This article describes the drugs that are in common use in metastatic colorectal cancer and discusses how healthcare providers might use the available evidence to make cost-effectiveness decisions
MB BS MRCPI MRCP(UK)
Department of Medical
GSc MB ChB MRCP PhD
David J Kerr
MA MD DSc FRCP(Glas Lond) FMedSci
Rhodes Professor of
Cancer Therapeutics and
Department of Clinical
University of Oxford
European Society of
Colorectal cancer (CRC) is one of the commonest cancers of the western world. It is associated with significant morbidity and mortality and incurs large costs to healthcare providers in terms of skilled surgery, expensive drugs and specialist medical and nursing care. Twenty percent of patients presenting with CRC will already have metastatic disease, and a further 30-40% of the remainder, despite best efforts, will eventually recur and die of their disease.
Overall five-year survival for metastatic CRC is less than 10%, with more than 80% of patients dying with hepatic metastases. Less commonly, patients can present with pulmonary metastases, bony metastases or local relapse within the abdomen. With best supportive care alone, median survival for metastatic CRC depends on disease burden and is in the range of 3-9 months.
Treatment of metastatic CRC aims to palliate symptoms, to improve quality of life and to prolong survival. A minority of selected patients with advanced disease may be potentially curable with metastatectomy, but palliative systemic therapy remains the therapeutic mainstay.
What drugs are available for metastatic colorectal cancer?
For many years, since its clinical introduction in the 1950s, 5-fluorouracil (5FU) was the only drug with demonstrable activity against colorectal cancer. However, the last decade has delivered two new cytotoxic agent, irinotecan and oxaliplatin, and three new biological agents, bevacizumab, cetuximab and panitumumab, all of which have shown some activity and have gained licences in this setting. Table 1 describes the major trials that have assessed these drugs and shows how the gradual evolution of treatment has led to improvements in response rates, median progression-free survival and expected overall median survival for metastatic CRC.
5-Fluorouracil and the oral fluoropyrimidines
5-Fluorouracil (5FU) is the backbone of most active combinations in CRC treatment. 5FU is an active thymidylate synthase inhibitor, thus preventing DNA synthesis. It also inhibits DNA synthesis by incorporating as fluorodeoxyuridine triphosphate (FdUTP) and inhibits RNA synthesis by incorporating as fluoridine triphosphate (FUTP), leading to transcription errors and arrest. When folinic acid, also known as leucovorin (LV), is administered in combination with 5FU, a stable ternary complex, fluorodeoxyuridylate (FdUMP)-TSleucovorin, is formed, which increases the inhibition of TS, and which doubles the response rate of the combination to 21%, compared with 11% for singleagent 5FU.1 The side-effects of 5FU include nausea and vomiting, soreness of the mouth (mucositis), diarrhoea, neutropenia and, rarely, chest pain believed to be a consequence of coronary artery spasm. Most of these side-effects can be managed by dose reduction; however, the angina does not appear to be doserelated, and its presence usually necessitates stopping treatment with 5FU.
Given the short half-life of 5FU (10-15 minutes), and its mechanistic focus on the S-phase of the cell cycle, it is perhaps not surprising that response rates are improved when it is given by continuous infusion when compared with bolus intravenous administration. There are obvious practical limitations with intravenous infusion of cytotoxics. Indwelling catheters can be difficult to place in some patients and are frequently associated with thrombosis and infection. However, direct conversion of 5FU to oral administration is difficult because its absorption is variable and systemic effects are unpredictable, in part due to the presence of dihydropyrimidine dehydrogenase (DPD) in the gastrointestinal mucosa, the enzyme that actively catabolises 5FU. Two strategies have attempted to overcome this problem:
* Production of a designer oral prodrug of 5FU (capecitabine) that will not be metabolised to the active drug until after absorption.
* Co-administration of inhibitors of DPD to decrease local catabolism and improve oral bioavailability (UFT).
