Metastatic colorectal cancer is a particularly frequent and severe cancer. Patients die mainly from metastatic disease; however, the survival of these patients has dramatically improved with the progress in chemotherapeutic regimens as new routes of administration and introduction of more potent cytotoxic agents administered in sequential 5-FU-folinic acid-irinotecan/5-FU-folinic acid-oxaliplatine strategies. Biologic therapies have been also developed targeting two different pathways, angiogenesis and the epidermal growth factor receptor. Their combination with chemotherapy leads to improved progression-free survival and overall survival in some cases as the addition of cetuximab in wild-type K-Ras tumors. The objectives of this expert conference were to review the different options, the available prognostic or predictive factors to optimally guide the treatment.
Colorectal cancer (CRC) is a particularly frequent cancer with 1 200 000 new cases each year worldwide  and a poor survival with only 55% of patients alive 5 years after diagnosis, in Western countries. Patients die mainly from metastatic disease; however, the survival of these patients with metastatic colorectal cancer (mCRC) has dramatically improved over the last decade. The median overall survival (OS), which was <6 months with best supportive care, reached 10–12 months with bolus 5-fluorouracil (5-FU) and up to 14 months with the addition of leucovorin (LV) in the bolus and infusional LV5FU2 regimen. With the introduction of the sequential 5-FU-folinic acid-irinotecan (FOLFIRI)/5-FU-folinic acid-oxaliplatine (FOLFOX) regimens, median OS reached up to 20 months as demonstrated by Tournigand et al.  in 2004. During this period, better understanding of tumor biology led to the development of biologic therapies targeting two different mechanisms, angiogenesis (bevacizumab) and epidermal growth factor receptors (EGFRs) (cetuximab and panitumumab).
In 2004, two major publications demonstrated the relevance of these approaches. The addition of bevacizumab to a bolus irinotecan-based chemotherapy showed promising efficacy in improving response rate, progression-free survival (PFS) and OS . However, subsequent reports have been less positive when bevacizumab was combined with oxaliplatin-based regimens, such as XELOX (capecitabine-oxaliplatine) and FOLFOX, although the primary end point of the study (PFS) was reached . Nevertheless, bevacizumab in combination with oxaliplatin-based chemotherapy after irinotecan failure has been demonstrated to improve OS .
The combination of cetuximab and irinotecan in progressive disease after previous irinotecan-based chemotherapy has demonstrated high response rate and significantly prolonged time to progression, offering a new option in the second-line or an even more refractory setting . Finally, cetuximab when used in first-line treatment in combination with either oxaliplatin- or irinotecan-based regimens also significantly improved response rate, PFS and OS in patients with wild-type K-Ras tumors [7, 8].
Consequently, we have now various regimens in our therapeutic armamentarium, but we do not know yet how to choose the best option among these strategies to optimally treat patients. The objectives of this expert conference were to address challenging queries and primarily whether treatment can be standardized for all patients. If not, what can drive decisions? Which prognostic or predictive factors should be used to guide treatment?
The conference was comprised of a panel of seven experts chosen because of their knowledge in the field and their global recognition. Various European countries (France, The Netherlands, Spain and the UK) and the United States were represented. The panelists reviewed and discussed the most relevant data from published trials and addressed the following topics:
The changes from the use 5-FU alone to the era of targeted therapies, with special emphasis on patients with isolated liver metastases.
The relevance of prognostic or predictive markers.
The contribution of targeted therapies in general and in patients with isolated liver metastases.
The possibility of defining guidelines to treat patients on the basis of an optimal use of all available agents.
The final manuscript was reviewed and approved by all panel members.
the changes from the use 5-FU alone to the era of targeted therapies
Since the introduction of 5-FU, progress was first achieved by optimizing its route of administration and by adding LV, leading to bolus regimens and the bolus and infusional regimen LV5FU2. The next step was the addition of irinotecan or oxaliplatin to 5-FU/LV, ultimately using all drugs sequentially, aiming for the ‘concept of all three drugs’ being available and offered to everyone through lines of treatment.
Lastly, the discovery of EGFR or vascular endothelial growth factor (VEGF) inhibitors allowed the development of biochemotherapy combinations. Although using each class of agents, either alone or with chemotherapy has generally met with success, preliminary clinical trials combining these bioagents together indicate possible negative interaction and these combinations are therefore not warranted in practice (Panitumumab Advanced Colorectal Cancer Evaluation  and Capecitabine-Irinotecan-Oxaloplatine-2 ).
