Background
Access to the host vascular system and the formation of tumor blood supply represent a limitation for the progression of solid tumors. A switch to an angiogenic phenotype occurs early in tumorigenesis and is induced by hypoxia, metabolic stress, or oncogene activation. The vascular endothelial growth factor (VEGF) family of ligands and their receptors play a crucial role in tumor angiogenesis and neovascularization. VEGF stimulates proliferation of endothelial cells, vascular permeability, and attracts bone marrow-derived endothelial precursors to the tumor microenvironment. Increased levels of pro-angiogenic factors and tumor vascularization are associated with increased risk of tumor metastases and inferior survival of patients with metastatic colon cancer and other types of cancer [
1].
Bevacizumab (bev) is a recombinant, humanized, monoclonal antibody that inhibits VEGF receptor signaling by binding to VEGF-A. Bevacizumab was the first angiogenic inhibitor shown to prolong survival in advanced cancer [
2]. Bev in combination with chemotherapy significantly improves progression-free survival and overall survival in the first- and second-line treatment of metastatic colorectal cancer (mCRC) [
3,
4].
Currently, tumor KRAS gene status remains one of the most important predictive biomarker used in management of colorectal carcinoma. Beside its established role in anti-EGFR therapy, KRAS mutation may also affect the clinical outcome of anti-angiogenic therapy. Activating KRAS mutation in carcinoma cells may induce angiogenesis via several mechanisms. Oncogenic mutations of Ras intracellular signal transducers can increase VEGF production by tumor cells [
1,
5]. Moreover, activating Ras mutation resulted in up-regulation of pro-angiogenic interleukin-8 (CXCL-8) leading to recruitment of endothelial cells, tumor vascularisation and tumor growth
in vitro [
6] and more aggressive biological behavior
in vivo [
7]. Additionally, Ras signaling promoted angiogenesis through repression of anti-angiogenic thrombospondin-1 [
8]. Thus, Ras oncogenes may contribute to tumor progression by both a direct effect on tumor cell proliferation as well as indirectly by facilitating tumor angiogenesis in a paracrine fashion
.
The aim of the present observational study was to assess the role of KRAS status in mCRC patient treated in the the first-line with anti-angiogenic agent bevacizumab combined with oxaliplatin- or irinotecan-based chemotherapy in current clinical practice. Furthermore, we aimed to identify risk factors related to disease characteristics and treatment within the context of KRAS status.
Methods
Data collection
Data from 1622 mCRC patients with known KRAS status who received first-line bevacizumab plus FOLFOX, XELOX, FOLFIRI or XELIRI between January 2005 and April 2013 were mined from the non-interventional post-registration database collecting epidemiological and clinical data of patients with advanced and metastatic CRC on targeted treatment (CORECT registry). The protocol was approved by the independent ethics committee at each participating centre (Ethics Committee (EC) of the Ceske Budejovice Hospital, EC of the Chomutov Hospital, EC of the General University Hospital in Prague, EC of the Jihlava Hospital, EC of the Liberec Regional Hospital, EC of the Masaryk Hospital in Usti nad Labem, EC of the Masaryk Memorial Cancer Institute in Brno, EC of the Na Bulovce Hospital in Prague, EC of the Na Homolce Hospital in Prague, EC of the Novy Jicin Hospital, EC of the Pardubice Regional Hospital, EC of the St. Anne’s University Hospital (UH) in Brno, EC of the Thomayer Hospital in Prague, EC of the Tomas Bata Regional Hospital in Zlin, EC of the UH Brno, EC of the UH Hradec Kralove, EC of the UH in Motol, Prague, EC of the UH Olomouc, EC of the UH Ostrava, EC of the UH Pilsen) and complied with the International Ethical Guidelines for Biomedical Research Involving Human Subjects, Good Clinical Practice guidelines, the Declaration of Helsinki, and local laws. The CORECT database includes data of approximately 96% of all mCRC patients treated with targeted therapies in the Czech Republic. The security of individual records is guaranteed via de-identified data collection and the system meets all valid rules on the protection of personal data. Only authorized users can access the system. No exclusion criteria were applied in data input, mining or analysis.
Evaluated parameters
KRAS testing
Tumor tissue samples were tested in local referral laboratories to identify mutations in codons 12 and 13 of exon 2 of the KRAS gene using established methods according to American Society of Clinical Oncology guidelines. In the Czech Republic in 2009, 60.2% colorectal tumors were tested as wild-type KRAS (wtKRAS) and the incidence of mutation detected was as follows: G12D (11.9%), G12V (8.8%), G13D (5.8%), G12C (3.3%), G12A (2.8%), G12S (2.0%), G12R (1.0%), G13C (0.3%), G13V (0.1%) [
9].