Capecitabine is a 5FU prodrug that, after oral administration, is absorbed as a 5-deoxy-5-fluorocytidine (5-DFCR) derivative and is subsequently metabolised in the liver by carboxylesterase to 5-DFCR. 5-DFCR is converted to 5-DFUR by cytidine deaminase located in the liver and in tumour tissues, and the final metabolic step to 5FU occurs at the site of the tumour under the action of thymidine phosphorylase conferring a degree of tumour selectivity.
Two large randomised phase III studies have reported an improved overall objective tumour response rate in the first-line treatment of metastatic CRC with capecitabine compared with the common bolus regimen (Mayo clinic 5FU/LV). This improved response rate did not translate into increased progression-free or median overall survival.[2,3] However, capecitabine produced lower incidence of nausea, stomatitis, alopecia and neutropenic sepsis, and there were demonstrable health-economic benefits. There was an increased incidence of hand-foot syndrome, but this is usually managed easily by appropriate dose reductions and most clinicians would accept that capectabine is an excellent oral substitution for single-agent 5FU. It does, however, have the same cardiotoxic profile of 5U, and patients who will be administering their medication at home should be warned that they should stop treatment immediately and seek medical attention if chest pain occurs.
UFT is a mixture of uracil and tegafur. Tegafur is a prodrug of 5FU, and uracil is an inhibitor of DPD and so increases the availability of 5FU. UFT has been compared with the Mayo clinic bolus 5FU/LV regimen in phase III studies in metastatic colorectal cancer patients. These studies reported no differences between the two regimens in terms of overall tumour response rates, time to tumour progression and median overall survival.[5,6]
This thymidylate synthetase inhibitor can be used as an alternative to 5FU/LV or capecitabine. In phase III trials in the advanced disease setting, median survival with raltitrexed was comparable to that of bolus or infusional 5FU/LV. It went out of favour when its toxicity profile was questioned after a phase III trial in the adjuvant setting. Although it is given as a single short infusion once every three weeks, which makes it extremely practical, it can produce a toxic combination of diarrhoea, dehydration and neutropenia, which, if not treated promptly, can lead to death. However, for patients who cannot tolerate fluoropyrimidines because of coronary artery spasm, raltitrexed (tomudex) does appear to be a sensible alternative as it does not appear to include this particular side-effect.
Newer cytotoxic drugs for colorectal cancer
With the evolution of new drugs for CRC came the complex question of how these drugs should be combined and in what order. However, most of the clinical community would agree that the two major novel cytotoxics, oxaliplatin and irinotecan, both work most effectively when administered in combination with 5FU as first-line therapy.
Oxaliplatin is a third-generation platinum derivative. It binds to DNA and leads to DNA adduct formation, subsequently causing irreversible DNA replication error and apoptosis. As a single agent in metastatic disease it has only a 10% response rate.8 However, in combination with 5FU it demonstrates true synergy and response rates increase to 40-50%.[8,9] Two seminal phase III trials in the first-line treatment of CRC have reported, and both compared infusional 5FU/LV with (FOLFOX) or without oxaliplatin. Both described a superior overall response rate and progression-free survival for FOLFOX.[10,11] (Overall response rate of 51% and 53% vs 16% and 22% and progression-free survival of 8.7 months and 9 months vs 6.1 and 6.2 months). Based on these studies FOLFOX is now a standard first-line treatment for CRC and has been approved by the UK National Institute for Health and Clinical Excellence (NICE). However, the addition of oxaliplatin does incur the additional toxicity of neuropathy, with about one in 10 patients suffering grade 3 sensory neuropathy at cessation of treatment.
The combination of oxaliplatin and capecitabine (XELOX) has also been assessed and compared with the FOLFOX regimen in a large phase III trial randomising more than 2,000 patients. Progression-free survival was similar for the two groups – eight months for XELOX and 8.5 months for FOLFOX (HR 1.04, 97.5% CI 0.88-1.12) – as was overall survival – 19.8 months for XELOX and 19.6 months for FOLFOX (HR 0.99, 97.5% CI 0.88-1.12). Overall high-grade toxicity rates were similar, but the rates of grade 3/4 neutropenia were higher with FOLFOX, and the rates of grade 3 hand-foot syndrome and diarrhoea were greater with XELOX. These data and the results of two other phase III trials[12,13] demonstrate that 5FU and capecitabine can be used interchangeably with oxaliplatin, and the choice will often be left with the individual patient and clinician depending on the need of avoiding specific side-effects and the practicality of delivery in a particular situation.