Several clinical trials were designed to demonstrate the feasibility of using an oral fluoropyrimidine and capecitabine instead of i.v. 5-FU. Among them:
The Three Regimens of Eloxatin Evaluation trial designed to compare the safety and efficacy of three oxaliplatin plus fluoropyrimidine regimens (modified FOLFOX, bolus 5-FU plus oxaliplatin and capecitabine–oxaliplatin combination) first indicated that such option was acceptable .
A meta-analysis of six phase III trials comparing the efficacy of capecitabine (3097 patients) and i.v. 5-FU (3074 patients) in mCRC and gastric cancer supported the conclusion that capecitabine is not inferior to 5-FU. In the CRC population, PFS was not statistically different from 5-FU regimens, leading to the conclusion that oral capecitabine is at least as effective as i.v. 5-FU .
The introduction of oxaliplatin has raised the issue of cumulative neuropathy. To reduce this toxicity, treatment can be administered intermittently. The options are stopping oxaliplatin and continuing with the fluoropyrimidine only (‘chemo-light’ concept) as in the OPTIMOX1  and CONcePT  trials or to stop all drugs, as in the OPTIMOX2 study . With the introduction of biologic agents, the question of stopping chemotherapy while continuing the biologic agents as maintenance therapy is being addressed by the Double inhibition, Reintroduction, Avastin study.
Another potential way to reduce neuropathy is to use preventive agents. The N04C7 and the CONcePT phase III randomized trials demonstrated the efficacy of i.v. calcium (Ca) and magnesium (Mg) as a neuroprotectant when added to FOLFOX [18, 19]. Due to negligible toxicity, low cost and lack of interference with chemotherapy, the i.v. administration of Ca and Mg can be considered as an option when administering oxaliplatin-based therapy.
The last question at this stage is the selection of the optimal first-line regimen for a given patient. First-line treatment should be decided on the basis of objectives we want to achieve for the patient. Therefore, the first visit should determine whether the patient is symptomatic or not, presents with a low volume and indolent tumor or high volume and aggressive disease. In all cases, the question of potential surgical cure, either up front or after primary chemotherapy, has to be raised. Accordingly, the strategy should be defined within collaboration in a multidisciplinary team including gastroenterologists, medical oncologists, surgeons and radiologists.
In the case of a patient presenting with an asymptomatic indolent disease, with no option of further resectability, there is no absolute need to get an early or complete response. At this point, age, comorbidities, convenience, quality of life, patient preference and treatment cost should be seriously considered. In these patients, treatment options are either single-agent fluoropyrimidine or combination chemotherapy, either alone or in combination with biologic agents. Intermittent treatment may be recommended as these patients are those who benefit the most from this strategy.
When the patient presents with symptomatic and aggressive disease, rapid tumor shrinkage can relieve symptoms. To achieve this goal, the best option is to propose a multidrug therapy up front (FOLFOX or FOLFIRI and, less frequently, FOLFOXIRI or 5-Fu-Folinic acid-Irinotecan-Oxaliplatin) with, when possible, the addition of biologic agents, which may increase response rate, prolong PFS and symptom-free survival aiming for a better quality of life.
For patients with potentially curative disease by surgical intervention, in case of initially resectable disease, perioperative (neoadjuvant and adjuvant) combination chemotherapy may be considered. This strategy has been well defined as perioperative treatment, with the paramount European Organisation for Research and Treatment of Cancer (EORTC) study.
If initial disease is unresectable, the recommended treatment is similar to that used in patients with aggressive disease. The concept of keeping drugs for later recurrence is less warranted than in patients with incurable metastatic disease.
In conclusion, the choices for first-line treatment in mCRC have become increasingly complex. Standard chemotherapy regimens are unlikely to vary in the short term and progress is more likely to emerge from predictive biomarkers development as this could influence the choice for biologic therapies. Prediction of treatment benefit and tolerability requests newer measures, algorithms and prognostic factors, beyond performance status and laboratory values. However, it merges that further resectability is a new objective in the decision-making process.
patients with isolated liver metastases
Patients with mCRC presenting with initial liver metastases account for 50% of cases; half of them have isolated liver metastases. Treatment and survival vary tremendously according to whether the disease is resectable (10%–20% of cases) or not (80%–90%) at baseline. Only patients undergoing resection may expect an OS up to 40% at 5 years and 26% at 10 years , while patients with unresectable disease have a very poor 5-year OS.
