The eligibility criteria included histologically or cytologically proven, unresectable, locally advanced stage IIIB and/or stage IV NSCLC. No prior palliative chemotherapy was permitted for the patients with advanced disease. Recurrent disease was defined as evident tumor progression after surgical resection or radiation treatment. The patients in this study were required to be at least 18 years of age, have a Karnofsky performance status (PS) of 80% or more and a life expectancy of at least 12 weeks, and the presence of at least one uni-dimensionally measurable lesion was mandatory. Additional eligibility criteria included adequate bone marrow reserves (a white blood cell count ≥3×10⁹/L, platelets ≥100×10⁹/L, hemoglobin ≥100g/L and a hematocrit ≥30%), and also adequate liver and renal function (creatinine <1.5 times the upper limit of normal).

Patients were excluded if they had any active infection, any uncontrolled central nervous system metastasis requiring emergency radiotherapy and/or corticosteroids, serious concomitant systemic disorders, a second primary malignancy (except in situ carcinoma of the cervix or non-melanomatous skin cancers) or any severe cardiovascular disease. Patients with peripheral neuropathy grade 2, according to the National Cancer Institute Common Toxicity Criteria (NCI-CTC) Version 2.0, were also excluded from the study. Any patients who were pregnant or breast-feeding were excluded from the study.

The patients were randomized to receive either the DC or EC chemotherapeutic regimen by using a randomized drawing from an envelope, and these regimens were randomly assigned to the patients at a 1:1 ratio. In the DC arm, the patients received docetaxel 75 mg/m², and this was immediately followed by cisplatin 75 mg/m²; both drugs were administered as 1-hour intravenous (i.v.) infusions on day 1. Oral dexamethasone 8 mg (or an equivalent corticosteroid) was administered twice a day for 3 days beginning a day prior to each docetaxel infusion so as to prevent any anaphylactic reaction and/or fluid retention that would be caused by docetaxel. In the EC arm, the patients were given etoposide 100 mg/m² as a 1-hour i.v. infusion on days 1 to 3, and they were given cisplatin 75 mg/m² as 1-hour i.v. infusion on day 1. Cisplatin was infused with adequate hydration, diuretics and mannitol. Chemotherapy was repeated every 3 weeks in both arms and it was continued until there was objective evidence of disease progression, until a patient withdrew consent for further treatment or in the cases of early withdrawal from the trial because of an unacceptable adverse event or non-compliance with the trial protocol. Prophylactic use of granulocyte colony-stimulating factor (G-CSF) or antibiotics was not permitted. In the event of hematological and severe non- hematological toxicities that occurred at any time during treatment, the patients were treated by adjusting the dose of the chemotherapeutic agents. With regard to the hematological toxicity, a complete blood cell count with the differential and platelet counts were performed on each day of treatment. During the course of treatment, two dose reductions for each chemotherapeutic agent were allowed in the event of grades 3/4 hematological or non-hematological toxicities. All the patients were assessed in the follow-up phase for their disease progression and thereafter until death.

The clinical response evaluation procedure included all the laboratory or imaging studies that had shown abnormal findings prior to treatment. Tumor responses were assessed every two cycles by performing simple chest X-ray and CT scans of the chest and/or abdomen along with bone scans unless there was obvious evidence of disease progression after one cycle of therapy. The treatment responses were evaluated using the WHO criteria. Briefly, a complete response (CR) was defined as the absence of disease at all previously known tumor sites for at least four weeks. A partial response (PR) was defined as a 50% reduction in the sum of the products of the perpendicular diameters of all measurable lesions, and this response lasted at least four weeks. Progressive disease (PD) was defined as either a 25% increase in the area of any one lesion over the prior measurement, or the development of one or more new lesions. Stable disease (SD) was defined as any change in the previous lesion that did not fit into either the PR or PD categories. All the patients were evaluated for adverse reactions after the first dose of therapy according to the NCI-CTC grading system, version 2.

The sample size calculation was based on superiority testing for the response rate. When assuming that the treatment groups were allocated an equal number of subjects and that comparisons were based on the response rate, the null hypothesis of an equal response rate in the two groups was tested against the alternative hypothesis of a higher response rate in the DC arm. Assuming a response rate of 14% in the EC arm and 41% in the DC arm, according to a review of the literature, the correct sample size was calculated to be 33 subjects. This allowed for detection of the targeted alternative hypothesis with at least 80% power and a one-sided 5% level of significance.

The intent-to-treat (ITT) population included all the patients who were randomly assigned and treated, and who had undergone at least one disease assessment with using the same lesion imaging procedure that was used at the study's baseline. Per protocol analysis was performed on those subjects who were eligible and assessable for response without them having incurred any major protocol deviations. The primary analysis was based on the ITT population.

