A three-arm phase III randomised trial comparing combinations of platinum derivatives, ifosfamide and/or gemcitabine in stage IV non-small-cell lung cancer

J. P. Sculier+, J. J. Lafitte, J. Lecomte, T. Berghmans, J. Thiriaux, M. C. Florin, A. Efremidis, C. G. Alexopoulos, P. Recloux, V. Ninane, P. Mommen, M. Paesmans and J. Klastersky §

Institut Jules Bordet, Bruxelles, Belgium

Received 19 October 2001; revised 27 November 2001; accepted 19 December 2001.


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Objective

To determine, in stage IV non-small-cell lung cancer (NSCLC), if the combination of gemcitabine—a new active drug—with ifosfamide (IG) or with the cisplatin-carboplatin association (CCG) will improve survival (primary end point) in comparison with a first-generation regimen, cisplatin–carboplatin–ifosfamide (CCI).

Patients and methods

A total of 284 chemotherapy-naïve patients with metastatic NSCLC were randomised. Four were ineligible and 16 were not assessable for responses. Cisplatin was given at 60 mg/m2 on day 1, carboplatin AUC 3 mg.min/ml on day 1, ifosfamide 4.5 g/m2 on day 1 and gemcitabine 1 g/m2 on days 1, 8 and 15. Courses were repeated every 4 weeks. Response was assessed after three courses and chemotherapy was continued in responding patients until best response. There were 94 eligible patients in the CCI arm, 92 in CCG and 94 in the IG arm.

Results

The objective response rates for CCI, CCG and IG were 23% [95% confidence interval (CI) 15% to 32%], 29% (95% CI 20% to 39%) and 25% (95% CI 16% to 33%), respectively (P = 0.61). Median survival time was 24, 34 and 30 weeks, respectively (P = 0.20). One-year survival was 23, 33 and 35%, and 2-year survival was 11, 14 and 17%, respectively. In some subgroups (older patients, women), there was a significant survival advantage for CCG and IG compared with CCI. Toxicity was tolerable: severe alopecia was less frequent in the CCG arm, and IG was associated with significantly more thrombopenia while CCG was associated with more leucopenia.

Conclusion

In stage IV NSCLC, treatment with regimens including the new drug gemcitabine were associated with a better but not statistically significant observed survival compared with a classical first-generation cisplatin-containing regimen. The non-platinum combination of gemcitabine was as effective as its combination with platinum.

Key words: chemotherapy, cisplatin, gemcitabine, ifosfamide, non-small-cell lung cancer, randomised trial


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
The introduction of active chemotherapy in the treatment of non-small-cell lung cancer (NSCLC) has allowed some improvement in the survival of metastatic (stage IV) disease [1, 2]. Cisplatin is considered to be the main active drug. Other active ‘first-generation’ drugs include vindesine, vinblastine, mitomycin and ifosfamide [3]. A benefit has been obtained when cisplatin is combined with some of them. A major issue with this latter drug is the dosage, because higher doses of cisplatin are associated with delayed cumulative toxicity: our group has shown [4] that the risk of developing a World Health Organization (WHO) grade 2 neurological (polyneuropathy), auditory (hypoacousia) or renal (renal failure) toxicity after six courses of high-dose cisplatin (120 mg/m2) is ~25%. This toxicity precludes further therapy and probably reduces the impact of cisplatin therapy on survival. Replacement of cisplatin by its better-tolerated analogue, carboplatin, has been disappointing because it induces less favourable results, at least in terms of response rate [5]. However, if moderate dosages of carboplatin (200 mg/m2) and cisplatin (60 mg/m2) are combined [4], the same activity as high doses of cisplatin (120 mg/m2) is maintained in terms of response and survival, with a reduced risk of long-term renal, neurological or auditory toxicity (6% after six courses of therapy). In this regimen, carboplatin 100 mg/m2 is considered equivalent to cisplatin 30 mg/m2. As discussed previously [4], this equivalence is extrapolated from experimental data, as no clinical data are available. When combined with ifosfamide, this regimen induces a much better response rate [6]. Combination of cisplatin and ifosfamide with or without mitomycin can be considered one of the best first-generation regimens [79].

New active drugs against NSCLC were identified in the 1990s: gemcitabine, paclitaxel, vinorelbine, docetaxel and irinotecan [10]. Among them, gemcitabine appears to be particularly promising because of its relatively better tolerance. It closely resembles cytosine–arabinoside (Ara C) in that it acts as a pyrimidine antimetabolite. Multiple phase II trials, summarised in a recent systematic review of the literature [10], have shown that gemcitabine is active against NSCLC, with a response rate of ~20%, and that its toxicity profile is favourable in comparison with the cisplatin–etoposide regimen [11, 12]. Gemcitabine has been combined with various other cytotoxic agents, including cisplatin, carboplatin and ifosfamide [10]. In a pilot study, we demonstrated that it can be administered with an acceptable toxicity in association with a combination of moderate doses of cisplatin and carboplatin or with ifosfamide [13].