Irinotecan (IR) is a topoisomerase I inhibitor. Topoisomerase I (an enzyme) produces single-strand breaks allowing the uncoiling of the supercoiled DNA, a requirement for DNA repair. As a single agent in the first-line setting of metastatic CRC, irinotecan produces an objective response rate of 10-20%. In second-line treatment, after 5FU failure, single-agent IR improves one-year survival rates compared with best supportive care alone (36.2% vs 13.8%; p = 0.0001), and median overall survival is improved (9.2 vs 6.5 months). Quality of life was also improved in this study. The dose-limiting toxicity of irinotecan is diarrhoea, which typically starts five to seven days after treatment. Patients can also suffer from an acute cholinergic reaction during treatment comprising bradycardia, diarrhoea and abdominal cramps. This can be effectively prevented or treated with subcutaneous atropine.
Two large phase III trials have demonstrated the superior activity of irinotecan/5FU/FA over 5FU/FA alone, with improvement of response rate, progression-free survival and overall survival.[16,17] One of these trials used infusional 5FU and one used bolus 5FU. Although both trials demonstrated improved outcome with the addition of irinotecan, infusional rather than bolus-based regimens are the most widely accepted across Europe. Although irinotecan does not incur the neuropathy induced by oxaliplatin, its overlapping toxicity of diarrhoea with 5FU means that care must be taken with this combination, and patients must be counselled about rapid referral to hospital if diarrhoea cannot be controlled by simple measures alone. This caution is even more important when irinotecan is combined with capecitabine.
Fuchs et al18 have compared the safety and efficacy of irinotecan in combination with infusional 5FU/leucovorin (FOLFIRI) or bolus FU/LV (m IFL) or capecitabine (CapeIRI) in a total of 430 patients in the first-line setting of CRC. Median progression-free survival (PFS) was 7.6 months for FOLFIRI, 5.9 months for m IFL and 5.8 months for CapeIRI (p=0.015). Median overall survival (OS) was 23.1 months for FOLFIRI, 17.6 months for m IFL and 18.9 months for CapeIRI (p = 0.27). The authors suggested that capecitabine/irinotecan should not be used in the first-line setting and they subsequently altered their trial plan accordingly.
However, Koopman et al have reported results from the CAIRO study. This phase III trial randomised 820 patients to receive either first-line treatment with capecitabine, second-line irinotecan and third-line capecitabine plus oxaliplatin (sequential treatment; n = 410) or first-line treatment with capecitabine plus irinotecan and second-line capecitabine plus oxaliplatin (combination treatment; n = 410). Median overall survival was 16.3 (95% CI 14.3-18.1) months for sequential treatment and 17.4 (15.2-19.2) months for combination treatment (p = 0.3281). The frequency of grade 3/4 toxicity over all lines of treatment did not differ significantly between the two groups, except for grade 3 hand-foot syndrome, which occurred more often with sequential treatment than with combination treatment (13% vs 7%; p = 0.004). So this group found that the combination of capecitabine/irinotecan is potentially tolerable in this population. Again the decision as to whether to use 5FU or capecitabine in combination with irinotecan will fall to the individual clinician with an individual patient, but capecitabine/irinotecan is a combination that should only be used with caution, particularly if there is a chance of incipient bowel obstruction.
In the quest for active treatments for patients who cannot tolerate fluoropyrimidines, a randomised phase II trial has compared irinotecan plus oxaliplatin (IRINOX: n = 40: Arm A) versus bi-weekly 5FU/folinic acid with irinotecan or oxaliplatin (FOLFIRI or FOLFOX: n = 20 each: Arm B). Median progression-free and overall survival times were reported as 8.4 and 19 months, respectively, in arm A, and 8.1 and 20.4 months, respectively, in arm B. So efficacy did not appear significantly different. Interestingly, diarrhoea did appear to be exaggerated in the IRINOX arm, even though oxaliplatin is not felt to be a diarrhoeagenic drug, with rates of grade 3/4 diarrhoea of 32.5% compared with 7.5% in the 5FU-containing arms. So the IRINOX combination could be considered in patients where fluoropyrimidines cannot be tolerated, although a larger phase III trial is recommended.