The primary efficacy variable, i.e., the objective response (OR), was defined as the best overall response for the CR or PR with using the WHO criteria. Differences in the OR between the treatment arms were assessed with using one-sided 5% Chi-square tests. The results were presented in terms of the estimated difference in the response rates and the p value. Disease control was defined as the best tumor response for the CR, PR or SD that was confirmed and sustained for 4 weeks or longer.

The Overall Survival (OS) was calculated from the date of enrollment to the date of death or the date when the patient was last known to be alive. The time to the progression (TTP) of disease was calculated from the date of enrollment to the date of progression or death; data for the patients who were alive and relapse-free were censored as of the date of the last known follow-up visit. The OS and TTP were estimated with using the Kaplan-Meier method and these parameters were compared in the two treatment arms with using the log-rank test. The median TTP and the 95% confidence interval (CI) were also estimated. A Cox proportional hazard regression model was used to further explore the observed differences and to identify the baseline factors that may independently predict patient survival.

Between April 2000 and March 2002, a total of 78 previously untreated patients were identified and randomized. Of the 78 patients who were recruited for the study, 76 of them received chemotherapy and so they were included in the safety and efficacy analysis. Of the 78 patients, 2 (DC, n=2) were ineligible for analysis because of withdrawal of consent and proven brain metastasis, respectively. Seventy-six patients were treated in either the DC (n=38) or the EC (n=38) arm. Among the 38 treated patients in each arm, 4 patients in the DC arm and 5 in the EC arm were excluded from the response evaluation due to their discontinuation of chemotherapy after one cycle; this was attributable to consent withdrawal (DC, n=2; EC, n=3), lost follow-up (DC, n=1; EC, n=1) or other reasons (DC, n=1; EC, n=1). Thirty-four patients in the DC arm and 33 in the EC arm finally constituted the per-protocol population. All the reported data are for the ITT population (76 patients). The median follow-up period was 13 months. The baseline patient characteristics and the baseline disease characteristics were evenly distributed across both the treatment groups in terms of the PS, histology and disease status (Table 1).

A total of 154 cycles of DC and 136 cycles of EC were administered, respectively. The median number of cycles was 4 (range: 1~9 cycles) in each arm (Table 2).

The relative dose intensity was 0.947 for docetaxel and 0.944 for etoposide. The percentage of cycles that was delayed was 17% for the DC and 14% for the EC. Leucopenia was the most common cause of cycle delay.

The tumor response rates in the ITT population are presented in Table 3. In the DC arm, 14 of the 38 patients (36.8%) had a PR and 1 (2.6%) had a CR. In the EC arm, 7 patients (18.4%) had a PR, but none had a CR. The ORR observed in the DC arm was significantly higher than that in the EC arm (p=0.023). The observed disease control rate was 65.7% and 44.7% of the patients who received DC and EC, respectively.

The median OS of the patients treated with DC and EC was 12.1 months (95% CI: 9.2~17.6) and 8.7 months (95% CI: 5.9 ~15.1), respectively. Fig. 1 shows the OS in both groups (p=0.168). The 1-year survival rates for the patients in both the groups were similar (50.4% for DC vs. 43.5% for EC). The observed 2-year survival rate in the DC arm (28.5%) was higher than that in the EC arm (16.3%), but the difference was not significant. The survival for patients with stage III disease, as compared with those with stage IV disease, was significantly different (p=0.0095). When patient survival was compared based on the tumor histology, the patients suffering with adenocarcinoma showed significantly longer survival than did those patients suffering with the other histological tumor types (p=0.0454). Other prognostic factors such as age, gender and PS did not affect patient survival. The responders to DC demonstrated significantly longer survival than did the non-responders (p=0.017). On the other hand, the response induced by EC did not affect overall survival (p=0.163) (Fig. 2). The median TTP was 5.9 months (95% CI: 4.4~7.1) in the DC arm and 2.7 months (95% CI: 1.5~4.3) in the EC arm (p=0.119) (Fig. 3).

Seventy-six patients were evaluable for toxicity. Table 4 presents the patients who experienced treatment-related toxicities of NCI-CTC grade ≥3. Febrile neutropenia occurred in 3 patients in the DC arm, but this didn't occur in the EC arm. A higher incidence of neutropenia was recorded for the patients who received EC than for those patients who received DC; however, this difference was not statistically significant. The incidence of grades 3/4 non-hematological toxicities was similar in both arms, except for fatigue, which was predominant in the DC arm. No treatment-related deaths occurred in either of the treatment arms.