Patient Selection

Patients were recruited from 83 centers in North America, Europe, the Southeast Asia, and Australia, randomly assigned in a 1:1 ratio, and stratified according to duration of response to first-line therapy (progression ≤ 6 months or > 6 months), sex, and presence or absence of liver metastases. Eligible patients were those with limited or extensive-stage SCLC who had documented complete or partial response to first-line therapy with disease recurrence after ≥ 90 days and met the following inclusion criteria: ≥ 18 years old, only one prior chemotherapy regimen, bidimensionally measurable disease, an Eastern Cooperative Oncology Group performance status ≤ 2, WBC count ≥ 3,500/μL, neutrophils ≥ 1,500 μL, platelets ≥ 100,000 μL, hemoglobin ≥ 9.0 g/dL, serum creatinine ≤ 1.5 mg/dL, bilirubin ≤ 2.0 mg/dL, and alkaline phosphatase, AST, and ALT ≤ 2× the upper limit of normal or ≤ 5× the upper limit of normal with liver metastases. Patients with CNS metastases were eligible only if they were asymptomatic without corticosteroids. Prior surgery was allowed if ≥ 4 weeks had passed, as was immunotherapy (≥ 3 months) and radiotherapy (≥ 24 hours). Concurrent chemotherapy, immunotherapy, or radiotherapy was not permitted. Concurrent radiation for palliation of bone or brain lesions was not allowed unless discussed with the medical monitor.

The protocol was approved by the institutional review board at each site. Each patient provided written informed consent.

Chemotherapy Regimens

Oral topotecan 2.3 mg/m²/d or IV topotecan 1.5 mg/m²/d (as a 30-minute infusion) was administered on days 1 through 5, every 3 weeks. Patients with complete or partial response continued treatment until disease progression or for two courses beyond best response. Patients with stable disease were recommended to receive at least four courses¹² because some responses have not been noted until completion of four to eight courses. Oral topotecan was supplied as capsules containing topotecan hydrochloride equivalent to 0.25 mg or 1.00 mg of the anhydrous free base (GlaxoSmithKline, Middlesex, United Kingdom).

During course 1, patients randomly assigned to oral topotecan were administered capsules under clinical supervision for each of the 5 days. During subsequent courses, only day 1 was orally administered in the clinic; capsules were dispensed for the remaining 4 days to be taken at home. Compliance was documented through pill counts.

Dose Modifications

After course 1, dose escalation was permitted if no toxicity more than grade 2 occurred during the previous course. Oral topotecan dose was increased in increments of 0.4 mg/m² to a maximum of 3.1 mg/m²/d, and IV topotecan was increased in increments of 0.25 mg/m² to a maximum of 2.0 mg/m²/d. Severe or prolonged neutropenia or severe thrombocytopenia was managed via dose reduction. Minimum doses permitted were 1.5 mg/m²/d for oral topotecan and 1.0 mg/m²/d for IV topotecan; delays of more than 2 weeks at these doses resulted in study withdrawal.

Assessments

Patients had to have lesions that were bidimensionally measurable according to WHO criteria.¹³ After baseline evaluation, lesions were assessed either at the end of each course (if evaluated by photography or physical examination) or at the end of alternate courses (if evaluated by computed topography or magnetic resonance imaging, x-ray, or ultrasound). The same method of evaluation was to be used throughout the study.

All responses were verified by a central radiologist independent of patient management decisions and blinded to study treatment. By implementing these stringent criteria, treatment responses have been shown to correlate with outcomes.¹⁴

Time to response was measured from first topotecan dose to first documented complete or partial response in patients who achieved a response; duration of response was measured from when response was first documented to disease progression. Time to progression was measured from first topotecan dose to progression; survival was measured from first dose until death. Follow-up was conducted at each study site to document dates of disease progression, any poststudy therapies administered, and date of death.

Blood counts were obtained on days 8 and 15 and before initiation of each subsequent course. Blood chemistry was evaluated on day 15, and if there was deterioration from baseline values, the evaluation was repeated before the next course. Toxicities (graded according to the National Cancer Institute Common Toxicity Criteria) were used to assess whether dose adjustments were indicated.