Trials indicate that survival of CRC patients correlates with the number of cytotoxic agents that are used during their treatment pathway. However, the most appropriate sequencing and combinations of these agents is still a contentious subject. A phase II trial combining all three cytotoxic agents (FOLFOXIRI) assessed a total of 32 patients with unresectable metastatic CRC. They received irinotecan 165 mg/m2 on day 1, oxaliplatin 85 mg/m2 on day 1, l-LV 200 mg/m2 day 1 and 5FU 3,200 mg/m2 as a 48-h infusion starting on day 1, repeated every two weeks.
This dose of oxaliplatin is the same as that used in the standard two-weekly regimen in combination with 5FU, and the dose of irinotecan is only 8% lower than the standard dose used. Grade 4 neutropenia (34%), grade 3 diarrhoea (16%), grade 3 stomatitis (6%) and grade 2-3 peripheral neurotoxicity (37%) were reported, and G-CSF was used in 23% of cycles. An overall response rate was 72%. However, in a randomised phase III trial, no difference in overall survival, median time to disease progression and response rates were reported when FOLFOXIRI was compared with FOLFIRI, and patients treated with FOLFOXIRI had significant higher incidence of alopecia, diarrhoea and neurosensory toxicity.
Given the results of the CAIRO study above, it would seem sensible that the decision about the intensity of the first-line regimen given (single agent vs double vs triple) should be made after careful consideration of the individual patient, the extent of their disease, their performance status, their renal and hepatic function, the presence or absence of incipient bowel obstruction, and the requirement to produce a maximal initial response to allow potentially curable metastatectomy. For most clinicians and most patients first-line double-agent chemotherapy (eg, FOLFOX or XELOX) followed by second-line double-agent chemotherapy (eg, FOLFIRI or cape/IR) will be the likely choice.
This philosophy is supported by the GERCOR study which assessed FOLFIRI followed by FOLFOX-6 on progression in arm A and the reverse in arm B. Both sequences achieved similar response rates and overall efficacy with median survival of 21.5 months and 20.6 months in arms A and B, respectively. The UK MRC CR08 FOCUS trial also assessed the role of irinotecan or oxaliplatin combined with infusional 5FU/FA in firstand second-line treatment. Staged single-agent, staged combination therapy or first-line combination therapy was used. Only a slight increase in overall survival was seen with each combination therapy over the staged single-agent arm, again suggesting that a sequential approach starting with single-agent fluoropyrimidine, especially in old or frail patients, is justified.
Are drug holidays justified?
Chemotherapy holiday is thought to reduce the oxaliplatin toxicity without reducing its efficacy. It has been investigated in optimal treatment duration for metastatic CRC (OPTIMOX) 1 and 2 trials. OPTIMOX 1 showed similar disease control and survival outcomes but lower toxicity with an oxaliplatin-free interval after FOLFOX compared with continuous FOLFOX chemotherapy.
OPTIMOX 2 was designed to see whether the 5FU/LV maintenance could be omitted in favour of a “chemotherapy-free interval” after FOLFOX induction. A total of 202 patients with metastatic CRC were treated with six courses of modified FOLFOX; the same regimen was reintroduced at disease progression. In the interval period patients were randomly assigned to receive either no maintenance therapy or continuous therapy minus oxaliplatin. Maintenance therapy included a mean of eight cycles. Continuous 5FU/LV after oxaliplatin-based chemotherapy showed a better progression-free survival (8.2 vs 6.7 months, p = 0.08) and longer OS (26 vs 19 months, p = 0.0549) than the chemotherapy-free interval. The duration of disease control was not significantly different between groups but favoured maintenance therapy (12 versus 9 months, p = 0.39). However, there are clinical trials of 5FU-based regimens, which suggest that a “drug holiday” does not have a negative impact on overall survival. Further trials are needed in this area.
