The eligibility criteria are as follows: histologically confirmed adenocarcinoma of the stomach; previous treatment with other chemotherapy regimens; bidimensionally measurable lesions; no central nervous system metastases; no active infections; no serious or uncontrolled concurrent medical illnesses; no history of other malignancies; an Eastern Cooperative Oncology Group (ECOG) performance status of 0~2; and an age between 18 and 75 years. The institutional review board approved the protocol and written informed consent was obtained from each patient prior to study.

On day 1, oxaliplatin (85 mg/m²) was administered via an intravenous (i.v.) infusion in 500 ml normal saline or dextrose over a 2-hour period. On days 1 and 2, LV (20 mg/m²) was administered as an iv bolus, which was immediately followed by 5-FU (400 mg/m²) that was given as a 10 minute iv bolus, and 5-FU (600 mg/m²) as a continuous 22 hour infusion with a light shield. Dose modifications to oxaliplatin or 5-FU were made for patients showing hematological, gastrointestinal, or neurological toxicity based on the most severe grade of toxicity that was encountered during the previous cycle. The patients were assessed before the beginning of each 2-week cycle using the National Cancer Institute-Common Toxicity Criteria (NCI-CTC), with the exception in the case of neurotoxicity. In this case, the following oxaliplatin-specific scale was used: grade 1, paresthesias or dysesthesias of short duration, but resolving before the next cycle; grade 2, paresthesias persisting between cycles (2 weeks); and grade 3, paresthesias interfering with function. The treatment was delayed for up to 2 weeks for the following reasons: if the symptomatic toxicity persisted; if absolute number of neutrophils was <1,500/µl; or if the platelet count was <100,000/µl. The 5-FU dose was reduced by 25% for the subsequent courses if NCI-CTC grade 3 diarrhea, stomatitis, or dermatitis occurred. The dose of oxaliplatin was reduced by 25% in subsequent cycles if persistent paresthesias between the cycles or paresthesias with functional impairment lasting more than 7 days were encountered. The treatment was continued until there were signs of disease progression, the development of unacceptable toxic effects, or the patient refused further treatment.

Before each treatment course, a physical examination, routine hematology studies, chemistry, and chest X-ray were performed. The serum carcinoembryonic antigen (CEA) levels were determined after each cycle. CT scans were performed to determine the extent of the disease and the response after 4 cycles of chemotherapy, or sooner if there was evidence of any clinical deterioration.

The response was assessed using the WHO criteria (20). A complete response (CR) was defined as the disappearance of all evidence of the disease and the normalization of the tumor markers for a period of at least 4 weeks. A partial response (PR) was defined as a ≥ 50% decrease in the bidimensional tumor measurements, without the appearance of any new lesions or the progression of any existing lesion. Progressive disease (PD) was defined as any of the following: (1) a ≥ 25% increase in the sum of the products of all the measurable lesions, (2) the appearance of any new lesion, or (3) the reappearance of any lesion that had previously disappeared. Stable disease (SD) was defined as a tumor response that did not meet the criteria for a CR, PR, or PD.

The dose intensity (mg/m²/week) was calculated as the total cumulative dose divided by the treatment duration. The relative dose-intensity (RDI) was calculated by dividing the dose-intensity by the planned dose-intensity, and multiplying the result by 100. The planned dose-intensities, which is expressed as milligrams per square meter per week, were 1,000 for 5-FU and 42.5 for oxaliplatin.

This trial was designed to detect a response rate of 25% compared with a minimal, clinically meaningful response rate of 5%. The two-stage optimal design proposed by Simon was used for this trial (21), with a statistical power of 90% for accepting the hypothesis and a 5% significance for rejecting the hypothesis. The total sample size required allowing for a follow-up loss rate of up to 10% was 33 patients with measurable disease. The variables included in this study were gender, age, previous adjuvant chemotherapy, and the baseline carcinoembryonic antigen (CEA) levels. The response rates with respect to the different variables were compared using Fisher's exact test. The time to progression and the overall survival were calculated using the Kaplan-Meier method. The time to progression was calculated from the date therapy had begun to the date of disease progression, and the overall survival was calculated from the date the therapy was begun to the date of death. All the data were analyzed using SPSS software (version 10.0, Chicago-IL)

Between December 2003 and December 2004, a total of 33 patients were treated at the department of Internal Medicine, Dong-A University Medical Center, Busan, Korea. Three patients were either to follow-up or had fewer than 2 cycles of chemotherapy. Therefore 30 patients were evaluated for their response and 33 patients were evaluated for their toxicity.

Of the 30 patients who could be evaluated for their tumor response, 8 achieved a partial response, with an overall response rate of 26.7% (95% confidence interval (CI): 20.5~32.7%), 15 (50%) showed stable disease, and 7 (23.3%) progressed during the course of treatment. The patients' age, gender, previous regimen, CEA level before chemotherapy, and the site of metastasis did not affect the response rate (Table 1). The median follow-up duration was 7.1 months, and the time to progression and overall survival were 3.5 months (95% CI: 2.6~4.4 months)(Fig. 1) and 7.9 months (95% CI: 5.9~9.9 months), respectively (Fig. 2). Docetaxel with cisplatin or irinotecan with 5-FU were administered to the 26 patients who progressed. Of these, only one patient achieved stable disease and others had progressive disease.

A total of 178 chemotherapy cycles were administered to the 33 patients. The median number of cycles was six ranging from 1 to 9. Table 2 summarizes the incidence of hematological toxicity and non-hematological toxicity. The major grade 3/4 hematological toxicity encountered included neutropenia (45.4%) and thrombocytopenia (3.0%). There were only 2 cycles of neutropenic fever. The most common non-hematological toxicity encountered was grade 1/2 nausea/vomiting (18.2%), diarrhea (12.1%), and neuropathy (15.2%). There was no treatment interruption or discontinuation for peripheral neuropathy. The relative dose intensities of oxaliplatin and 5-FU were 94% and 83% of the planned doses. There were no treatment related deaths. Most cycle delays (58 of 178) were due to treatment-related toxicity, and 10.6% (19 of 178) were for personal convenience. Of the 33 patients who received more than one cycle, 13 (39%) underwent at least one dose reduction among a total of 58 cycles (33%).