Health-related quality of life (HRQOL) was assessed using the Functional Assessment of Cancer Therapy–Lung (FACT-L).¹⁵ FACT-L is a 44-item self-report instrument that includes four generic dimensions of HRQOL and a subscale that measures symptoms specific to lung cancer (Lung Cancer Subscale [LCS]). Generic dimensions include functional well-being, physical well-being, social and family well-being, and emotional well-being. LCS measures the severity of the following seven common symptoms: shortness of breath, cough, tightness in chest, difficulty breathing, appetite loss, weight loss, and lack of clear thinking. The Trial Outcome Index (TOI) is derived by adding the physical and functional well-being LCS subscale scores. FACT-L has been well validated, translated for use in many different countries,^15-17 and used extensively in clinical trial and other study settings. Patients completed the FACT-L as the initial event of their first clinic visit (eg, before course 1, or at baseline), before each subsequent course, and at study end.

Primary end point was response rate. Secondary end points were response duration, time to progression, time to response, survival, quality of life (QOL), and toxicities.

Statistical Methods

Response rate was compared between treatments, and the estimated percent difference in response rates between regimens was calculated with 95% CIs. Sample size was not based on formal statistical criteria; rather, it was based on the feasibility of patient accrual and study completion. A study population of 150 patients per treatment arm provided 71% power that the 95% CI would exclude more than 10% difference in favor of IV treatment.

Time to event data (time to response, response duration, time to progression, and survival) were summarized using Kaplan-Meier survival methods. A hazard ratio for treatment in the presence of covariates (ie, duration of prior response, sex, and liver metastases) using Cox's proportional hazards model was generated for the survival end point.

QOL data were evaluated by calculating the total FACT-L score and the 21-item TOI.¹⁵ Scores recorded before each course of treatment after the first course were compared with baseline scores. A repeated measures analysis was performed to compare the rate of change between the two treatment groups for each dimension or subscale.

Patients

Between January 25, 1999, and October 20, 2001, 309 patients were randomly assigned (Fig 1); 153 patients were assigned to oral topotecan, and 151 patients were assigned to IV topotecan (five patients received no treatment). Because there was no predetermined treatment duration, all patients who had least one full course of treatment were considered to have completed the study unless withdrawn as a result of an adverse event, protocol deviation, patient request, or loss to follow-up. Proportions of patients completing therapy were similar for both oral (124 patients, 81%) and IV topotecan (128 patients, 85%) treatment arms. Withdrawals for adverse events were also similar in both groups (12% and 13% of patients on oral and IV topotecan, respectively).

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 1.

CONSORT diagram. IV, intravenous; AEs, adverse events; ITT, intent-to-treat; QOL, quality of life.

Demographics and baseline characteristics were well matched between groups (Table 1). A majority of patients (approximately 70%) had extensive disease, and liver metastases were present in approximately 28% of patients. Forty-three patients had documented baseline brain or leptomeningeal metastases (18 patients on oral topotecan, and 25 patients on IV topotecan). All patients had responded to previous cancer chemotherapy; 129 patients (84.3%) in the oral topotecan arm and 125 patients (82.8%) in the IV topotecan arm had received prior combination chemotherapy that included platinum (cisplatin or carboplatin). Approximately 10% of patients in both treatment groups had a treatment-free interval of less than 90 days at study entry.

Delivered Chemotherapy

In each treatment group, patients received a median of four courses of topotecan, with at least 40% of patients in either group continuing to receive treatment beyond course 4. Median dose-intensity was 3.74 mg/m² for the oral topotecan arm and 2.31 mg/m² for the IV topotecan arm (ratio = 1.61), which reflects the difference in oral and IV doses (ratio = 1.53).

Dose escalations above the starting dose occurred for 36% of patients treated with oral topotecan and 19% of patients receiving the IV regimen. More than 90% of initial dose escalations were made at the end of courses 1 or 2. Dose reductions were made for 31% of patients in the oral group and 35% of patients in the IV group. A majority of the dose reductions were made at the end of course 1, primarily as a result of hematologic toxicity. Of the 55 patients (36%) who had a dose escalation of oral topotecan to 2.7 mg/m²/d, 25 had a further dose escalation up to 3.1 mg/m²/d. For IV topotecan, of the 28 patients (19%) who were dose escalated to 1.75 mg/m²/d, 13 had a further dose escalation to 2.0 mg/m²/d. No patient was withdrawn from the study or excluded from any analysis as a result of lack of compliance.