Treatment
Choice of chemotherapy backbone regimen was at the physicians’ discretion according to national guidelines based on European Society for Medical Oncology guidelines. The dosage of chemotherapeutic agents were determined according to the body surface area and dose reduction was recommended only in cases of severe (grade 3) adverse events to assure optimal chemotherapy dose intensity. The chemotherapy regimens were as follows: FOLFOX4 (oxaliplatin 85 mg/m2 IV day1; leucovorin 200 mg/m2 IV days 1 and 2; 5-FU bolus 400 mg/m2 IV days 1 and 2; 5-FU 600 mg/m2 IV 22-hour continuous infusion days 1 and 2 every 2 weeks), FOLFIRI (irinotecan 180 mg/m2 IV day 1; leucovorin 200 mg/m2 IV day 1 and 2; 5-FU 600 mg/m2 IV 22-hour continuous infusion days 1 and 2 every 2 weeks), XELOX (oxaliplatin 130 mg/m2 IV day 1; capecitabine 1000 mg/m2 twice daily PO for 14 days every 3 weeks), or XELIRI (irinotecan 250 mg/m2 IV day 1; capecitabine 1000 mg/m2 twice daily PO for 14 days every 3 weeks). Bevacizumab was administered at a dosage of 5 mg/kg IV every 2 weeks or 7.5 mg/kg IV every 3 weeks depending on the chemotherapy regimen. Dosage of bevacizumab was not reduced.
Clinical response and follow-up
Data were reported to the registry every 6 months. Patients’ response to treatment and tumor measurements were determined by independent evaluation by two local radiologists according RECIST version 1.0. Patients were followed-up until death or loss to follow-up.
Statistical analyses
Standard descriptive statistics were used to characterize the sample data set. Differences in initial categorical parameters were assessed using the Pearson chi-square test. Comparisons of the subgroups in continuous variables were based on the Mann–Whitney test. Both overall survival (OS) and progression-free survival (PFS) were estimated using the Kaplan-Meier method. OS was defined as the time from bevacizumab treatment initiation to death due to any cause. PFS was defined as the time from bev treatment initiation to progression or death due to any cause. Statistical significance of the differences in Kaplan-Meier estimates was assessed using the log-rank test. For all point estimates the 95% confidence interval (95% CI) was calculated. The univariate Cox proportional hazards model was used to evaluate the influence of all potential predictive and prognostic factors on the survival measures and subsequently the multivariable Cox proportional hazards model was used to quantify the influence of KRAS status on survival in the presence of other potential predictive and prognostic factors. A level of significance α = 0.05 was used in decision on statistical significance.
Discussion
The present data demonstrates that mCRC patients treated with bevacizumab plus combination chemotherapy have similar outcomes regardless of KRAS mutation status. Currently, standard first-line treatment for patients with mCRC includes combination chemotherapy with either an anti-EGFR agent (cetuximab or panitumumab), or bevacizumab. Adding bevacizumab to chemotherapy prolongs survival in advanced colorectal cancer patients. However, not all mCRC patients respond to this treatment and identification of a subgroup of patients who will benefit most from bevacizumab currently represents a challenge. Data on the prognostic and/or predictive significance of KRAS status in mCRC patients treated with first-line bevacizumab-containing regimens were obtained mostly as secondary endpoints in trials comparing bevacizumab vs anti-EGFR therapy efficacy (see Additional file
1: Table S1). In general, these studies failed to demonstrate KRAS status as a predictive factor for anti-VEGF therapy in the first-line treatment of mCRC, consistent with the present study. The results of the MACRO trial indicated that KRAS status may be a prognostic factor in mCRC patients treated with bevacizumab; however, that study setting did not allow the evaluation of the predictive role of KRAS status [
10]. Evaluation of OS in the studies evaluating the prognostic role of KRAS in patients treated with first-line bevacizumab, including the present study (summarized in Additional file
1: Table S1), is difficult because the patients’ survival is influenced by subsequent line of therapy and patients with wtKRAS tumors could benefit often from anti-EGFR therapy after progression on bev/chemotherapy. Accordingly, the insignificant prolongation of overall survival by 2.3 months in wtKRAS patients over the mutated KRAS subgroup may be to some extent attributed to subsequent anti-EGFR therapy (Table
1).
The adverse events from bevacizumab treatment observed in this study are consistent with those published in post-registrational trials, including BEAT, BRiTE and ARIES [
11-
13]. We confirmed that severe (grade 3–5) events, including arterial thromboembolic events, proteinuria, gastrointestinal perforation and wound-healing complications were rare with incidence and spectrum of adverse events comparable across all age groups [
14]. Taken together, the adverse events related to bevacizumab are acceptable and well tolerated.