The first option to improve OS in this patient population is to expand surgery indications. Patients with more than three nodules may benefit from surgery with a long-term survival reaching >28% at 5 years and 13% at 10 years. Patients with large metastases (≥50 mm) can also benefit from resection with a survival reaching 33% at 5 years and 21% at 10 years . More recently, some data have demonstrated that even patients with R1 resection may have the same survival that those with R0 resection with 57% versus 61% at 5 years and 37% versus 43% at 10 years, respectively, provided that perioperative chemotherapy was combined with surgery .
Another way to improve resectability comes from specific techniques like portal vein embolization, the combination of hemiliver resection with and radiofrequency of nonresectable controlateral lesions or a two-stage hepatectomy.
In case of unresectable metastases, primary chemotherapy is the first option. Optimized chemotherapy regimens have led to an increase in the chance of response and resectability and consequently OS over the last decade [22, 23].
Primary chemotherapy for patients with unresectable disease may allow at least 15%–30% of patients to become candidates for optimal resection . With this downsizing chemotherapy, OS has been greatly optimized for patients who benefit from this strategy . There is a clear relationship between response rate to chemotherapy and the resection rate . Resectability should therefore be an end point for any strategy in mCRC. The more effective the chemotherapy, the better is the chance for surgery. The combination of all these strategies can optimize resectability in up to 50% of patients with liver disease only.
In patients with primary resectable disease, neoadjuvant and adjuvant chemotherapy approaches have proven beneficial. The recently published Eloxatin® for Peri-Operative Use (EPOC) prospective randomized trial conducted by EORTC showed that 12 cycles of perioperative FOLFOX4 used as neoadjuvant (six cycles) and adjuvant (six cycles) therapy increased the PFS of resected patients with a statistically significant benefit of 9.2% at 3 years in those patients who underwent resection . The CRUK (Cancer Research UK) new EPOC study proposes to compare two treatment arms containing FOLFOX with or without cetuximab in neoadjuvant and adjuvant treatments.
The timing between chemotherapy and surgery is a key parameter for optimal outcome of patients. The longer chemotherapy is administered and the higher the number of treatment lines, the lower is the survival after resection . Moreover, looking for complete clinical response does not translate necessarily into pathological complete response and leads to the risk of tumor progression after an initial response ; in addition, it may increase difficulties for the surgeon to detect missing lesions and the risk of postoperative complications related to hepatotoxicity of prolonged chemotherapy [27, 28]. The determination of the good therapeutic window, as soon as the metastases become resectable, requires optimal collaboration between the oncologist and the surgeon.
The ‘Oncosurgery’ approach is becoming a standard of care irrespective of primary resectability and may allow a long-term benefit in selected patients with an expected OS of ∼30%–40% at 5 years.
the relevance of prognostic or predictive markers in CRC
Valuable biomarkers should be on the basis of scientific evidence, reproduced in different tumor types and understood mechanistically. Their measure should be reproducible with high sensitivity and specificity and should have a clinically relevant impact on treatment.
Progress attributed to biomarkers is likely related to success in drug development and registration.
K-Ras: a determinant factor in CRC.
To date, no biomarker demonstrated a predictive impact on antiangiogenic treatment outcome. On the contrary, in the EGFR signaling pathway, K-Ras is a key central downstream effector. K-Ras gene status, wild type versus mutated, highly influences the efficacy of the two drugs, cetuximab and panitumumab, targeting EGFR.
Most patients (up to 60%) present with a wild-type K-Ras tumor. The K-Ras mutation confers a modified conformation to the K-Ras protein that prevents its inactivation .
implications for K-Ras status and EGFR-targeted therapies in mCRC.
EGFR inhibitors do not affect the pathway in case of a K-Ras mutation . This resistance was indicated in single-arm studies retrospectively analyzed with regard to K-Ras status. Basically, no response to cetuximab or panitumumab was noted in patients with mutated K-Ras tumors [31–33]. The same conclusion was made in randomized trials comparing a treatment with cetuximab or panitumumab to best supportive care [34, 35]. In these studies, EGFR inhibition offered a modest benefit in term of PFS in the general patient population, but the results appeared much better in patients with a wild-type tumor [36, 37]. A benefit in OS was observed in the cetuximab noncrossover study. In first line, the randomized Cetuximab Combined with Irinotecan in First-Line Therapy for Metastatic Colorectal Cancer (CRYSTAL) and Oxaliplatin and Cetuximab in First-Line Treatment of mCRC (OPUS) trials conducted with cetuximab demonstrated the same outcomes with large differences between wild-type and mutated tumors, particularly in response rate, PFS and OS [7, 8]. The meta-analysis results of these two studies strengthen the findings obtained from the CRYSTAL and OPUS patient populations with K-Ras wild-type tumors. Adding cetuximab to chemotherapy significantly reduces the risk of disease progression by 34% (HR 0.66, 95% CI 0.55–0.80, P < 0.0001), increases the chance of response by greater than twofold (odds ratio 2.16, 95% CI 1.64–2.86, P < 0.0001) and improves OS by 19% (HR 0.81, 95% CI 0.69–0.94, P = 0.006) in the first-line treatment of mCRC . Moreover, the addition of cetuximab to chemotherapy did not affect quality of life in patients with K-Ras wild-type tumors .
