective study (16% vs. 84%) and other published data (16–28% mutations in the codon 13 KRAS gene).28, 29, 30 However, given that they have a shorter survival than patients with mutations in codon 12, but respond in a higher percentage to the first treatment, we believe that results of this analysis will help in the treatment of these patients, and in particular in reducing the disease burden and improving the quality of life.

In our study, we also analysed TTP of the disease in patients with operable disease. There were 64 such patients, with operable primary tumour without metastases, 16 patients with rectal cancer and 48 with colon carcinoma. Median time to recurrence of the disease in patients after primary treatment of operable disease with a non-mutated KRAS gene was shorter than in patients with mutated KRAS gene (20 vs. 21 months), but the difference was not significant (p = 0.88). We assume that the V600E mutation in the BRAF gene, which was found in four patients with a non-mutated KRAS gene contributes for such results. With this, we confirmed the prognostic importance of the V600E mutation in the BRAF gene for early progression and worse outbreaks in patients with a non-mutated KRAS gene.31, 32 The TTP of the disease in patients primarily treated for operable disease with operation was longer in those who had mutations in the KRAS codon 12 than in those with mutations in codon 13 (21 vs. 16 months), otherwise the difference was not significant (p = 0.53).

Regarding the presence of the V600E mutation in the BRAF gene in our study, as previously stated it was found in four patients, representing 3% of all patients. According to the literature, the frequency of this mutation is from 5 to 10%.2, 3, 9, 13 Also, according to our previous retrospective analysis, the V600E mutation in the BRAF gene was present at a higher percentage of patients, in 7.4%.24 It is assumed that the specificity of the patients involved with such primary tumour characteristics may be due to the fact that the V600E mutation was present in a lower percentage. All four patients with the V600E mutation had colon cancer, three patients were previously on adjuvant treatment, and one had a primary disseminated disease. In a retrospective analysis of the study involving 94 patients with rectal carcinoma, none of the V600E mutations in the BRAF gene were detected in either patient.33 This suggests that this mutation is probably less important in the development of rectal cancer. The time before progression of the disease in patients after primary treatment is longer for those who did not have the V600E mutation in the BRAF gene than in the V600E mutation (20 to 12 months), but the difference was not significant (p = 0.14). The cause is most likely a small number of patients with this mutation, but such result suggests a worse prognosis of these patients and an earlier recurrence of the disease. In the survival analysis, patients without V600E mutations had a significantly longer median OS than those with V600E mutation (59.2 vs. 27.6 months; p = 0.05). In a retrospective analysis of 176 patients, 13 patients with the V600E mutation in the BRAF gene had a significantly lower expected median survival than patients without this mutation (44.9 vs. 107.4 months; p = 0.04.24 Both studies, both the previous retrospective and the current prospective, confirm the prognostic importance of the V600E mutation in the BRAF gene for predicting a worse disease outcome with shorter patient survival.

Furthermore, we examined the response rate according to the type of systemic treatment in patients with non-mutated KRAS gene. There were 88 such patients. Three hundred and sixty patients with a non-mutated KRAS gene received the first line of systemic treatment with EGFR-inhibitor cetuximab in combination with oxaliplatin or irinotecan and fluoropyrimidines, 35 patients with a non-mutated KRAS gene received the first line of systemic treatment with angiogenesis inhibitor bevacizumab in combination with oxaliplatin or irinotecan and fluoropyrimidines. A higher percentage of objective responses (both complete and partial), both for treatment and stability of the disease was achieved in patients with non-mutated KRAS gene receiving cetuximab in combination with oxaliplatin or irinotecan and fluoropyrimidines.

The definitive conclusions about this analysis cannot be made within our research, since this is a small number of patients, and the number of patients was not balanced in number. In addition, the types of systemic chemotherapy that patients received differ. However, it was found that patients with a non-mutated KRAS gene had a better response to systemic first-line treatment involving cetuximab regardless of the type of systemic chemotherapy as patients with the non-mutated KRAS gene that received angiogenesis inhibitor bevacizumab in combination with systemic chemotherapy.

Colorectal adenocarcinoma is a heterogeneous disease. We do not know all the factors that influence the course of the disease and the response to specific systemic therapy. The determination of other mutations in the KRAS gene, such as mutations in codon 61 and 146, and mutations in the NRAS gene, has been standard in clinical practice since 2014. The frequency of mutations is about 15% and the patients with a non-mutated RAS gene have about 40% of this mutation. In our retrospective analysis, we found 17% of these additional mutations, which is consistent with reports from previous literature and they are predictive for response to EGFR inhibitor therapy. Patients treated with first-line anti-EGFR targeted therapy without these additional mutations in RAS gene, had longer survival and time to progression compared to the patients who had them. Our results confirmed predictive value of these mutations in RAS gene for anti-EGFR targeted therapy. However, our group of patients is moderate, but it is consisted with previously published data.33, 34

Adverse effects of systemic treatment were evaluated according to NCI-CTC, version 4.03. They were mostly of Grade 1 and 2. There were 2% of G3 diarrhea and hypermagnesemia and 1% of G3 bleeding. One patient had G4 adverse effect (febrile neutropenia). In our study, adverse effects were less common than in previous published literature.35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47 None of the patients died of adverse effects.

However, all patients with RAS non-mutated genes do not respond to treatment with anti-EGFR inhibitors. Other biomarkers have to be determined. One of them is mismatch repair- deficiency (MSI-H), which is present in 3.5% to 5.0% of mCRC.48, 49 Immunotherapy with checkpoint inhibitors is recommended in second-line therapy in these patients. In clinical trials, there are other biomarkers in response to specific systemic therapy, such as amfiregulin and epiregulin, PI3K and PTEN mutations, but are still not recommended in regular clinical practice.50 In order to made definite reasoning on this issue we need more information and more studies, both, retrospective and prospective should be performed.

According to the results of our prospective study G3 of tumour differentiation (poorly differentiated adenocarcinoma) and N2 status of regional lymph nodes (more than 3 metastatic regional lymph nodes) show a greater biological aggressiveness of the disease and shorter survival after the first relapse, but do not affect the efficacy of the first-line systemic therapy.

The mutations in the KRAS gene have predictive value for the response to treatment with EGFR inhibitors, but do not have prognostic significance for the survival of patients with disseminated colorectal adenocarcinoma. The KRAS mutations in codon 12 and 13 differently affect the survival of these patients and the response rate of first-line systemic therapy.

The V600E mutation in the BRAF gene does not have the prognostic significance for the decision on the type of first-line systemic therapy, but it predicts a poor prognosis of the disease and shorter survival of these patients at first relapse.

According to our restrospective analysis of additional mutations in NRAS gene and in KRAS codon 61 and 146 are predictive for treatment with anti-EGFR targeted therapy.