On the basis of the aforementioned data, we designed a three-arm, phase III, randomised trial to compare the combination of gemcitabine with ifosfamide (IG) or with carboplatin–cisplatin (CCG) to our standard regimen, cisplatin–carboplatin–ifosfamide (CCI). Primary end point was survival improvement, whereas secondary end points were impact on response rates and toxicity comparisons.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Selection criteria
To enter the study, patients with histologically proven NSCLC had to fulfil all the following criteria: inoperable stage IV or stage IIIB (according to Union Internationale Contre le Cancer, International Society of Surgery, 1987 classification) with malignant pleural effusion; presence of a measurable or assessable lesion; no prior history of malignancy except non-melanoma skin cancer, in situ carcinoma of the cervix or ‘cured’ malignant tumour (>5-year disease-free survival); no prior chemotherapy; Karnofsky performance status (PS) >=60; good renal (serum creatinine level <1.5 mg/dl and/or creatinine clearance >=60 ml/min), hepatic (serum bilirubin level <1.5 mg/dl) and haematological [white blood cell (WBC) count >=4000/mm3 and platelet count >=100 000/mm3] functions; absence of hypoacousia and peripheral neuropathy; no recent (<3 months before the date of treatment) myocardial infarction and no active congestive heart failure or cardiac arrhythmia requiring medical treatment; no uncontrolled infectious disease; and no other serious medical or psychological factors that may prevent adherence to the treatment schedule. Patients had to be available for follow-up and informed consent had to be provided. The protocol had to be approved by the ethical committee of each participating centre.

Treatment
Eligible patients were randomised between three arms: (i) a CCI regimen consisting of moderate doses of cisplatin (60 mg/m2) and carboplatin (AUC 3 mg.min/ml) combined with ifosfamide (4.5 g/m2), which was given on day 1; (ii) a CCG regimen consisting of moderate doses of cisplatin (60 mg/m2 on day 1) and carboplatin (AUC 3 mg.min/ml on day 1) combined with gemcitabine (1 g/m2 on days 1, 8 and 15); and (iii) an IG regimen consisting of a combination of ifosfamide (4.5 g/m2 on day 1) plus gemcitabine (1 g/m2 on days 1, 8 and 15).

All drugs were given intravenously (i.v.). Carboplatin (Paraplatin®, Bristol-Myers Squibb, New York, NY) was administered in ready-to-use solution over 30 min, just before cisplatin infusion. Dosage (AUC 3 mg.min/ml) was calculated according to the Calvert formula and glomerular filtration rate according to the Cockcroft formula, as described previously [14]. Cisplatin (Platinol®, Bristol-Myers Squibb, New York, NY) was administered over 1 h in 250 ml NaCl 3%. Platinum derivatives administration, in the CCI arm, was followed by i.v. infusion of 2000 ml NaCl 0.9% with 3 g KCl over 16 h and preceded by the infusion of ifosfamide (Holoxan®, ASTA Medica, Frankfurt, Germany) in 1000 ml NaCl 0.9% over 3 h. In the CCG arm, it was preceded by the infusion of 1 l dextrose 5% and NaCl 0.45% over 3 h, and was followed by the infusion of 3 l dextrose 5% and NaCl 0.45% over 18 h. In the IG arm, ifosfamide was infused over 18 h in 1 l dextrose 5%. Mesna (Uromitexan®, ASTA Medica, Frankfurt, Germany) was administered to avoid urotoxicity of the dosage of 1 g/m2 in short i.v. infusion prior to ifosfamide infusion. This was followed, in the CCI regimen, by 6 g/m2 in a 21 h infusion (1.5 g/m2 over 3 h with ifosfamide and 4.5 g/m2 in saline over 18 h) and, in the CCG arm, by 6 g/m2 in saline over 18 h. Gemcitabine (Gemzar®, Eli Lilly, Indianapolis, IN) was administered in a 30 min infusion in 250 ml NaCl 0.9% prior to the administration of the other cytotoxic drugs. To control emesis, administration of a 5-HT3 antagonist (granisetron) was recommended.

Courses were given every 4 weeks. Tumour response was assessed after three full courses. In case of stable disease or progression, the treatment was discontinued. Responding patients were given additional courses until best response, progression (PG) or major toxicity.