A prospective single-arm study not only reached the same conclusions on outcomes with regard to K-Ras mutation status but also demonstrated that other biomarkers could predict the response to cetuximab, such as the ligands amphiregulin and epiregulin [40, 41]. Interestingly, a signature of genes could more accurately differentiate patients within the wild-type population who could benefit from cetuximab. Some of these genes correlate with genes found in series drawn by other investigators .
Other pathways may be activated in advanced CRC like c-MET, platelet derived growth factor receptor and insulin-like growth factor 1-receptor that bypass the EGFR pathway. PTEN deregulation, PI3K mutations or B-Raf mutations may also play a role in the quality of response to cetuximab [43–45]. New compounds targeting these specific pathways are needed, especially in mutated K-Ras tumors.
Although progress has been made by recent research on biomolecular and clinical development, especially with the targeted therapies, further development should help to drive therapeutic decisions on the basis of new specific predictive biomarkers.
the contribution of targeted therapies
A number of targets play an important role in the tumor cell development. In CRC, several targeted therapies are still under clinical investigation, but two main pathways have been already successfully targeted, VEGF with antiangiogenic compounds (bevacizumab) and the epidermal growth factor through its receptor inhibition (cetuximab, panitumumab).
Three important trials have demonstrated that the addition of bevacizumab to chemotherapy in the first-line setting increases the activity without significantly worsening the safety profile of the cytotoxic drugs [3, 4, 46]. The addition of bevacizumab to irinotecan-based chemotherapy has shown promising efficacy with a median PFS up to 10 months and a median OS of 20+ months . However, more recent data have shown less significant differences when bevacizumab was combined to oxaliplatin-based regimens . Some data from this study indicated that in first-line therapy, bevacizumab should not be stopped before progression. Other studies have indicated that bevacizumab should be continued after progression in combination with another chemotherapeutic regimen . Although this observation came from nonrandomized registrational studies, prospective randomized trials are ongoing.
Well-known biomarkers, such as K-Ras, B-Raf and p53 mutation status do not influence the outcome of patients treated with a bevacizumab-based treatment.
Bevacizumab has specific side-effects like hypertension, proteinuria, thromboembolic events especially arterial, bleeding that could be minor like epistaxis or major with hemorrhage, gastrointestinal perforation, wound healing and postoperative bleeding. These effects are rarely severe, as demonstrated in the Bevacizumab Expanded Access Trial (BEAT) extended access program . Risk factors for arterial thromboembolic events include history of prior arterial thromboembolic events (stroke or heart attack) and an age of ≥65 years .
The human EGFR signaling pathway plays a key role in tumor growth and progression in numerous cancers. In CRC, EGFR signaling is deregulated. HER-1/EGFR overexpression correlates with disease progression, poor prognosis and reduced sensitivity to chemotherapy. Several agents target EGFR, including small-molecule tyrosine kinase inhibitors and mAbs. Cetuximab is the studied mAb in mCRC, already registered in both first- and second-line settings. Panitumumab was recently also approved, but its use is limited, as a single agent, when previous therapeutic options have failed.
In the first-line studies with cetuximab, CRYSTAL and OPUS, clinical benefit was observed for patients treated in first line; moreover, the outcome was strongly correlated with the K-Ras status [7, 8].
Testing for multiple other indicators (PTEN loss, EGFR ligands, PI3K mutations, B-Raf mutations, EGFR gene copy number, dual-specificity phosphatases dual specificity phosphatases …) may increase predictive power for resistance or response to treatment. The EGFR ligands, epiregulin and amphiregulin, may characterize EGFR-dependent tumors which could be more sensitive to cetuximab [40–42]. These new predictive factors are awaiting validation in clinical trials.