Targeting pathways crucial to oncogenesis has been a focus of interest for the last 5-10 years. Three targeted agents are now licensed for use in colorectal cancer – bevacizumab, cetuximab and panitumumab.
Vascular endothelial growth factor (VEGF) is overexpressed in CRC and is believed to be essential in angiogenesis and metastatic evolution. Bevacizumab is a humanised monoclonal antibody that binds and inhibits the action of VEGF. Although the first phase III trial assessing the addition of bevacizumab to a bolus 5FU/IR regimen indicated an increase in median overall survival by five months, more recent trials with oxaliplatin have not borne out this profound early promise. In the first-line phase III trial reported by Cassidy et al, the improvement in progression-free survival with the addition of bevacizumab to either FOLFOX or XELOX chemotherapy was only 1.4 months (from 8.0 months to 9.4 months). Similarly, median overall survival was only marginally increased: 21.3 months versus 19.9 months, although this did reach statistical significance (p = 0.0023).27 Suggestions for causes of these less impressive results with oxaliplatin include the fact that many of these patients did not carry on with bevacizumab until progression, and the chemotherapy regimens used in this trial are more effective as a baseline standard arm combination, making incremental activity of the biological agent perhaps more difficult to achieve. NICE has not approved the use of bevacizumab in the treatment of CRC in the UK.
Cetuximab (CTX) targets the epidermal growth factor receptor (EGFR), which is overexpressed in CRC. It has some antitumour activity by itself but mainly has synergistic effects when combined with chemotherapeutic agents. In the first-line phase III CRYSTAL trial, patients received FOLFIRI or FOLFIRI plus weekly cetuximab intravenously. Median progression-free survival increased from 8.0 months to 8.9 months (p = 0.0479) with the addition of CTX, which is statistically but perhaps not clinically significant.
However, the authors went on to assess RAS mutation status of the patients’ tumours. They found that patients with mutated RAS tumours fared badly with CTX, with PFS of 7.6 months with the addition of CTX compared with 8.1 months for chemotherapy alone. In contrast, patients with wild-type RAS tumours did gain some benefit, with PFS of 9.9 months with CTX compared with 8.7 months without (p = 0.017).28 Similar results were seen in the OPUS study, where bevacizumab was added to FOLFOX chemotherapy. These results provide a prime example of personalised medicine.
This drug may well be cost-effective if it is applied to the right subpopulation based on tumour RAS testing. However, at the present time, NICE has refused approval based on the disappointing results from broad nonspecific populations.
Panitumumab is a fully human monoclonal antibody directed against the epidermal growth factor receptor (EGFR). In a phase III trial, panitumumab plus best supportive care (BSC) was compared with BSC alone in patients with metastatic CRC who had progressed after standard chemotherapy. Panitumumab significantly prolonged PFS (HR 0.54, 95% CI 0.44-0.66, p < 0.0001). Median PFS time was 8.0 weeks (95% CI 7.9-8.4) for panitumumab and 7.3 weeks (95% CI 7.1-7.7) for BSC. Mean (standard error) PFS time was 13.8 (0.8) weeks for panitumumab and 8.5 (0.5) weeks for BSC. Objective response rates also favoured panitumumab over BSC; after a 12-month minimum followup, response rates were 10% for panitumumab and 0% for BSC (p < 0.0001). No difference was observed in OS (HR 1.00, 95% CI 0.82-1.22), which was confounded by similar activity of panitumumab after 76% of BSC patients entered the crossover study.
Common toxicities include skin toxicity, hypomagnesaemia and diarrhoea. Other phase III studies also reported improved PFS and higher health-related quality of life in favour of panitumumab. Panitumumab monotherapy efficacy in metastatic CRC is confined to patients with wild-type (WT) KRAS tumours. KRAS status should be considered in selecting patients with mCRC as candidates for panitumumab monotherapy. It has been studied in a phase III study where the treatment effect on PFS in the WT KRAS group (HR 0.45, 95% CI 0.34-0.59) was significantly greater (p < 0.0001) than in the mutant group (HR 0.99, 95% CI 0.73-1.36). Median PFS in the WT KRAS group was 12.3 weeks for panitumumab and 7.3 weeks for BSC. Response rates to panitumumab were 17% and 0% for the WT and mutant groups, respectively. WT KRAS patients had longer overall survival (HR 0.67, 95% CI 0.55-0.82, treatment arms combined). Consistent with longer exposure, more grade III treatment-related toxicities occurred in the WT KRAS group. No significant differences in toxicity were observed between the WT KRAS group and the overall population.