The administration of the next course of chemotherapy required that haematological (neutrophils >2000/mm3 and platelets >100 000/mm3) and renal (serum creatinine <1.5 mg/dl) functions had recovered. If the delay between two courses was >6 weeks, the patient was taken off treatment. Gemcitabine was not administered on days 8 and 15 if neutrophils were <500/mm3 or platelets <50 000/mm3. If neutrophil nadir was <500/mm3 and platelet nadir <25 000/mm3, carboplatin dosage was reduced to AUC 1.5 mg.min/ml and ifosfamide dosage to 3 g/m2. If such toxicity occurred again, carboplatin was stopped and ifosfamide was given at 1.5 g/m2. If serum creatinine peak increased to between 1.5 and 3.0 mg/dl, cisplatin dosage was reduced to 50%. If creatinine t was >3.0 mg/dl, cisplatin was stopped. If hearing loss occurred or WHO grade 2 neurotoxicity was exceeded, cisplatin was also stopped.

Work-ups
The initial work-up consisted of: complete history and physical examination with weight, height and surface-area measurements; recording of performance status; chest X-ray and computed tomography (CT) scan; fibreoptic bronchoscopy with biopsy; bone scintigraphy; liver and adrenal CT scan or echography; brain CT scan or nuclear magnetic resonance (NMR); blood chemistries including complete blood-cell count, electrolytes, serum creatinine and liver function tests; and electrocardiogram.

Blood chemistries, chest X-ray and clinical examination were repeated before each course. Restaging based on all tests performed during the initial work-up was repeated after the first three courses, and also every three courses if treatment continued.

After treatment completion, patients were followed up every 2 months for the first 6 months, and then every 3 months, with clinical examination, chest X-rays and biological tests.

Criteria of evaluation
Patients were evaluated for response after completion of three courses of chemotherapy. Responses were assessed by at least three independent observers during regular meetings of the group. Complete remission (CR) was defined as the disappearance of all signs of disease, for at least 4 weeks. Partial response (PR), in measurable disease consisted of >=50% decrease in total tumour load as established in two observations not less than 4 weeks apart, in the absence of new lesions or progression of any existing lesion. Tumour load was estimated as the tumour area calculated by the multiplication of the longest diameter by the greatest perpendicular diameter. In assessable disease, PR was defined as an estimated decrease in tumour size of >=50%. PG was defined as an increase of >25% in one or more measurable or assessable lesions, or the appearance of new lesion(s). All other circumstances were classified as no change (NC). Patients with early death (ED) due to disease progression before evaluation, those with toxic death due to chemotherapy or those with early chemotherapy stopping for toxicity were considered as ‘treatment failure’ and incorporated into the evaluable patient data.

Duration of response was calculated from the day of registration until the date of the first observation of PG. Survival time was dated from the day of registration. WHO criteria were used to assess toxicity.

Primary end point and sample-size determination
The primary objective of the trial was to compare survival times between the three regimens used. The sample-size evaluation was based on this end point. On the basis of the group’s previous experience, a median survival time of 24 weeks was expected for the CCI regimen (arm A). It was judged that an extension to 44 weeks would be clinically meaningful. With these assumptions, and using a two-sided log rank test for the primary comparison of the three study regimens ({alpha} = 5%, ß = 20%), 81 eligible patients needed to be randomised in each arm of the trial and analysis could be performed after observation of 162 deaths [15]. A preliminary analysis of toxicity was performed after inclusion of the first 75 patients in order to confirm the data of our previous pilot study [13]. No interim analysis for survival was planned or performed.

Randomisation procedure
Randomisation was performed centrally using the minimisation technique and stratified according to centre, Karnofsky PS, presence of brain metastases and prior chest irradiation. Treatment allocation was obtained by calling the ELCWP (European Lung Cancer Working Party) data centre.

Statistical methodology
Survival curves were estimated by the method of Kaplan–Meier. The log-rank test was used to compare survival curves. P values for testing differences between proportions were calculated with chi-square tests or with Fischer’s exact tests. A multivariate analysis for adjustment of the treatment effect, taking into account prognostic factors, was performed by fitting the data with a Cox model for duration of survival and a logistic regression model for objective response. The result of a statistical test was considered significant when P <0.05. All reported P values are two sided.

The evaluation of chemotherapy intensity was performed by calculating absolute dose intensity (ADI). The ADI was defined as the ratio of the received dose per square metre of body surface to the actual duration of treatment: it is expressed in mg/m2/week. All relevant formulae have been published previously [16]. Comparisons of the distributions of the dose-intensity variables between regimens were made using Mann–Whitney tests.


