The prognosis of patients with metastatic renal cell carcinoma (MRCC) remains poor, with an estimated 5-year survival rate of 0–20%.1 Intravenous (IV) interleukin-2 (IL-2) and interferon α (IFN-α) have shown objective responses in 10–25% of patients.2–5 In 1998, our group reported the results of a large, multicenter, Phase III trial (the CRECY trial) that included 425 patients with MRCC who were treated with IL-2 and/or IFN-α.5 Those patients were assigned randomly to receive subcutaneous (SC) IFN-α, a continuous infusion of IL-2, or a continuous infusion of IL-2 associated with SC IFN-α. The results of that trial were in favor of an association of IL-2 and IFN-α (18.6% response) compared with the rates obtained with IL-2 alone (6.5%) or IFN-α alone (7.5%).5 Moreover the 1-year event free survival rate also was improved significantly by the combination of cytokines (20%). Conversely, toxicities induced by continuous infusion were frequent and severe. Among 140 patients who were treated with IL-2 and IFN-α, 470 events of Grade 3 or 4 toxicity were reported. Eight patients died during treatment or within 1 month of treatment from causes unrelated to MRCC.
In the 1990s, the use of SC IL-2, either alone6–10 or in association with IFN-α7, 11–15 and/or 5-fluorouracil (5-FU), was investigated in patients with MRCC.7, 13, 16–24 Most of these Phase II studies showed a sustained efficacy of IL-2 with a 10–25% objective response rate. In addition, the adverse events of IL-2 were reduced significantly. For this reason, most existing IL-2-based treatments administered to patients with MRCC use the SC route in routine practice, as approved by health authorities in some European countries, such as France. Nevertheless, when it was tested in our group, which included large numbers of patients, the results obtained with SC IL-2 associated with IFN-α with or without 5-FU were disappointing.19, 25 The reason for this discrepancy may be the intermittent use of our IL-2 schedule and the low dosage. Indeed, IL-2 doses were lower than in previous cytokine combination treatments with continuous IL-2 infusion, such as the CRECY trial5 or other SC IL-2 regimens.6, 7
Because of the disappointing results obtained with SC IL-2-based regimens in our group, we decided to evaluate the most beneficial treatment of the CRECY trial using identical cytokine doses and schedules but with SC administration rather than IV administration of IL-2. Here, we report the results of a multicenter Phase II trial testing this regimen in patients with MRCC.
MATERIALS AND METHODS
Eligible patients had histologically proven renal cell carcinoma with progressive, distant, metastatic disease. Patients were adults age < 80 years who had an Eastern Cooperative Oncology Group (ECOG) performance status ≤ 1. Eligible individuals had measurable metastatic disease without brain metastases, as determined from a mandatory computed tomography (CT) scan. Patients who presented with more than one metastatic site, including the liver, and who had less than 12 months from the time they underwent nephrectomy to the discovery of metastases were excluded from the trial, according to our previous results.5 Patients should have received no previous systemic therapy for their disease and no radiotherapy during the previous 6 weeks. They had normal blood cell counts (white blood cells ≥ 4 × 109/L, hematocrit > 30%, and platelets ≥ 120 × 109/L), normal bilirubin level (< 20 μmol/L), and creatinine concentration < 180 μmol/L. Patients were not suffering from a current symptomatic cardiac disease and had a left ventricular ejection fraction > 50%. Adequate organ functions were required without respiratory, neurologic, or psychiatric disorders. Patients had a life expectancy of at least 12 weeks. Patients with severe infection, a positive human immunodeficiency virus test, or chronic hepatitis were excluded along with patients who were on corticosteroids. Patients had no history of organ allograft, previous cytokine treatment, or other malignancies. Pregnant or lactating women were excluded.
In addition, clinical history and physical examination were recorded for each patient. Preinclusion staging included cerebral, thoracic, and abdominal CT scans and a bone scan. Written informed consent was obtained from each patient before entry into the trial. In agreement with French laws, the trial received approval from the Lyon Ethics Committee.
