Colorectal cancer (CRC) is a major public health problem. Its incidence in France is increasing 1 with approximately 40 000 new cases per year 2 and prognosis remains poor 3 . Over 50% of patients will develop metastases and will be candidates for palliative chemotherapy 4 . Bevacizumab is a monoclonal antibody directed against the vascular endothelial growth factor (VEGF). It has proven efficacy in the treatment of metastatic CRC when combined with chemotherapy 5 6 7 . Irinotecan, infusional 5-fluorouracil (FU), leucovorin (LV) (FOLFIRI®) and bevacizumab (FOLFIRI® + B) offered better outcomes when compared to irinotecan plus infusional fluorouracil (FU)/leucovorin (LV) (FOLFIRI®), irinotecan plus bolus FU/LV (mIFL), and irinotecan plus oral capecitabine (CapeIRI) in a randomized trial 5 . However, a relatively high rate of ≥ Grade 3 hypertension was observed.

Several gene polymorphisms may interfere with anticancer drug activity, and thus affect drug efficacy and toxicity. For FU, a thymidylate synthase promoter 28-bp tandem repeat (rs34743033) is associated with lower efficacy and increased toxicity 8 . For irinotecan, a UGT1A1 promoter TA repeat (rs8175347) is a risk factor for toxicity 9 . For Bevacizumab, several VEGFA single nucleotide polymorphisms (SNP) are known to influence VEGFA plasma concentrations 10 and to be associated with CRC risk 11 . In addition, prognostic 12 13 and predictive 14 15 roles of VEGFA variants have been identified in various studies.

The principal objective of this phase II trial was to evaluate efficacy of first-line treatment with FOLFIRI® + B for metastatic CRC patients in terms of response rates. Secondary objectives were to assess overall and progression-free (PFS) survivals, response duration, and toxicity. We also explored common gene polymorphisms known to interfere with the metabolism and/or activity of FOLFIRI® + B, located respectively in the thymidylate synthase (TYMS), UDP-glucuronosyltransferase 1A1 (UGT1A1) and VEGFA genes, looking for associations between these polymorphisms and the clinical parameters of toxicity and efficacy of the treatment.

Patients for this open-label, single arm, phase II trial were recruited from Institut Bergonié, the University Hospital of Bordeaux, and five general hospitals and private clinics in South-West France.

Inclusion criteria were: histopathologically-proven adenocarcinoma of the colon or rectum, non-resectable metastatic disease; no prior chemotherapy other than adjuvant chemotherapy (provided it had been discontinued > 6 months before study entry); Eastern Cooperative Oncology Group (ECOG) performance status ≤ 2; age ≥ 18 years; measurable metastatic disease per Response Evaluation Criteria in Solid Tumors (RECIST Version 3.0) 16 ; adequate hematological function [hemoglobin ≥10 g/dl, absolute neutrophils count ≥1.5 × 10⁹/l, platelets ≥100 × 10⁹/l]; adequate renal function [no proteinuria and creatinine ≤1.25 × the upper limit of the normal value (ULN)]; adequate hepatic functions [total bilirubin ≤1.25 × ULN, aspartate amino-transferase (AST) and alanine aminotransferase (ALT) ≤3 × ULN, in case of liver metastases, total bilirubin ≤1.5 × ULN and AST and ALT ≤5 × ULN]; intervals since inclusion of 4 weeks for eventual surgery or radiotherapy; ability to comply with scheduled follow-up and management of toxicity.

Exclusion criteria included: histology other than adenocarcinoma; non-measurable disease; adjuvant chemotherapy within 6 months or containing Bevacizumab; unresolved bowel or partial bowel obstruction; history of chronic diarrhea; severe gastrointestinal toxicity while receiving FU; current uncontrolled infection; serious illness or medical condition; previous abdominopelvic radiation therapy; known Gilbert’s syndrome; arterial thromboembolism accident or myocardial infarction within preceding 6 months; history of cancer other than colorectal, except for curatively treated non-melanoma skin cancer or in-situ cervical cancer; concomitant treatment with any other investigational drug; and pregnancy.

The protocol was approved by the regional Ethics Review Committee (Comité de Protection des Personnes du Sud-Ouest et d’Outre-Mer) and registered with clinicaltrials.gov (NCT00467142, registration date April 25, 2007). Each patient provided written informed consent. For the complementary pharmacogenetic study, patients were enrolled on a voluntary basis and gave specific consent.

The primary objective of this trial was to assess the objective response rate after treatment with FOLFIRI® + B in metastatic CRC. We observed 47.5% objective partial responses with no complete responses at six months. These results appear slightly higher than the most recent phase III trial report with 142 patients showing a response rate of 40.1% 20 , similar to as reported by Souglakos et al. 21 (45.5%), and higher than reported in a recent phase III trial (36.8% 22 , although they are lower than in previous studies (eg. BICC-C 5 or Kopetz et al. 23 ) with 59% and 65% objective responses respectively.

With a median follow-up time of 43.6 months, the median PFS was 10.3 months, and one-year PFS was 35%. This supports previous reports after FOLFIRI® + B treatment with PFS reported between 9 24 to 11.6 months 25 , and 12.8 months 23 . The median OS in the present trial was 25.7 months, with a one-year survival of 85%. Once again, these rates support previously and recently published rates with OS reported between 22 months 22 23.7 months 25 , 25.7 21 , 28 months (one-year OS of 87%) 26 and 31.3 months 23 . In a recent observational report involving over 240 patients, median PFS was reported at 10.2 months and median OS at 25.5 months, once again providing further support for these patterns 27 . Our results show that a curative hepatic surgery could be carried out for 11 patients (18.6%) that were judged to be unresectable before FOLFIRI + B treatment. The rate of hepatic surgery with FOLFIRI® + B was not reported in the BICC-C study 5 26 , and it was 6.5% in the Beat Study Cohort and 9.3% in Kopetz et al.’s study 23 .