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Patient population and characteristics
A total of 284 patients were randomised between January 1998 and March 2000. Four (1.4%) were ineligible for the study (three in the CCG arm and one in the IG arm) for the following reasons: small-cell lung cancer histology, prior chemotherapy administration, increased bilirubinemia prior to randomisation, absence of informed consent. Of the 280 patients assessable for survival, 16 (5.7%) were not assessable for response (seven in the CCI arm, four in the CCG arm and five in the IG arm) due to loss of follow-up (n = 4), major protocol violation (n = 7), death prior to treatment (n = 1), over-long delay between two courses of therapy (n = 1) and patient’s refusal of further treatment (n = 3). Thus, 264 patients were assessable for response.

Characteristics of the eligible patients are listed in Table 1. The three arms were well balanced for the main patient characteristics. There were 94 patients in the CCI arm, 92 in the CCG arm and 94 in the IG arm. The majority of the patients were male (85%) and had a good Karnofsky PS (>=80 in 73%). Only six (2%) presented with stage IIIB disease, by virtue of the presence of a malignant pleural infusion, and 10 (3.5%) had prior chest irradiation. Histology was adenocarcinoma in 51% and squamous cell carcinoma in 30%. Thirty per cent of patients had brain metastases.


View this table:
[in this window]
[in a new window]
 
Table 1. Patient characteristics
 
Median follow-up duration was 98 weeks (range 0–159). At the time of the present analysis, 223 patients were dead, 45 were alive and 12 had been lost to follow-up for survival evaluation.

Tumour response
As shown in Table 2, there was a 25% [95% (CI) 16% to 34%] overall response rate for CCI regimen (23% of eligible patients; 95% CI 15% to 32%), 31% (95% CI 21% to 40%) for CCG (29% of eligible patients; 95% CI 20% to 39%) and 26% (95% CI 17% to 35%) for IG (24% of eligible patients; 95% CI 16% to 33%). These differences were not statistically significant (P = 0.68 for evaluable patients and P = 0.61 for eligible patients). In none of the subgroups (sex, age, Karnofsky PS, type of lesion, weight loss, histology, stage, presence of brain metastases, prior chest irradiation, WBC and neutrophil count) was there a significant difference in response rates between the three arms.


View this table:
[in this window]
[in a new window]
 
Table 2. Evaluation of response
 
By univariate analysis performed on all eligible patients (Table 3), no significant variable was identified. A logistic regression model using a backward selection with sex and Karnofsky PS failed to reveal significant independent predicting factors.


View this table:
[in this window]
[in a new window]
 
Table 3. Univariate prognostic factors analysis
 
Duration of response
Duration of response was not significantly different between the arms with a median of 29 weeks (95% CI 20–39 weeks) for CCI, 54 weeks (95% CI 33–73 weeks) for CCG, and 38 weeks (95% CI 27–50 weeks) for IG (P = 0.28). When arms were compared two by two, there was also no significant difference: P = 0.16 for CCI versus CCG, P = 0.20 for CCI versus IG and P = 1 for CCG versus IG. Progression-free survival was also not significantly different between the three treatments.

Survival
There was no difference (P = 0.20) in survival between the three study arms (Figure 1): median survival was 24 weeks (95% CI 19–29 weeks) for CCI, 34 weeks (95% CI 28–39 weeks) for CCG arm and 30 weeks (95% CI 27–32 weeks) for IG, respectively, with 1-year survival rates of 23, 33 and 35%, and 2-year survival rates of 11, 14 and 17%, respectively. When survival distributions between arms were compared two by two, results were also not statistically significant: P = 0.14 for CCI versus CCG, P = 0.12 for CCI versus IG, and P = 0.85 for CCG versus IG. At the time of analysis, 78 deaths had been observed in the CCI arm, 72 in the CCG arm and 73 in the IG arm. Survival analysis for patient subgroups is summarised in Table 4. There were significant differences not favouring the CCI regimen for women and patients aged >=60 years. Responder survivals were not statistically different (P = 0.44), with median survival times of 71 weeks for CCI, 65 weeks for CCG and 62 weeks for IG.



View larger version (18K):
[in this window]
[in a new window]
 
Figure 1. Survival according to treatment.