IL-2 (Proleukin; Chiron Therapeutics, Suresnes, France) was administered SC at a dose of 9 × 106 IU/m2 twice daily for 5 days (Table 1). The treatment schedule consisted of two induction cycles and four additional cycles, with a 3-week rest between cycles. Each induction cycle consisted of two 5-day courses of IL-2 separated by a 9-day break. Each additional cycle was a 5-day treatment followed by a 3-week rest. Concurrently, IFN-α (Roferon; Roche, Neuilly sur Seine, France) was administered SC at a dose of 6 × 106 IU per day three times per week during the induction cycles and the additional cycles.
Table 1. Schedule and Dose of Treatmenta
|Induction cycle (× 2)|| |
| IL-2||1–5 and 15–20|
| IFN-α||1, 3, 5, 8, 10, 12, 15, 17, and 19 (TIW)|
|Additional cycle (× 4)|| |
| INF-α||1, 3, and 5|
Recommendations were made in the study protocol to adapt the schedule and the dose to toxicity. These recommendations were identical at least to the recommendations made by the Groupe Français d'Immunothérapie for previous trials (SCE study; unpublished data),19, 25 including IL-2 and/or IFN-α. Modifications of the IL-2 and IFN-α schedules were as follows: IL-2 and IFN-α were stopped temporarily in patients with hypotension ≥ Grade 3, supraventricular arrhythmia, Grade 3 central nervous system toxicity, total bilirubin > 40 μmol/L, serum creatinine > 400 μmol/L, bacterial infection, and dyspnea at rest. Treatment was started again at the full dose when toxicities decreased to Grade 1. In patients who had a recurrence of these toxicities ≥ Grade 3, IL-2 was reduced definitively by 50%. In patients who had a recurrence of these toxicities ≥ Grade 3, IL-2 was stopped definitively. Recommendations to stop IL-2 definitively were given for the following events: severe allergy, severe cardiac dysfunction or ischemia, Grade 4 neurologic disorders, and increases in bilirubin or in creatinin values that did not decrease at least to Grade 1. In patients with Grade 3 neutropenia, platelet counts (50 × 109 L), or transaminases (5× normal value), IFN-α was stopped until levels achieved Grade 1 and were reduced by 50% in patients who had a recurrence of these toxicities.
Evaluation of Treatment
Tumor response, including thoracic and abdominal CT scans and a bone scan, was evaluated after 12 weeks and 27 weeks of treatment using the World Health Organization (WHO) criteria.26 A complete response (CR) was defined as the complete disappearance of all measurable and evaluable tumor sites for at least 4 weeks. The duration of CR was calculated from the date of first documentation of CR to the date of first evaluation of disease progression. A partial response (PR) was a decrease ≥ 50% in the sum of products of the greatest perpendicular tumor dimensions that lasted for at least 4 weeks, with no increase in known lesions and without the appearance of any new lesion. The duration of PR was calculated from the first day of treatment. When evaluation showed a decrease < 50% and an increase < 25% in the size of lesions, patients were considered to have stable disease (SD). Disease was considered progressive (PD) when any existing lesion increased by ≥ 25% or when a new lesion appeared. The results of the successive bone scans were considered a sign of PD if new spots appeared, a sign of SD if new spots did not appear, and a sign of CR when all existing spots disappeared. Patients who presented with CR, PR, or SD were evaluated further every 2 months during the first year and every 4 months thereafter. The follow-up evaluation consisted of a clinical examination; laboratory studies (hematologic, creatinine, hepatic, and thyroid parameters); cerebral, thoracic, and abdominal CT scans; and a bone scan.
Survival duration was evaluated from the beginning of treatment to the date of the last contact or the date of death. Progression free survival was calculated from the beginning of treatment to the date of last follow-up or the date PD was identified. Toxicities encountered were classified according to the WHO grading system.