Prospective data on the optimal sequence of chemotherapy regimen used with targeted therapy of metastatic colorectal cancer is lacking. It has been demonstrated that first-line regimens combining fluoropyrimidines with oxaliplatin or irinotecan are of comparable efficacy [
15,
16]. Accordingly, the optimal choice for the chemotherapy to be used in combination with bevacizumab in the initial treatment of mCRC should be based on the characteristics of the patient and tumor, as well as the drug efficacy and toxicity profile. FOLFOX or XELOX is considered as a convenient first-line treatment option for mCRC patients as they cause less diarrhea, nausea, and alopecia than FOLFIRI or XELIRI, despite a potential for dose-limiting neurotoxicity. In the present study, patients who received in the first line bev/OX-based chemotherapy tended to have longer overall survival over those treated with irinotecan-based chemotherapy with bevacizumab. As we did not observe a difference in performance status (ECOG of 0 or 1) between patients who started treatment with bev/OX versus bev/IRI-based therapy in the first-line setting (data not shown), the trend toward better OS in bev/OX subgroup is unlikely to be attributed to better performance status in these patients. Interestingly, the inferior efficacy of treatment with the irinotecan backbone was restricted to the subgroup of patients with KRAS mutation, with median OS in the bev/OX-based subgroup being 5 months longer than in the bev/IRI-based treatment. Based on
in vitro experiments showing that KRAS mutation is a predictor of oxaliplatin sensitivity in colon cancer cells, it has been suggested that mutant KRAS CRC patients might benefit more from receiving first-line oxaliplatin-based regimens [
17]. Although we have not observed a better response in KRAS mutated mCRC patients treated with bev/OX-based therapy over those with wtKRAS, as it was reported when patients with advanced CRC were treated in the first line with FOLFOX-6 [
18], the reported sensitivity of mutated KRAS carcinoma cells to oxaliplatin may explain the improved clinical outcome when KRAS mutant patients were treated with bev/OX-based over bev/IRI-based therapy.
We have confirmed that patients presenting with synchronous metastases have an inferior prognosis compared to patients with metachronous metastases. Among the evaluated parameters, metastatic involvement of multiple organs at time of treatment initiation was the strongest prognostic factor reducing both PFS and OS. Within the subgroup of patients with metastatic disease limited to one distant organ we did not observed difference in clinical outcome with regard to hepatic or extrahepatic (including pulmonary) involvement. Interestingly, together with Rossi
et al. [
19], we observe trend toward longer both PFS and OS in patients with extrahepatic disease and KRAS mutation than in wtKRAS subgroup when treated with bevacizumab plus chemotherapy. Apart from this fact, we and others [
20-
22] have reported that KRAS mutation in colorectal cancer itself is associated with pulmonary metastasis. Findings from the VICTOR trial showing that KRAS mutant tumors are associated with an increased risk of lung relapse in CRC patients supported the role of chest imaging in surveillance of colorectal cancer patients, particularly of those with resected primary mutated KRAS carcinoma [
20,
22]. The reason for increased incidence of lung metastases in KRAS mutated colorectal tumors remains unknown at this moment. Considering the decreased proportion of lymph node metastasis in mutated KRAS patients compared to wtKRAS subgroup (Table
1, [
23,
24]), it seems that carcinoma cells with activating mutation in KRAS may exhibit a more hematogenous metastatic spread rather than along a lymphogenous path. Survival of tumor cells within the bloodstream and adhesion in the vasculature at the metastatic sites depend on tumor cell – platelet interactions [
25]. We hypothesize that activating mutation of KRAS inducing expression of molecules responsible for interaction with platelets, such as tissue factor [
26], cyclooxygenase and metalloproteinase-9 [
27], or cathepsin B [
28] might contribute to increased protection of these carcinoma cells against shear stress as well as to enhanced adhesion properties which in turn leads to onset of pulmonary metastasis of mutated KRAS carcinoma cells and higher metastatic activity in general [
29].