 

View this table:
[in this window]
[in a new window]
 
Table 4. Sub group analysis of survival
 
Some variables were identified by univariate analysis (Table 3) as significantly associated with improved survival: Karnofsky {Pi}{Sigma} >=80, weight loss <5%, absence of brain metastases, WBC <10 000/mm3 and neutrophils <7500/mm3. A Cox model, using all variables with a P value <0.05 in univariate analysis, was used for 236 patients (187 deaths) for whom data were available, and it identified two significant independent factors for survival: Karnofsky PS (hazard ratio 1.92 in favour of PS >=80; 95% CI 1.41–2.71; P <0.001) and neutrophil count (hazard ratio 1.45 in favour of count <7500/mm3; 95% CI 1.08–1.95; P = 0.02).

Toxicity
As summarised in Table 5, the main toxicities were in emesis, alopecia, leucopenia and thrombocytopenia. Leucopenia was more frequent in the IG arm while thrombopenia occurred more often in the CCG arm. Alopecia was less common in the CCG arm. There were 13 toxic deaths: five in the CCI arm (one of pericardial bleeding, four of febrile neutropenia), four in the CCG arm (all of febrile neutropenia) and four in the IG arm (one of encephalopathy and three of febrile neutropenia). In addition, seven sudden deaths were observed and were related to cardiovascular events; three in the CCI arm and four in the CCG arm.


View this table:
[in this window]
[in a new window]
 
Table 5. Main toxicity for the first six courses per patient per arm
 
Dose intensity
Overall treatment duration was not significantly different between the three treatment arms (P = 0.07), with a median duration of 84 days (range 0–202 days) for CCI, 86 days (range 28–233 days) for CCG and 84 days (range 28–221 days) for IG. Twenty-six, 26 and 24 patients received at least six courses of chemotherapy, respectively.

Patients in the CCI arm received a significantly higher dose-intensity of cisplatin than in the CCG arm (14.9 versus 14.6 mg/m2/week; P = 0.007), but the difference was not significant for carboplatin (24.4 versus 21.9 mg/m2/week; P = 0.13). Patients in the CCI arm also had a significantly higher dose-intensity of ifosfamide than those in the IG arm (1.1 versus 1.04 g/m2/week; P = 0.02). There was no significant dose-intensity difference for gemcitabine between CCG and IG (0.60 versus 0.58 g/m2/week; P = 0.53).

Treatment at relapse
Second-line chemotherapy (left to the discretion of the local investigator) was administered to 36 patients in the CCI arm, to 33 in the CCG arm and to 38 in the IG arm. It consisted principally of taxanes (21 patients) or gemcitabine (11 patients) after CCI, of taxanes (24 patients) after CCG and of taxanes (18 patients) or platinum-based regimens (18 patients) after IG.


    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
The two arms including gemcitabine, a second-generation cytotoxic agent for NSCLC, were associated with better survival in comparison with the standard-arm combination of platinum derivatives and ifosfamide. The difference was not statistically significant (P = 0.20) for the overall three-arm comparison. Some tests gave better results in some subgroups (women, patients older than 60 years) but they must be considered as exploratory only. The main explanation for not reaching statistical significance is probably a lack of trial power. Assuming that the observed difference is a true one, 120 patients per arm should have been randomised to detect it. This represents a 30% increase in our sample size.

The reference regimen used in this trial was a combination of cisplatin, carboplatin and ifosfamide, based on the results of a previous trial [6] as mentioned in the introduction. This regimen is similar to those (cisplatin and mitomycin plus ifosfamide or vindesine or vinblastine) shown to be superior compared with the standard of the 1980s, cisplatin plus etoposide [7]. We have not included mitomycin in the regimen because there are no data to demonstrate a potential benefit when it is added to a cisplatin combination [17]. Indeed, all the randomised trials testing this failed to show any advantage. It can be argued that carboplatin was not necessary either. A recent randomised trial by our group [9], comparing MIP (mitomycin, ifosfamide and cisplatin) and CarboMIP (MIP plus carboplatin) demonstrated no difference between the two combinations. These results were not available when we designed the present study and thus we could not integrate the information into our protocol decisions.

While there are published trials comparing the combination of cisplatin and second-generation drugs (vinorelbine, paclitaxel and gemcitabine) with cisplatin plus etoposide or teniposide [1821], without survival benefit, very few studies have so far been published that compare these with classical standards like MIP or MVP (mitomycin–vindesine or vinblastine–cisplatin). Crino and colleagues compared the cisplatin–gemcitabine regimen with MIP [22] and found no significant advantages in terms of survival or quality of life (the primary end point of the trial), despite a better response rate in favour of the new drug regimen. Patient selection in the Crino trial differed from our study in that it included patients with stage IIIB disease. As no subgroup analysis was reported, it is difficult to perform an extensive comparison of the results of the two trials.