Definitely, these drugs and several other new compounds will soon become available, leading to further potential survival increase in CRC.
the contribution of targeted therapies in isolated liver metastases
The chemotherapy regimens FOLFOX or FOLFIRI, when combined with cetuximab, can result in >50% response rates in wild-type K-Ras tumors. Three main randomized clinical trials ColorEctal LIver Metastases (CELIM), OPUS and CRYSTAL have evaluated the resection rate of initially unresectable patients [7, 8, 50]. The CELIM study compared two treatment arms both containing cetuximab combined with FOLFIRI or FOLFOX6 [50, 51]. After eight cycles (4 months), in technically unresectable disease, treatment was continued for four further cycles. In that study, the response rate reached almost 80% allowing surgery in 43% of the patients among whom 34% had R0 resection. In the CRYSTAL and OPUS studies, the addition of cetuximab to cytotoxic backbone led also to an increased resection rate, although the numbers of resected remained low.
In the case of nonresponse to primary chemotherapy either FOLFOX or FOLFIRI, the addition of cetuximab in second line can increase by 50% the number of patients converted to resectability .
Other well-conducted trials, but with modest number of patients, have tested different chemotherapies with targeted agents. Most of them failed to demonstrate an enhanced resection rate, pointing out that the secondary resectability should be the first end point.
Bevacizumab in combination has been studied in few trials. Demonstration of its safety in the neoadjuvant setting was reported in a phase II study .
Moreover, the aim of the first BEAT (MO18024) observational study was to provide safety experience for physicians dealing with advanced unresectable CRC. The results showed that 9.4% of patients with multiple metastatic sites underwent surgery; liver resection was possible in 12% with 10% R0 resections [48, 54]. A phase II study is ongoing with R0 resection rate as primary end point.
how to optimally treat mCRC patients?
With multiple options and complex decisions and end points, trying to define guidelines/recommendations needs first a careful analysis of each patient/tumor characteristics and a precise definition of the treatment goals as highlighted in the discussion.
The first question to be addressed is the possibility or not to consider cure as a potentially achievable objective. In other words, would it be possible or not to propose primary or secondary surgery but not only to patients presenting with isolated liver metastases?
If the answer is YES, then the up-front treatment will have to provide, as quickly as possible, maximum tumor shrinkage without delaying surgery. This can be optimally achieved using a biochemotherapy combination. On the basis of available published large randomized trials, it seems likely that the best chemotherapy component is certainly an irinotecan- or oxaliplatin-based regimen with cetuximab in patients with wild-type K-Ras tumors. It is then logical to recommend K-Ras status determination for these patients.
Even if the answer is NO, the choice of the up-front treatment is key, as it will have a paramount impact on the subsequent options. This is true, not only regarding the chemotherapy component but also regarding the biologic one. As a matter of fact, the data so far generated with bevacizumab in the first-line setting demonstrate that the expected clinical benefits vary with the combined chemotherapy regimen and as a consequence, this biologic agent cannot be considered as a standard for all patients. Here also, K-Ras status determination is a key decision-making tool allowing the possibility to consider cetuximab as a component of the up-front regimen in patients with wild-type K-Ras tumor rather than bevacizumab.
As to the first question being addressed, it is then possible to consider:
the use of 5-FU or an oral fluoropyrimidine;
continuous versus intermittent treatment;
neuroprotective measures when using oxaliplatin …
The main panel conclusions were
Not all patients should be treated the same, there is no standard treatment primarily in the up-front setting and hope for cure has to be the objective at the introduction of treatment in each patient.
mCRC is still an area under investigation and today's dogmas may be obsolete tomorrow. Nevertheless, to offer, on a day-to-day basis, our patients the best possible chance to stay alive and to get a chance for cure, we need to use for all, the tools that are already available.
A multidisciplinary management of the patients from the onset of their disease is a prerequisite condition to offer them the best strategy and the optimal timing.
Merck-Serono to R.A.; Roche, Merck-Serono, Amgen to E.V.C.
RA received lecture fees from Merck-Serono, Roche and Pfizer. JLR received lecture fees from Merck. JT received advisory board fees from Amgen, Merck-Serono, Roche and Imclone. GP has no conflict of interest. DH and JPS have given no information about conflict of interest.
The authors thank Anne Ponzio-Prion, Institut Gustave Roussy, France, for her assistance in this publication.
- chemotherapy regimen
- biological therapy
- drug administration routes
- neoplasm metastasis
- receptor, epidermal growth factor
- liver metastases
- k-ras oncogene
- colorectal cancer metastatic
- predictor variable
- molecular targeted therapy
- progression-free survival