In randomized trials in patients with metastatic CRC, bevacizumab has been shown to improve response rates, OS and PFS when combined with chemotherapy regimens like bolus FU/LV 28 , irinotecan plus bolus FU/LV (IFL) 6 and oxaliplatin plus infusional FU/LV (FOLFOX) 29 . In a systematic review for patients with advanced CRC receiving first- or second-line fluoropyrimidine-based chemotherapy, the addition of bevacizumab improved PFS and OS, although toxicity was also increased 30 . A more recent systematic review and meta-analysis including over 3000 patients from randomized trial 31 shows a distinct advantage for PFS when bevacizumab is added. In subgroup analyses, the effect was strongest for FU- and irinotecan-based chemotherapy regimens and less marked in oxaliplatin-based regimens.

Severe toxic effects were mainly hematologic and less frequently gastrointestinal. The rate of grade ≥3 toxicities was low. At 19.4%, neutropenia was the most frequent severe toxicity, as has been described in other trials 20 , although the rate was lower than described in other studies at (53.6%) 5 or (40%) 23 . The rate of severe diarrhea (11.3%) was similar to the BICC-C study (10.7%) 5 and higher than in Kopetz’s study (2%) 23 . Considering adverse events which could be related to bevacizumab, we had no gastrointestinal perforation compared to 0% 5 23 to 2% 25 reported in the literature; no severe bleeding compared to 0% 5 23 -3% 25 , no Grade 3/4 proteinuria, similar to 1% in Van Cutsem et al. 25 ; no severe hypertension, lower than reported in the literature at 5% 25 to 12.5% 5 and 19% 23 ; and 1.6% Grade 3/4 venous thromboembolism events compared to 1% 25 and 19% 23 . It should be noted that for the 11 patients who were operated on in this trial, no delayed wound healing was noted, with bevacizumab being stopped six weeks before surgery.

Associations between TYMS polymorphisms and FU efficacy and toxicity in CRC have been reported in several studies 32 33 34 35 but not all 36 . In our study, no influence of TYMS polymorphisms on the efficacy or toxicity of the treatment was demonstrated. The limited size of the population tested as well as the fact that the patients did not receive exclusively FU probably explain the lack of associations. The toxicity of irinotecan has been associated with a polymorphism of the SN-38-detoxifying enzyme, UGT1A1, located in the promoter of the gene 37 . However, the risk of experiencing irinotecan-induced hematological toxicity appears to be a function of the dose administered 37 ; the risk is higher at the dose usually prescribed in the US (340 mg/m² every 3 weeks) than at the dose prescribed in Europe (180 mg/m² every two weeks). This may also explain the lack of associations in our study.

We found an association between OS and the presence of a polymorphic exonic synonymous variation in the VEGFA gene. Interestingly, the rs25648 variation is one of the three SNPs influencing VEGF serum levels 38 , as well as VEGFA mRNA levels in colorectal adenocarcinoma 39 . One could hypothesize that the higher levels of VEGFA produced by the variant allele limits the efficacy of the anti-VEGF antibody and, therefore, has an impact on patient survival. However, the association of this polymorphism with outcome was only significant for OS and not for PFS, which would indicate a general prognostic impact rather than a predictive role for bevacizumab efficiency. Studies involving a larger number of patients should be undertaken with a special focus on this polymorphism which has not been studied extensively up to now.

The fact that a synonymous polymorphism in the coding sequence of VEGFA exerts an effect on VEGF levels in plasma may be explained by a difference in the 3-dimension structure or the half-life of the transcribed mRNA, which may introduce differences in its handling by the translation machinery. Alternatively, the difference in tRNA availability because of the use of a rare codon may lead to a decrease in translation rate, and thus to a reduced production of the protein.

Some limits should be taken into account when interpreting this data. Firstly, the non-comparative nature of this non-randomized phase II trial should be kept in mind when comparing efficacy results. It should also be noted that although we found a significant effect for the primary endpoint of objective response, no complete responses were observed.

The present results support the growing body of evidence from phase II 24 , phase III and observational studies indicating that FOLFIRI® + B is an active and safe treatment for first-line treatment of metastatic colorectal cancer, with almost half of patients showing an objective response and comparatively long median OS. Further, almost 1/5 initially unresectable patients became resectable after treatment, offering potential for longer survival. It has a good safety profile, with relatively low rates of thromboembolism compared to other alternative chemotherapy associations. The association between the genetic polymorphism rs25648 and improved OS is encouraging, but needs to be confirmed in further trials.

CRC: Colorectal cancer; VEGF: Vascular endothelial growth factor; FOLFIRI + B: Irinotecan, leucovorin, fluororacil + bevacizumab; FU: Infusional 5-fluorouracil; OS: Overall survival; PFS: Progression-free survival; RECIST: Response Evaluation Criteria In Solid Tumors; DNA: Deoxyribonucleic acid.

The authors declare that they have no competing interests.

YB conceived of the study and drafted the manuscript. CL and SMP participated in the design and coordination of the study and statistical analysis. MP participated in the design of the study, performed the statistical analysis and helped to draft the final manuscript. RB, PT and LC recruited patients to the study. VLM and SA carried out the molecular genetic studies and participated in the sequence alignment. DB, RBr and MF participated in the design of the study. JR carried out the molecular genetic studies, participated in the sequence alignment, and drafted the manuscript. All authors read and approved the final manuscript.