Our study is one of the first randomised trials assessing a non-cisplatin (or carboplatin) regimen. A recently published randomised trial compared cisplatin–docetaxel to docetaxel–gemcitabine and found no difference between the two [23]. As no first-generation standard control arm was used, it is difficult to assess the practical message of this trial because we do not know the exact value of the cisplatin–docetaxel combination in this indication. It is interesting to observe that the ifosfamide–gemcitabine combination was equivalent to the CCG regimen and better than the CCI regimen. Since platinum salvage chemotherapy was administered in <20% of the non-responders (18/94), this cannot be the only explanation for the observed effect. The trends observed in our trial are at variance with those reported during the 2001 ASCO meeting, where paclitaxel–gemcitabine [24] and gemcitabine–vinorelbine combinations [25] were associated with worse results than cisplatin-including regimens. The use of ifosfamide in our non-platinum combination could potentially explanation the better results.

In terms of toxicity, the IG arm was associated with more severe leucopenia while the CCG one was associated with more severe thrombopenia than can be explained by an effect of carboplatin. Alopecia was less common with CCG. We observed seven sudden deaths, probably due to acute cardiovascular complications. Since they were only observed in the platinum-containing arms, there may have been an adverse effect of platinum compound, as found previously in a phase II trial assessing cisplatin, 5-fluorouracil, mitomycin and vindesine [26] with a 10% sudden death rate. In the present trial, the rate was 4 and 3% for the CCG and CCI arms, respectively; because this may have been due to toxicity, we have considered these patients as assessable for response as those with toxic death. It could be argued that another potential advantage of non-platinum regimens could be the avoidance of this lethal cardiovascular toxicity.

We have not evaluated quality of life of the patients in our trial because it is time-consuming and difficult to perform adequately. Most studies that try to assess this variable in fact report analysis of subsets of patients, and results have to be generalised for the whole patient population. Practically, quality of life assessment is performed mainly for drug registration purposes. In large trials, it is often restricted to symptoms and toxic effects evaluation [8], and analysis is difficult to interpret due to the amount of missing data.

Finally, while we did not find in the present trial significant predicting factors for responses, we did identify performance status and neutrophil count as independent variables for survival. These latter findings have consistently been found to be a significant predictor in our previous studies [9, 27, 28]. We had a high rate of brain metastases but, as in our prior reports, their presence was not a prognostic factor, showing that inclusion of such patients if they have an adequate PS is not a source of bias. Based on these results, we now use neutrophil count in our stratification procedure for randomisation.

In conclusion, this trial in metastatic NSCLC has shown some survival advantages for the combination of gemcitabine with platinum derivatives or with ifosfamide in comparison with a first-generation drugs regimen with platinum derivatives and ifosfamide. Statistical significance has not been reached in the overall survival comparison, probably due to the insufficient power of the trial. Interestingly, the non-platinum combination IG provided similar results to the platinum regimen CCG, suggesting that new drug combinations not including platinum can be considered for advanced NSCLC. This observation requires further confirmation by other randomised controlled trials.


    Acknowledgements
 
The following institutions participated in the trial: Institut Jules Bordet (J.P. Sculier, M. Paesmans, T. Berghmans, P. Mommen, J. Klastersky), Brussels, Belgium; C.H.U. de Charleroi (J. Thiriaux, J. Lecomte), Charleroi, Belgium; C.H.U. de Lille, Hôpital Calmette (J.-J. Lafitte), Lille, France; Evangelismos General Hospital (C.G. Alexopoulos, M. Vaslamatzis), Athens, Greece; Hellenic Cancer Institute (A. Efremidis, G. Koumakis), Athens, Greece; C.H. de Douai (M.C. Florin, E. Maetz), Douai, France; C.H. de Tivoli (J. Michel), Tivoli, Belgium; Hospital de Sagunto (V. Giner-Marco), Valencia, Spain; Hôpital d’Hayange (M.C. Berchier), Hayange, France; C.H. de Roubaix (F. Kroll), Roubaix, France; C.H.U. A. Vésale (D. Brohée), Montignies-le-Tilleul, Belgium; C.H.I. de Montfermeil (C. Zacharias), Montfermeil, France; R.H.M.S (A. Tagnon, V. Richard), Tournai, Baudour, Belgium; Hôpital de Warquignies (P. Libert, M. Richez), Warquignies, Belgium; C.H. de Mons (P. Recloux), Mons, Belgium; C. H. Dr Schaffner (J. Amourette), Lens, France; Klinika Radiotherapie a Onkologie (J. Baumöhl), Kosice, Slovakia; C.H. de Tourcoing (X. Ficheroulle), Tourcoing, France; Cabinet de Pneumologie (Y. Watrigant), Tourcoing, France; Hôpital Duchenne (J.L. Crepin), Boulogne-sur-Mer, France; Hôpital Brugmann (A. Drowart), Brussels, Belgium; Hôpital Erasme (S. Luce, F. Branle), Brussels, Belgium; Clinique Teissier (G. Demarc, M.-L. Line) Valenciennes, France; Polyclinique de Henin Beaumont (R. Roboubi), Henin Beaumont, France; C.H. d’Arras (J.-F. Bervar), Arras, France.


