IL-2 in combination with IFN-αand 5-FU versus tamoxifen in metastatic renal cell carcinoma: long-term results of a controlled randomized clinical trial

We conducted a prospectively randomized clinical trial to compare the efficacy and safety of subcutaneous interferon-α2a, subcutaneous interleukin-2 and intravenous 5-fluorouracil as home therapy against oral tamoxifen in 78 patients with progressive metastatic renal cell carcinoma. Treatment courses consisted of interferon-α2a 5 × 106 IU m−2day 1 weeks 1 + 4; days 1, 3, 5 weeks 2 + 3; 10 × 106 IU m−2days 1, 3, 5 weeks 5–8; interleukin-2 10 × 106 IU m−2twice daily days 3–5 weeks 1 + 4; 5 × 106 IU m−2days 1, 3, 5 weeks 2 + 3; and 5-fluorouracil 1000 mg m−2day 1 weeks 5–8. The tamoxifen group received tamoxifen 80 mg twice daily over 8 weeks. Among 41 patients treated with interleukin-2, interferon-α2a and 5-fluorouracil there were 7 complete (17.1%) and 9 partial responders (21.9%), with an overall objective response rate of 39.1% (95% CI, 24.2–55.5). An additional 15 patients (36.6%) were stable throughout therapy. The overall survival was 24 months (range 5–76+). In 37 patients receiving tamoxifen no objective remissions occurred. 13 patients (35.1%) had stable disease and 24 patients (64.9%) showed continued disease progression. The overall survival was 13 months (range 3–73+). In summary, this home-based therapy regimen of interferon-α2a, interleukin-2 and 5-fluorouracil demonstrated significant therapeutic efficacy in patients with progressive renal cell carcinoma when compared to hormonal therapy. © 2001 Cancer Research Campaign  http://www.bjcancer.com

The human investigations were performed after approval by a local Human Investigations Committee and in accordance with an assurance filed with and approved by the Department of Health and Human Services; written informed consent was obtained from all patients prior to entry into the study.

Study design
Patients were stratified according to known clinical predictors (Elson et al, 1988;Tourani et al, 1998) to allow for equal risk distribution in both treatment arms (Table 2, Figure 1). They were subsequently randomized to receive IL-2 and IFN-α combined with 5-FU (Arm A) or tamoxifen (Arm B); randomisation was performed on a per centre basis to rule out centre-related statistical bias. There was no statistically significant inbalance of risk factors and risk scores when comparing both treatment arms.
In the immunochemotherapy group (Arm A) 34% of patients received one cycle, 43% of patients obtained 2 cycles and 23% received 3 cycles. In the hormonal therapy group (Arm B) 72% of patients passed through one cycle, 19% obtained 2 cycles, 6% received 3 cycles, and 3% obtained 4 cycles. Re-evaluation of the patients tumour status was performed between treatment  q Erythrocyte sedimentation rate > 70 mm 1 h -1 q ECOG performance status > 0 q Lactic dehydrogenase > 280 U l -1 q more than one metastatic site q Neutrophilic granulocytes > 6000 µl -1 q interval from diagnosis of primary tumour to q Hemoglobin < 10 g dl -1 treatment of metastatic disease ≤ 2 years q extrapulmonary metastases only q bone metastases Individual risk was defined as Individual risk was defined as LR low risk in patients without any of these factors LR low risk in patients with either one or none of these factors HR high risk in patients with one or more of these factors HR high risk in patients with more than one of these factors a There was no statistically significant inbalance of risk factors and risk scores when comparing both treatment arms.
cycles. All patients were examined by means of computed tomography.
After the first 8 weeks of therapy patients primarily treated with p.o. tamoxifen crossed over to immunochemotherapy upon disease progression wherever performance status allowed further systemic therapy. Concomitant medication was given as needed to control adverse effects of immunochemotherapy.

Assessment of response, survival and toxicity
Response to therapy was evaluated according to World Health Organization (WHO) criteria: complete response -disappearance of all signs of disease for a minimum of 8 weeks; partial response -50% or more reduction in the sum of products of the greatest perpendicular diameters of measurable lesions, no increase in lesion size and no new lesions; stable disease -less than a partial response with no disease progression for at least 8 weeks; progressive disease -25% or more increase in sum of products in the longest perpendicular diameters of measurable lesions or the development of new lesions. Duration of response was measured from start of therapy. In case of patients with early progression, progression-free survival was calculated at 0 months. Survival was measured from start of therapy to date of death or to the last known date to be alive.
Systemic toxicity was evaluated at weekly intervals using a grading system adapted from the WHO.
Treatment efficacy was assessed on intent-to-treat basis.