There has been a huge expansion in the repertoire of available drugs for metastatic CRC over the last 10 years. The drive now should be towards personalised medicine, both on clinical grounds with respect to performance status and so on, to decide how many cytotoxic drugs should be combined in the first-line setting, and on molecular grounds, to decide which novel biological agents should be employed.
Without elegantly designed trials to tease out responding populations, bodies governing the dispersion of financial resources will continue to refuse to fund expensive drugs on the basis of marginal incremental benefits in broad-based populations.
1. Thirion P, et al. J Clin Oncol 2004;22:3766-75.
2. Van Cutsem E, et al. J Clin Oncol 2001;19:4097-106.
3. Hoff PM, et al. J Clin Oncol 2001;19:2282-92.
4. Twelves C, et al. Eur J Cancer 2001 Mar;37(5):597-60.
5. Douillard JY, et al. J Clin Oncol 2002;20:3605-15.
6. Carmicheal J, et al. J Clin Oncol 2002;20:3617-25.
7. Papova I, et al. Eur J Cancer 2008. doi:10.1016/ j.ejca.2008.07.002.
8. Misset JL, et al. Crit Rev Oncol Hematol 2000;35:75-93.
9. Bleiberg H, et al. Semin Oncol 1998;25(5):32-9.
10. de Gramont A, et al. J Clin Oncol 2000;18:2938-47.
11. Ravaioli A, et al. J Clin Oncol 2002;20:2545-50.
12. az-Rubio E, et al. J Clin Oncol 2007;25:4224-30.
13. Porschen R, et al. J Clin Oncol 2007;25:4217-23.
14. Conti JA, et al. J Clin Oncol 1996;14:709-15.
15. Cunningham D, et al. Lancet 1998;352:1413-8.
16. Saltz LB, N Engl J Med 2000;343:905-14.
17. Douillard JY, et al. Lancet 2000;355:1041-7.
18. Fuchs CS, et al. J Clin Oncol 2007;25:4779-86.
19. Becouarn Y, et al. Ann Oncol 2007;18:2000-5.
20. Masi G, et al. Ann Oncol 2004;15:1766-72.
21. Souglakos J, et al. Br J Cancer 2006;94: 798-805.
22. Tournigand C, et al. J Clin Oncol 2004;22:229-37.
23. Seymour MT. J Clin Oncol ASCO Annu Meet Proc 2005;23:(Abstract 3518, p 250s).
24. Maindrault-Goebel F. Proc ASCO 2006, 24 (Suppl 18S), Abstr. 3504.
25. Maughan TS. Lancet 2003;361:457-64.
26. Cassidy J, et al. J Clin Oncol 2008;26:2006-12.
27. Saltz LB, et al. J Clin Oncol 2008;26:2013-9.
28. Van Cutsem, et al. J Clin Oncol 2008;26:(May 20 Suppl; Abstr 2).
29. Bokemeyer C. J Clin Oncol 2008;26(May 20 Suppl; Abstr 4000)
30. Van Cutsem E, et al. J Clin Oncol 2007;25:1658-64.
31. Siena S, et al. Br J Cancer 2007;97:1469-74.
32. Amado RG. J Clin Oncol 2008;26:1626-34.
33. Saltz LB, et al. N Engl J Med 2000;343:905-14.
34. Douillard JY, et al. Lancet 2000;355:1041-7.
35. Giacchetti S, et al. J Clin Oncol 2000;18:136-47.
36. deGramont A, et al. J Clin Oncol 2000;18:2938-47.
37. Goldberg RM, et al. Proc Am Soc Clin Oncol 2002;21:A511.
38. Kufe DW, et al. Cancer Medicine. 6th ed. BC Decker; 2003.