The present study is a retrospective analysis, and thus an unintentional selection bias for a subset of patients is possible. However, the parameters of our analysis that confer substantial reliability to the presented results are, for example, unselective multicenter input of evaluated data and the proportion of tumors with wtKRAS vs mtKRAS, mirroring the proportion of KRAS mutation previously detected in the Czech Republic during a one year survey [
9]. We excluded a potential bias caused by inclusion patients with known KRAS status only, as no difference was observed between PFS of mCRC patients treated in the first-line with bevacizumab and OX- or IRI-based chemotherapy with KRAS-known (11.5 months, 95% CI 11.0-12.1) and KRAS-unknown (11.6 months, 95% CI 11.0-12.2) which is further underlined by KRAS-unselected published data on first-line bevacizumab in mCRC patients based on CORECT registry [
14]. An existing limitation of the present study is that we had no data specifying the type of KRAS mutation, NRAS mutation or data on BRAF mutation. Mutation of BRAF and KRAS are, in the vast majority cases, mutually exclusive, and thus, the wtKRAS subgroup in the present study included an unknown number of BRAF mutated cases who might have worse survival rates. However, estimating from our findings of BRAF mutations in 3.8% of mCRC patients (unpublished data from diagnostic testing), contamination of the wtKRAS subgroup with mutant BRAF patients is unlikely to have been substantial. Regarding specific mutations in the KRAS gene, there is emerging data showing that mutation in codon 13 leads to a more aggressive disease course with extensive synchronous metastases compared to colorectal carcinoma with mutation in codon 12 [
30]. Mutation in codon 13 is reported to confer worse prognosis and outcome than alteration in codon 12 [
31]. On the other hand, when treated with cetuximab and chemotherapy, patients with mutation in codon 13 demonstrated a trend toward longer PFS than the codon 12-mutated subgroup but also a better outcome than patients with tumors without KRAS mutation [
30,
31]. KRAS codon 12 (Cys12 = G12C) or codon 13 (Asp13 = G13D) mutation resulted in different vascular strategy
in vitro; the neovascularization in Asp13 cells was associated with high levels of VEGF-A but less effective vasculature compared to their Cys12 counterpart [
32]. Moreover, a trend toward increased vascular invasion in codon 13-mutated tumors has been observed in colorectal cancer patients [
33]. Thus, the different biological role of KRAS mutation in codon 13 may have also implications for anti-angiogenic treatment with bevacizumab.
Acknowledgements
Maintenance of the CORECT database is supported by unconditional grants from Roche, Merck, and Amgen. The funding bodies had no role in design, collection, analysis or interpretation of the data, writing of the manuscript or the decision to submit the manuscript for publication. We would like to thank all physicians who provided data for the CORECT registry and namely the following heads of the comprehensive cancer centers involved in the CORECT project: Jitka Abrahamova, Praha; Jiri Bartos, Liberec; Otakar Bednarik, Brno; Martina Chodacka, Chomutov; David Feltl, Ostrava; Jindrich Finek, Plzen; Vaclav Janovsky, Ceske Budejovice; Jana Katolicka, Brno; Milan Kohoutek, Zlin; Milan Lysy, Usti nad Labem; Jiri Petera, Hradec Kralove; Lubos Petruzelka, Praha; Jana Prausova, Praha; Lubomir Slavicek, Jihlava; Renata Soumarova, Novy Jicin; Vladimira Stahalova, Praha; Martin Safanda, Praha; Jaroslav Vanasek, Pardubice. We also thank to Dr. Michael A. Sheard for critical reading of the manuscript and his beneficial comments. The work was supported by the European Regional Development Fund and the state budget of the Czech Republic for the Regional Centre for Applied Molecular Oncology – RECAMO (CZ.1.05/2.1.00/03.0101), by MH CZ - DRO (MMCI, 00209805) and the project BBMRI_CZ (LM2010004).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
Development of IT infrastructure and organization of the project CORECT is supported by investigator-initiated, unrestricted, research grant provided by Roche, Merck, and Amgen. The funding bodies had no role in design, collection, analysis or interpretation of the data, writing of the manuscript or the decision to submit the manuscript for publication. TP and LD are researchers responsible for CORECT database at Institute of Biostatistics and Analyses, Masaryk University. BB has received honorarium for a lecture from Roche. LO has received speakers’ honoraria from Roche and Merck. IK, RV and BM have received speakers’ honoraria from Roche, Merck, Amgen and have acted on advisory board for Roche, Merck and Amgen. LZD has received honorarium for a lecture from Merck. ZB, JH, and DV have declared no financial or non-financial competing interests.
Authors’ contributions
BB conceived of the study, participated in its design, contributed to data interpretation, and drafted the manuscript. ZB performed data extraction and data analysis, prepared figures and tables, contributed to data interpretation, and drafted the manuscript. JH drafted and edited the manuscript. LO reviewed and edited the manuscript. IK reviewed and edited the manuscript. BM contributed to data interpretation, reviewed and edited the manuscript. TP supervised data collection and management of the CORECT registry, and contributed to data extraction and data analysis. LD and RV initiated and organized the CORECT registry, supervised data collection and management, and reviewed the manuscript. DV contributed to data interpretation, reviewed and edited the manuscript. RV initiated and organized the CORECT registry, supervised data collection and management, and reviewed the manuscript. LZD contributed to conceive of the study design, participated in study coordination, contributed to data analysis and interpretation, drafted and finalized the manuscript. All authors read and approved the final manuscript.