The role of surgical resection in the subset of patients with resectable hepatic metastases converted from initially non-resectable liver metastasis was still not clearly established. To further explore the oncologic results of surgical versus non-surgical methods for the treatment of this subset of patients, we designed and conducted the present randomized prospective study beginning in 2002. The present study was based on the following arguments against the predominant survival benefits of surgical resection in previous reported series: (1) The initially non-resectable liver metastasis was basically a disseminated disease, even though some metastases were highly responsive to chemotherapy and become resectable; (2) Since the evaluation of resectability was based on the imaging studies, it was difficult to consider the surgical resection as "curative" for the resectable hepatic metastases converted from non-resectable ones, given the limitation of the current imaging stools of high-technology; (3) The resectable hepatic metastases after neoadjuvant chemotherapy might represent a subset of hepatic metastases biologically highly responsive to chemotherapy and the time-to-progression for these metastases might be fairly long after a response. Additionally, these metastases might be also biologically responsive to other cytotoxic or targeted therapies that justified the patients' continuous adoption of non-surgical treatment.

In 2000, de Gramont et al. and Giacchetti et al. reported that combination chemotherapy including modulated infusional 5-Fu plus irinotecan and oxaliplatin could achieve a response rate of approximately 50% and a median survival of in excess of 20 months. These encouraging results implied that chemotherapy is likely to play an increasingly important role in decreasing the size of metastasis to allow for liver resection. Interestingly, Adam and Bismuth et al. reported that the 5-year overall survival of 50% (95% confidence interval 33-68%), observed in patients with resection following neoadjuvant chemotherapy, was comparable to patients with primarily resectable hepatic metastasis as reported by Scheele et al. (39%, 469 patients), Fong et al. (37%, 1001 patients), Nordlinger et al.(28%, 1568 patients), and Fiqueras et al. (36%, 235 patients). Recently, Zelek et al. reported that intravenous and hepatic artery infusion of irinotecan/ 5-fluorouracil/ leucovorin could facilitate complete resection of initially non-resectable hepatic metastases. Pozzo et al. indicated that neoadjuvant treatment of unresectable liver disease with irinotecan/ 5-fluorouracil/ leucovorin enabled a significant portion of patients to undergo surgical resection. Alberts et al. showed that FOLFOX-4 therapy allowed for successful resection of disease in a portion of patients initially not judged to be optimally resectable but with a high recurrence rate after surgery. Masi et al. reported that neoadjuvant approach with irinotecan plus FOLFOX-4 had significant antitumor activity and provided a radical surgical resection of initially unresctable hepatic colorectal metastases with promising median survival of patients. However, most studies published till now are retrospective analyses without a control group or case series with limited patient number and/or short time of follow-up. Therefore, the role of surgical resection in the subset of patients with resectable hepatic metastases converted from initially non-resectable liver metastasis was still not clearly established. To further explore the oncologic results of surgical versus non-surgical methods for the treatment of this subset of patients, we designed and conducted the present randomized prospective study beginning in 2002. The present study was based on the following arguments against the predominant survival benefits of surgical resection in previous reported series: (1) The initially non-resectable liver metastasis was basically a disseminated disease, even though some metastases were highly responsive to chemotherapy and become resectable; (2) Since the evaluation of resectability was based on the imaging studies, it was difficult to consider the surgical resection as "curative" for the resectable hepatic metastases converted from non-resectable ones, given the limitation of the current imaging stools of high-technology; (3) The resectable hepatic metastases after neoadjuvant chemotherapy might represent a subset of hepatic metastases biologically highly responsive to chemotherapy and the time-to-progression for these metastases might be fairly long after a response. Additionally, these metastases might be also biologically responsive to other cytotoxic or targeted therapies that justified the patients' continuous adoption of non-surgical treatment.

Inclusion Criteria:

1. Secondary resectable hepatic metastasis converted by FOLFOX-4 regimen.
2. age between 18-80 y/o.
3. no extra-hepatic disease
4. life expectancy ≥ 3 months
5. Karnofsky performance status ≥ 50%.
6. WBC count ≥ 4,000/μl, platelet count ≥ 100,000/μl, serum bilirubin ≤ 2.0 mg/dl and normal serum glucose and electrolyte levels.-

Exclusion Criteria:

1. Secondary resectable hepatic metastasis with extra-hepatic disease
2. poor liver function including: serum bilirubin > 2.0 mg/dl, GOT,GPT >100 U/L
3. severe steatosis hepatitis or sinusoidal dilation
4. Karnoofsky performance status <50%.-