The study is a two-arm randomised phase III trial. Patients will be randomised to receive capecitabine (arm A) or S-1 (arm B). Bevacizumab may be added according to the choice of the investigator. Patients will be followed 3-weekly at the outpatient clinic, toxicity will be assessed according to study protocol guidelines. Patients will be evaluated every 3 cycles for response. Upon disease progression patients will be treated according to the local investigators

Capecitabine, an oral fluoropyrimidine, has shown a comparable efficacy but a better tolerability compared to bolus 5-FU/LV. However, capecitabine has a higher incidence of hand-foot syndrome (HFS). HFS is characterized by erythema, dysesthesia and/or paresthesia of the palms of the hands or soles of feet. In advanced stage, desquamation, ulceration and blistering can occur. Although HFS is not life threatening, it can cause significant discomfort and impairment of function, especially in elderly patients. This adverse event is becoming particularly relevant since many patients may require the administration of capecitabine over prolonged periods of time.

S-1 (Teysuno®) is an oral fluoropyrimidine that has shown comparable efficacy to 5FU and capecitabine in gastrointestinal cancers but is associated with a much lower incidence of HFS. Studies on S-1 have mainly been performed in Asian patients,which population has known differences in tumour biology and toxicity compared to Western population. S-1 has shown comparable efficacy to other fluoropyrimidines as monotherapy or in combination chemotherapy schedules in several gastrointestinal tumors. However, given the lack of data from prospective studies on S-1 as monochemotherapy in metastatic colorectal cancer in Western patients, this study is designed to compare S-1 and capecitabine monotherapy in terms of safety, with particular interest in HFS, in metastatic colorectal cancer patients.

Inclusion Criteria:

• Histological proof of colorectal cancer.
• Distant metastases (patients with only local recurrence are not eligible).
• Unidimensionally measurable disease (≥1 cm on spiral CT scan or ≥2 cm on chest X-ray; liver ultrasound is not allowed). Serum CEA may not be used as a parameter for disease evaluation.
• In case of previous radiotherapy, at least one measurable lesion should be located outside the irradiated field.
• Age ≥ 18 years
• Planned treatment with fluoropyrimidine monotherapy with or without bevacizumab.
• WHO performance status 0-2 (Karnofsky PS ≥70%)
• Adequate bone marrow function (Hb ≥ 6.0 mmol/L, absolute neutrophil count ≥1.5 x 109/L, platelets ≥ 100 x 109/L), renal function (serum creatinine ≤ 1.5x ULN and creatinine clearance, Cockroft formula, ≥30 ml/min), liver function (serum bilirubin ≤ 2 x ULN, serum transaminases ≤ 3 x ULN without presence of liver metastases or ≤ 5x ULN with presence of liver metastases).
• Life expectancy > 12 weeks.
• Negative pregnancy test in women with childbearing potential.
• Expected adequacy of follow-up.
• Institutional Review Board approval.
• Written informed consent.

Exclusion Criteria:

• Prior adjuvant treatment for stage II/III colorectal cancer completed within 6 months prior to randomisation.
• Any prior adjuvant treatment after resection of distant metastases.
• Any previous systemic treatment for metastatic disease.
• History or clinical signs/symptoms of CNS metastases.
• History of a second malignancy <5 years with the exception of adequately treated carcinoma of cervix or basal/squamous cell carcinoma of skin.
• Previous intolerance of capecitabine.
• Known dihydropyrimidine dehydrogenase (DPD) deficiency or treatment within 4 weeks with DPD inhibitors, including sorivudine or its chemically related analogues such as brivudine.
• Planned radical resection of metastases after downsizing by systemic treatment.
• Significant cardiovascular disease < 1 yr before randomisation (symptomatic congestive heart failure, myocardial ischemia or infarction, unstable angina pectoris, serious uncontrolled cardiac arrhythmia, arterial thrombosis, cerebrovascular event, pulmonary embolism).
• Any significant cardiovascular events during previous fluoropyrimidine therapy.