Efficacy

The overall response rate was 18.3% (95% CI, 12.2% to 24.4%) for patients who received oral topotecan and 21.9% (95% CI, 15.3% to 28.5%) for patients who received IV topotecan (Table 2). Difference in response rates (oral –IV) was −3.6% (95% CI, −12.6% to 5.5%). In the oral topotecan group, median time to response was 6.1 weeks, and median duration of response was 18.3 weeks. In the IV topotecan group, median time to response was 6.1 weeks, and median duration of response was 25.4 weeks. Median time to progression was 11.9 weeks in the oral topotecan group compared with 14.6 weeks in the IV topotecan group (Table 3). A plot of the Kaplan-Meier product-limit estimates for time to progression is provided in Figure 2. Of the 43 patients with baseline brain or leptomeningeal metastases, one patient on the IV topotecan arm experienced a partial response.

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 2.

Kaplan-Meier plot for time to progression in the intent-to-treat population. IV, intravenous.

Median survival time was 33.0 weeks (95% CI, 29.1 to 42.4 weeks) in the oral group and 35.0 weeks (95% CI, 31.0 to 37.4 weeks) in the IV group, with data censored for 13.7% and 10.6% of patients in the respective groups (Table 3). Cox proportional hazards regression showed no difference between the two groups (hazard ratio = 0.98; 95% CI, 0.77 to 1.25). A plot of the Kaplan-Meier product-limit estimates for survival is provided in Figure 3. Data collected during poststudy monitoring showed that similar proportions of patients in each group had received third-line chemotherapy (33% of patients in the oral group and 35% of patients in the IV group). At 1 year, the survival rate was 33% after treatment with oral topotecan and 29% after treatment with IV topotecan; the 2-year survival rates were 12% for oral topotecan and 7% for IV topotecan.

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 3.

Kaplan-Meier plot for survival in the intent-to-treat population. IV, intravenous.

Toxicity

The principal toxicity was neutropenia for both groups (Table 4). Fever and/or infection of at least grade 2 associated with grade 4 neutropenia, together with sepsis, occurred in 5% of courses in both groups. Antibiotic use was slightly higher for patients treated with IV topotecan compared with oral topotecan (56% v 41% of patients received a systemic antibiotic and 23% v 14% of patients received an IV antibiotic, respectively). Granulocyte colony-stimulating factor was administered to a higher proportion of patients in the oral than the IV group (25% v 16%, respectively), although the proportion of treatment courses was similar in both groups (9% v 7%, respectively). With the protocol-specified dose adjustments, there was no evidence of cumulative toxicity.

At time of analysis, 267 patients had died; 250 of these patients had died as a result of disease progression. Ten patients (six in the oral group and four in the IV group) died as a result of hematologic toxicity, septic shock related to treatment with topotecan, or of other causes where a relationship to topotecan could not be excluded. This treatment-related mortality is consistent with reports of deaths in studies on second-line therapy for SCLC, in which deaths caused by hematologic toxicity or septic shock related to treatment ranged from 0% to 12%.^18,19

Incidence of grade 4 thrombocytopenia was higher in the oral than in the IV group (Table 4), but platelet transfusions were used to a similar extent in both groups (oral group, 15%; IV group, 13%). Grade 3 or 4 anemia occurred in a slightly higher proportion of patients in the IV group; however, transfusions of RBCs were administered to a similar number of patients in each group (oral group, 37%; IV group, 43%).

Nonhematologic adverse events in the oral and IV groups involved mainly nausea (42.5% and 42.4%, respectively), fatigue (30.7% and 36.4%, respectively), and alopecia (25.5% and 29.8%, respectively). Diarrhea was one of the most commonly occurring events in the oral group; 35.9% of patients in the oral group experienced diarrhea (7.9% grade 3 or 4) compared with 19.9% of patients in the oral group (2.7% grade 3 or 4; Table 4). These events were predominantly grade 1 or 2, and there were no suggestions of cumulative toxicity. Diarrhea in the oral topotecan group was self-limiting or manageable with treatment, primarily oral loperamide.

HRQOL

All randomly assigned patients reported baseline HRQOL information using FACT-L and LCS, and 88% (oral) to 90% (IV) of patients reported HRQOL before at least one course of therapy. Questionnaire response was 75% and 78% for the oral and IV groups, respectively, after two courses of therapy. Rates at which patients failed to complete QOL assessment at one or more courses of therapy were similar in the two arms.

Least squares estimates for mean change from baseline indicated no statistical difference between treatment groups for subscale dimension scores and LCS, TOI, and FACT-L total scores. Only a small decline in HRQOL was noted for each treatment group compared with declines that may be expected in an untreated lung cancer population (ie, patients on best supportive care), which suggests that treatment (with oral or IV topotecan) may delay symptom progression, as was subsequently reported.²⁰ Mean change from baseline to last course also showed no statistical differences between the oral and IV topotecan arms.