The major end point of the study was the response rate. The secondary end points were progression free survival, overall survival, and toxicity. First, we assessed the response rate after 19 patients had been recruited so that we would have a 95% chance of detecting at least one response when the actual response rate was ≥ 15%.27 If at least one response occurred in the first 19 patients, then we planned to increase the number of patients to assess the response rate with 5% precision (i.e., 2 responses justified the inclusion of 45 more patients). The Kaplan–Meier actuarial method was used for the calculation of overall survival and progression free survival.
From April, 1997 to January, 1998, 66 patients with MRCC were enrolled in the trial in 15 centers. When the 19th patient was recruited, an analysis of files showed 2 PRs that were reported to the coordinating center. Consequently, patient enrollment was pursued as planned with 64 patients (19 + 45 patients; see Statistical Analysis, above). When the 64th patient was enrolled, inclusion was closed, although 2 patients who already had signed informed consent and had planned to start treatment were accepted as additional patients. All patients were evaluable for response and toxicity. The main characteristics of the patients are outlined in Table 2. Most patients had an ambulatory performance status: Thirty-one patients (47%) had an ECOG performance score of 0, and 30 patients (45%) had an ECOG score of 1. The time from diagnosis of renal cell carcinoma to the occurrence of metastases was < 12 months in 42 patients (64%). Most patients had undergone prior nephrectomy (59 patients). At the time of treatment, the metastatic disease was localized in the lung (49 patients), mediastinal or abdominal lymph nodes (39 patients), bone (24 patients), liver (10 patients), or at the site of nephrectomy (11 patients). Thirty-one patients (47%) had at least 2 tumor sites.
Table 2. Characteristics of Patients
|Assessable patients|| || |
| For toxicity||66||100|
| For response||66||100|
|Age (yrs) median (range)||58 (32–75)||—|
|Performance status [ECOG]|| || |
|Time from diagnosis to first metastases|| || |
| < 12 months||42||64|
| ≥ 12 months||17||26|
|Histology|| || |
| Clear cell||46||70|
| Not specified||16||24|
|Prior hormonotherapy or chemotherapy||1||2|
|Site of metastatic disease|| || |
| Mediastinal lymph nodes||26||39|
| Abdominal lymph nodes||13||20|
| Adrenal gland||11||16|
| Recurrence at nephrectomy site||10||15|
|No. of sites with metastases|| || |
| > 2||31||47|
Administration of Treatment and Toxicity
During the induction cycle, 35 patients (51%) received ≥ 80% of the planned dose of IL-2, whereas 43 patients (63%) received ≥ 80% of the planned dose of IFN-α. The main toxicities (Table 3) were fever in 66 patients (100%), decreased performance status in 52 patients (79%), nausea/emesis in 47 patients (70%), hypotension in 46 patients (70%), diarrhea in 41 patients (62%), cutaneous erythema or nodule at the site of injection in 37 patients (56%), and increased creatininemia in 28 patients (43%). Fifty-three patients (80%) had at least 1 Grade 3 toxicity related to the treatment. Twenty-two patients developed the following Grade 4 toxicities: hypotension in 16 patients (24%); decreased performance status in 4 patients (6%); dyspnea in 2 patients (3%); mucositis in 2 patients (3%); and fever, ventricular tachycardia, and anemia in 1 patient each. No toxic death was reported.