    Footnotes
 
+ Correspondence to: Dr J. P. Sculier, Institut Jules Bordet, 1, rue Héger-Bordet, B-1000 Bruxelles, Belgium. Tel: +32-2-541-31-85; Fax: +32-2-534-37-56; E-mail: sculier@bordet.be Back

§ Participants are listed in the Acknowledgements. Back


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1. Non-small Cell Lung Cancer Collaborative Group. Chemotherapy in non-small cell lung cancer: a meta-analysis using updated data on individual patients from 52 randomised clinical trials. Br Med J 1995; 311: 899–909.[Abstract/Free Full Text]

2. Sculier JP, Berghmans T, Castaigne A et al. Best supportive care or chemotherapy for stage IV non-small-cell lung cancer. In Van Houtte P, Klastersky J, Rocmans, P (eds): Progress and Perspectives in Lung Cancer. Berlin, Germany: Springer Verlag 1999; 199–207.

3. Donnadieu N, Paesmans M, Sculier JP. Chemotherapy of non-small-cell bronchial cancers. Meta-analysis of the literature as a function of the extent of the disease. Rev Mal Respir 1991; 8: 197–204.[ISI][Medline]

4. Sculier JP, Klastersky J, Giner V et al. Phase II randomized trial comparing high-dose cisplatin with moderate-dose cisplatin and carboplatin in patients with advanced non-small-cell lung cancer. European Lung Cancer Working Party. J Clin Oncol 1994; 12: 353–359.[Abstract]

5. Klastersky J, Sculier JP, Lacroix H et al. A randomized study comparing cisplatin or carboplatin with etoposide in patients with advanced non-small-cell lung cancer: European Organization for Research and Treatment of Cancer Protocol 07861. J Clin Oncol 1990; 8: 1556–1562.[Abstract]

6. Sculier JP, Paesmans M, Thiriaux J et al. Phase III randomized trial comparing cisplatin and carboplatin with or without ifosfamide in patients with advanced non-small-cell lung cancer. European Lung Cancer Working Party. J Clin Oncol 1998; 16: 1388–1396.[Abstract]

7. Crino L, Clerici M, Figoli F et al. Chemotherapy of advanced non-small-cell lung cancer: a comparison of three active regimens. A randomized trial of the Italian Oncology Group for Clinical Research (GOIRC). Ann Oncol 1995; 6: 347–353.[Abstract]

8. Cullen MH, Billingham LJ, Woodroffe CM et al. Mitomycin, ifosfamide, and cisplatin in unresectable non-small-cell lung cancer: effects on survival and quality of life. J Clin Oncol 1999; 17: 3188–3194.[Abstract/Free Full Text]

9. Sculier JP, Lafitte JJ, Paesmans M et al. Phase III randomized trial comparing moderate-dose cisplatin to combined cisplatin and carboplatin in addition to mitomycin and ifosfamide in patients with stage IV non-small-cell lung cancer. Br J Cancer 2000; 83: 1128–1135.[ISI][Medline]

10. Meert AP, Berghmans T, Branle F et al. Phase II and III studies with new drugs for non-small-cell lung cancer: a systematic review of the literature with a methodology quality assessment [in process citation]. Anticancer Res 1999; 19: 4379–4390.[ISI][Medline]

11. Manegold C, Bergman B, Chemaissani A et al. Single-agent gemcitabine versus cisplatin-etoposide: early results of a randomised phase II study in locally advanced or metastatic non-small-cell lung cancer. Ann Oncol 1997; 8: 525–529.[Abstract]

12. Perng RP, Chen YM, Ming-Liu J et al. Gemcitabine versus the combination of cisplatin and etoposide in patients with inoperable non-small-cell lung cancer in a phase II randomized study. J Clin Oncol 1997; 15: 2097–2102.[Abstract]

13. Berghmans T, Klastersky J, Markiewicz E et al. Cisplatin–carboplatin–gemcitabine or ifosfamide–gemcitabine in advanced non-small-cell lung carcinoma: two pilot studies. Anticancer Res 1999; 19: 5651–5655.[ISI][Medline]

14. Sculier JP, Paesmans M, Thiriaux J et al. A comparison of methods of calculation for estimating carboplatin AUC with a retrospective pharmacokinetic–pharmacodynamic analysis in patients with advanced non-small-cell lung cancer. European Lung Cancer Working Party. Eur J Cancer 1999; 35: 1314–1319.[ISI][Medline]

15. Machin D, Campbell MJ. Statistical tables for the design of clinical trials. Oxford: Blackwell Scientific Publications 1987.