Statistical analysis
The statistical end points in our analysis were (1) objective response and (2) overall and progression-free survival of patients. The potential response rate was assumed to be 35% (immunochemotherapy) and 5% (hormonal therapy), respectively. It was intended to find an immunochemotherapy-induced objective response benefit over tamoxifen at 30% with a probability of 95% (α = 0.05) and a sample power of 1-β = 0.80. The probability of survival was plotted over time according to the method of Kaplan and Meier. Statistical significance was assessed using the Log-rank, Mantel-Cox and the Breslow test. The Wilcoxon test was used to analyse differences between dependent nonparametric data.

RESULTS
To balance risk factors by treatment arms patients were stratified according to risk and, subsequently, randomised within risk adjusted groups. 78 patients were enrolled in this clinical trial and stratified for risk. 41 patients were randomized to receive combined immunochemotherapy and 37 patients were randomized to receive single-agent p.o. tamoxifen.  (Table 3). The overall objective response rate was 39.1% (95% CI, 24.2-55.5). In partial responders, the median response duration was 18 months (range 3-31), the complete response durations ranged from 3 to 76+ months (median not reached). Objective tumour regressions were seen in lung (14), lymph nodes (4), liver (1), bone (1), adrenal gland (1) and peritoneum (1). 15 patients (36.6 %) showed disease stabilization. The progression-free survival in this subgroup ranged from 1 to 20 months with a median of 7 months. 10 patients (24.4%) exhibited continuous disease progression despite therapy.
In the tamoxifen group, there was no objective response. 13 patients (35.1%) had stable disease with a median progression-free interval of 6 months (range 2-28). 24 patients (64.9%) showed continued progression of disease.

Overall survival
Overall median survival of all patients entered into the study was 18 months, whereby 17 patients continue to be alive at last follow up. In the immunochemotherapy regimen the median survival was 24 months (range 5-76+; 95% CI 11-37), with statistical significance (Log rank P = 0.0325) compared to the tamoxifen group with a median survival of 13 months (range 2-73+; 95% CI 8-18) (Figure 3).
Patients primarily treated with tamoxifen and crossed over to immunochemotherapy had a 5-year survival of 18.7% as opposed to 9.5% 5-year survival in patients without crossover (Figure 6).

Treatment toxicity
Combined immunochemotherapy was well tolerated and was always given in the outpatient setting. Side effects were mostly limited to WHO grade I and II with patients experiencing flu-like symptoms as malaise (grade I/II 78%; III 7%), chills (grade I/II

DISCUSSION
The use of biologic therapies has an established role in the treatment of metastatic renal cell carcinoma (Atzpodien et al, 1990;Wirth, 1991;Figlin et al, 1992;Stahl et al, 1992). However, very few randomized controlled trials have confirmed the therapeutic efficacy of rIL-2 based regimens in this disease.
The discrepancy in phase I/II results might be best explained by at least 3 uncontrolled variables: (A) patient selection and clinical/biological risk distribution; (B) variations in the rIL-2 based regimen (e.g., reconstitution of drug, route and dose of administration, use of rIL-2 with concomitant chemotherapy); and (C) different standards of patient care (e.g., surgery, supportive care upon IL-2 therapy).
After tamoxifen therapy some authors reported a low objective response rate of 1.7% (95% CI, 0.04-9.09%) (Schomburg et al, 1993) based on spontaneous regression of metastatic renal cell cancer. The lack of objective remissions in the present tamoxifen group was most likely due to the small number of patients treated.
Previous uncontrolled trials mostly focused on objective responses with limited follow-ups and survival of around 2 years. In the present study we, for the first time, reported the long-term randomized outcome of IL-2-based immunochemotherapy in a risk adjusted cohort of renal cell carcinoma patients.
The median overall survival of 24 months in the immunochemotherapy group was statistically superior to 13 months in the tamoxifen group. Notably, patients failing primary tamoxifen could reach a median survival of 19 months after crossing over to combined immunochemotherapy, and experienced no significant decrease in long-term survival when compared to the primary immunochemotherapy group. This may be due to the fact that only patients with good or fair performance status -and hence possibly slower rate of progression -were crossed over whereas those with lower performance status -and hence more aggressive disease -did not undergo secondary immunotherapy.
Interestingly, survival graphs of low-and high-risk patients, respectively, suggested that high-risk patients experience a more pronounced survival benefit from immunochemotherapy compared to oral tamoxifen.
In most patients reported here, immunochemotherapy-related toxicity was limited to WHO grades I and II, notably malaise, fever, nausea, diarrhoea, while dose-limiting WHO III/IV side effects occurred in no more than 10% of patients and always normalized after termination of therapy. Overall, the s.c. route of application reduced IL-2 toxicity when compared to i.v.-based IL-2 regimens (Sleijfer et al, 1992); also, low therapy-related toxicity as observed here may have been due to adequate patient selection as required for home therapy. In summary, this prospectively randomized clinical trial established the long-term therapeutic efficacy, safety and tolerability of s.c. IL-2, s.c. INF-α, i.v. 5-FU as a home therapy in prospectively controlled patients with metastatic renal cell carcinoma.