Table 3. Toxicity
|Fever||35 (53)||29 (44)||1 (2)|
|Decreased performance status||19 (29)||16 (24)||4 (6)|
|Diarrhea||12 (18)||11 (17)||0|
|Nausea/emesis||22 (33)||14 (21)||0|
|Mucositis||7 (11)||1 (2)||2 (3)|
|Local skin nodules||13 (20)||0||0|
|General skin disorders||30 (45)||4 (6)||0|
|Hypotension||15 (23)||14 (21)||16 (24)|
|Cardiotoxicity||0||5 (8)||1 (2)|
|Dyspnea||1 (2)||1 (2)||2 (3)|
|Neurologic toxicity||3 (5)||2 (3)||0|
|Infection||2 (3)||3 (4)||0|
|Hypercreatininemia||11 (17)||2 (3)||0|
|Hematological toxicity||12 (18)||4 (6)||1 (2)|
|Increased transaminases||4 (6)||0||0|
|Weight gain||1 (2)||1 (2)||0|
|Others||8 (12)||6 (9)||0|
Response to Treatment and Survival
The median follow-up was 43 months. At Week 12 after the start of treatment, 5 patients (7.6%; 95% confidence interval [95%CI], 2.5–16.8%) had achieved an objective response, with 2 CRs. The duration of responses was 7.5 months, 17.5 months, 27.4 + months, 32.1 months, and 43.2 months. Twenty-four patients had SD, whereas 38 had PD. The sites of response were lung, bone, lymph nodes, spleen, and recurrence at the nephrectomy site. The median survival for all patients was 14 months (95%CI, 11.3–16.7 months). The median progression free survival was 3 months (95%CI, 1.8–4.2 months).
The association of subcutaneous IL-2 and IFN-α used in this study replicated the IV IL-2 dosage and schedule of a previous trial (CRECY).5 It yielded a low response rate (7.6%) that was disappointing compared with the efficacy of previous treatments that delivered the same regimen with IV administration (18.6%).5 In this study, the 14-month median survival can be compared with the 14.5-month survival reported for all patients in the CRECY trial, regardless of the treatment arm (IFN-α, IV IL-2, or IV IL-2 and IFN-α). Nevertheless, if we consider only patients from the CRECY trial who were treated with IV IL-2 and IFN-α, their overall survival was longer (up to 17 months). Moreover, increasing SC IL-2 to the dose delivered in the IV IL-2 schedule that was used in the CRECY trial did not improve the results. A low response rate (< 10%) was obtained, as observed in previous large Phase II or Phase III studies by our group using SC IL-2 and IFN-α (SCE study; unpublished data).25
The different response rates raise the question of possible differences in the populations treated in each trial. Considering the major eligibility criteria, the CRECY trial was more restrictive compared with the current trial, leading to the inclusion only of patients age ≤ 70 years with serum creatinine < 150 μmol/L, which may have affected tolerance but not efficacy. If we consider the parameters that are known to affect the prognosis of patients with MRCC, then the only difference between patients in the CRECY trial and patients in the current study was the ECOG performance status: 77% of the patients in the CRECY trial had an ECOG performance status of 0 compared with only 47% of patients in the current study. Undoubtedly, this difference may have affected the response rate, although it probably is not sufficient to account for the 50% decrease in response rate, because all other major prognostic factors were identical: time from diagnosis to metastases, prior nephrectomy, and number of metastatic sites. The lack of randomization may have introduced bias; therefore, a true comparison of results from the current study and with results from previous trials is not possible.
Nevertheless, because the likelihood of a heavy population selection bias has been ruled out, this study raises the recurring question of the efficacy of SC IL-2-based treatments compared with IV IL-2-based treatments. Two previous Phase III studies from the National Cancer Institute (NCI)10 and the Cytokine Working Group15 addressed this point and showed an increased response rate in treatments with IV IL-2 compared with SC IL-2; however, the difference was not significant, and the follow-up was too short to draw conclusions about survival or response duration. The NCI study10 compared high-dose IV IL-2 with SC IL-2 according to the schedule reported by Buter et al.6 Of 109 patients who were evaluated at the time of that report, 16% had responded to treatment, including 7% CRs in the high-dose IV IL-2 arm and 11% objective responses in the SC IL-2 arm. The study from the Cytokine Working Group15 compared IV IL-2 plus SC IL-2 with IFN-α in 193 patients. The response rates were 25% in the IV IL-2 arm and 12% in the SC IL-2 plus IFN-α arm, with a median response durations of 10 months and 7 months, respectively.