16. Sculier JP, Paesmans M, Bureau G et al. Randomized trial comparing induction chemotherapy versus induction chemotherapy followed by maintenance chemotherapy in small-cell lung cancer. European Lung Cancer Working Party. J Clin Oncol 1996; 14: 2337–2344.[Abstract]

17. Sculier JP, Ghisdal L, Berghmans T et al. The role of mitomycin in the treatment of non-small-cell lung cancer: a systematic review with meta-analysis of the literature. Br J Cancer 2001; 84: 1150–1155.[ISI][Medline]

18. Comella P, Frasci G, De Cataldis G et al. Cisplatin/carboplatin + etoposide + vinorelbine in advanced non-small-cell lung cancer: a multicentre randomised trial. Gruppo Oncologico Campano. Br J Cancer 1996; 74: 1805–1811.[ISI][Medline]

19. Giaccone G, Splinter TA, Debruyne C et al. Randomized study of paclitaxel–cisplatin versus cisplatin–teniposide in patients with advanced non-small-cell lung cancer. The European Organization for Research and Treatment of Cancer Lung Cancer Cooperative Group. J Clin Oncol 1998; 16: 2133–2141.[Abstract]

20. Cardenal F, Lopez-Cabrerizo MP, Anton A et al. Randomized phase III study of gemcitabine–cisplatin versus etoposide–cisplatin in the treatment of locally advanced or metastatic non-small-cell lung cancer. J Clin Oncol 1999; 17: 12–18.[Abstract/Free Full Text]

21. Bonomi P, Kim K, Fairclough D et al. Comparison of survival and quality of life in advanced non-small-cell lung cancer patients treated with two dose levels of paclitaxel combined with cisplatin versus etoposide with cisplatin: results of an Eastern Cooperative Oncology Group trial. J Clin Oncol 2000; 18: 623.

22. Crino L, Scagliotti GV, Ricci S et al. Gemcitabine and cisplatin versus mitomycin, ifosfamide, and cisplatin in advanced non-small-cell lung cancer: a randomized phase III study of the Italian Lung Cancer Project. J Clin Oncol 1999; 17: 3522–3530.[Abstract/Free Full Text]

23. Georgoulias V, Papadakis E, Alexopoulos A et al. Platinum-based and non-platinum-based chemotherapy in advanced non-small-cell lung cancer: a randomised multicentre trial. Lancet 2001; 357: 1478–1484.[ISI][Medline]

24. Van Meerbeeck JP, Smit EF, Lianes P et al. A EORTC randomized phase III trial of three chemotherapy regimens in advanced non-small-cell lung cancer (NSCLC). ASCO Proceedings 2001; 20: 308a (Abstr).

25. Alberola V, Camps C, Provencia M et al. Cisplatin/gemcitabine (CG) vs cisplatin/gemcitabine/vinorelbine (CGV) vs sequential doublets of gemcitabine/vinorelbine followed by ifosfamide/vinorelbine (GV/IV) in advanced non-small-cell lung cancer (NSCLC): results of a Spanish Lung Cancer Group phase III trial (GEPC/98–02). ASCO Proceedings 2001; 20: 308a (Abstr).

26. Klastersky J, Sculier JP, Ries F et al. A four-drug combination chemotherapy with cisplatin, mitomycin, vindesine and 5-fluorouracil: a regimen associated with major toxicity in patients with advanced non-small-cell lung cancer. Lung Cancer 1994; 11: 373–384.[ISI][Medline]

27. Paesmans M, Sculier JP, Libert P et al. Prognostic factors for survival in advanced non-small-cell lung cancer: univariate and multivariate analyses including recursive partitioning and amalgamation algorithms in 1052 patients. The European Lung Cancer Working Party. J Clin Oncol 1995; 13: 1221–1230.[Abstract]

28. Sculier JP, Paesmans M, Ninane V et al. Evaluation of the TN sub-staging in patients with initially unresectable stage III non-small-cell lung cancer treated by induction chemotherapy. The European Lung Cancer Working Party. Lung Cancer 1998; 22: 201–213.[ISI][Medline]