Other Phase II trials testing IL-2 and IFN-α yielded 17–24% responses. Considering the results of Phase II trials from groups with a large previous experience with IV IL-2, like the current study, the Cytokine Working Group13 compared results from 47 patients who received IL-2 and IFN-α with results from 71 patients who received IV IL-2. The response rate was 17% in both arms (CR rate, 4% and 7%, respectively), and the median response duration was 12 months and 53 months, respectively, although there was a longer follow-up in the IV IL-2 arm.
SC IL-2, either alone or in combination, is used widely, especially outside clinical trials, because of its significant efficacy demonstrated in Phase II trials. However, until Phase III trials obtain definitive results, the response rates or the median durations of response or survival obtained with SC IL-2 regimens will not be sufficient to support their efficacy compared with IV IL-2 administered by bolus or continuous infusion,10, 13, 15 especially when trials with identical cytokine schedules and dosages are compared.5
When side effects are considered, SC IL-2-based therapy causes less acute and higher grade toxicity compared with IV IL-2-based therapy.6–14 Nevertheless, when the schedule and dosage of IL-2 and IFN-α used in this study were identical to those of a previous IV IL-2 and IFN-α regimen, the side effects were more severe than expected. In the current study, 66 patients had 140 Grade 3–4 toxic events, whereas 470 toxic events were reported by 140 patients in the CRECY trial.5 In our study, 46%, 45%, 30%, 21%, and 17% of patients had Grade 3–4 fever, hypotension, decreased performance status, nausea/emesis, and diarrhea, respectively; whereas the percentages were slightly higher among patients in the CRECY trial, who received continuous IV IL-2 and IFN-α infusion: 56%, 64%, 36%, 31%, and 25%, respectively When considering side effects with higher morbidity, 10% of patients in this trial had Grade 3–4 cardiotoxicity compared with only 8% of patients in the CRECY trial. Conversely, neurologic disorders (current trial vs. CRECY trial; 3% vs. 14.3%, respectively), infection (4% vs. 9%), and elevated transaminases (0% vs. 11.4%) were less frequent. A majority of patients who had Grade 3 or 4 toxicity events had to be hospitalized during part of the treatment program because of toxicity, although only 1-week hospitalization at the beginning of treatment was planned. The toxicities encountered in this study were more frequent and more severe compared with the toxicities in previous SC IL-2-based studies in which IL-2 was given at a lower dose.6–14, 25
When considering identical schedules of SC and IV IL-2 administration, our study showed that the reduced toxicity observed with low doses of SC IL-2 was not maintained. Physicians are faced with a therapeutic dilemma: either choose to obtain increased IL-2 efficacy despite severe toxicity, which would favor the use of IV regimens, or decide to keep toxicity to a minimum and use SC IL-2 despite the possibility of lower response rates.
In conclusion, this study did not reproduce a similar efficacy of SC IL-2 and IFN-α compared with IV IL-2 and IFN-α, as shown in previous trials. Moreover, increasing SC IL-2 dosage and intensifying the schedule of IL-2 and IFN-α did not induce a better response rate compared with lower doses. The controversy will not be resolved until results of ongoing Phase III trials demonstrate whether IV IL-2 is more efficient than SC IL-2 for increasing survival and/or prolonging the objective response. Therefore, the role of SC IL-2-based treatments for patients with MRCC remains to be defined.
The authors are indebted to the nurses of the medical oncology departments at the medical centers who provided the patients with excellent and compassionate care; to the clinical research technicians for data management; to Bernard Escudier, M.D. (Institut Gustave Roussy, Villejuif) for reviewing the article; and to Dorothée Quincy and Marie-Dominique Reynaud for assistance in preparing and editing the article.