Phase I/II study of paclitaxel, gemcitabine and vinorelbine as first-line chemotherapy of non-small-cell lung cancer

Original article Annals of Oncology 13: 1862 1867, 2002 DOI: 10.1093/annonc/mdf308 Phase I/II study of paclitaxel, gemcitabine and vinorelbine as first-line chemotherapy of non-small-cell lung cancer V. Lorusso 1 *, E. Crucitta 1, N. Panza 2, N. Silvestris 1, M. Guida 1, F. Carpagnano 3, S. Mancarella 1, D. Sambiasi 1 & M. De Lena 1 1 Operative Unit of Medical Oncology, Oncology Institute of Bari, Bari; 2 Medical Oncology, Caldarelli Hospital, Naples; 3 Thoracic Surgery Department, San Paolo Hospital of Bari, Bari, Italy Received 18 February 2002; revised 7 May 2002; accepted 7 June 2002 Introduction Background: The aim of our study was to determine the maximum tolerated dose of paclitaxel combined with a fixed dose of gemcitabine and vinorelbine in the treatment of non-small-cell lung cancer (NSCLC) and to evaluate in a phase II trial the efficacy of this combination. Patients and methods: Sixty-two patients with stage IIIB/IV NSCLC were treated with paclitaxel in escalating doses from 40 80 mg/m 2 combined with gemcitabine and vinorelbine at fixed doses of 1000 mg/m 2 and 25 mg/m 2, respectively. All drugs were given intravenously on day 1 and 8 every 3 weeks. Results: In a phase I trial, carried out on 21 patients, grade 4 neutropenia, as dose-limiting toxicity, occurred at the dosage level of paclitaxel 80 mg/m 2. In a phase II trial, with paclitaxel administered at 70 mg/m 2, 27 out of 41 (66%) assessable patients responded (10% complete responses and 56% partial responses). Objective response was observed in 13 of 16 patients (81%) with stage IIIB disease and in 14 of 25 (56%) with stage IV disease. The median time to treatment failure was 26 weeks (range 3 72 weeks; 32 weeks and 20 weeks for stages IIIB and IV, respectively) and median survival 62 weeks (range 4 176 weeks; 72 weeks and 56 weeks for stages IIIB and IV, respectively). One-year survival was 64% for all patients (72% for patients with stage IIIB and 52% for those with stage IV). Grade 3 and 4 neutropenia were observed in 11 (27%) and seven (17%) cases, respectively; grade 3 thrombocytopenia was observed in three patients (7%) and grade 3 anemia in four patients (10%). The most relevant non-hematological toxicity was grade 2/3 asthenia, which was observed in 12 patients (29%). Alopecia was almost universal, whereas nausea and vomiting were absent. Conclusions: The combination of paclitaxel, gemcitabine and vinorelbine is effective and tolerable in the treatment of NSCLC. The high activity and low toxicity of this regimen warrant randomized studies with platinum-containing combinations. Key words: chemotherapy, gemcitabine, non-small-cell lung cancer, paclitaxel, vinorelbine In the past there was considerable pessimism about the role of chemotherapy in non-small-cell lung cancer (NSCLC). This pessimism largely developed because alkylating agent-based therapies produced low response rates and considerable toxicity. Moreover, these drugs worsened survival at all stages, even when used as postoperative adjuvant therapy [1 3]. The treatment of lung cancer improved considerably in the 1980s, when cisplatin-based chemotherapy was shown to prolong survival and improve quality of life in patients with stage III or *Correspondence to: Dr V. Lorusso, Operative Unit of Medical Oncology, Oncology Institute of Bari, Via Amendola 209, 70126 Bari, Italy. Tel: +39-80-5555257; Fax: +39-80-5555471; E-mail: vitolorusso@inwind.it IV NSCLC as shown by randomized trials and meta-analyses of these randomized trials [4]. However, this benefit of improved survival is limited, and a real impact on quality of life is debatable [5, 6]. In fact, approximately one-third to onehalf of patients with NSCLC treated in randomized trials experience some form of severe side-effect, particularly with the platinum-based regimens [7]. In particular, in one phase II trial, although there was a high objective response rate (48%) with a platinum combination, the patients general deterioration negated any potential advantage achieved by most patients, even in those with objective responses [8]. In recent years, a number of new non-platinum agents have demonstrated significant activity in advanced NSCLC. Among them there are the taxanes (paclitaxel and docetaxel), vinorelbine and gemcitabine [9]. Along with activity, in 2002 European Society for Medical Oncology

1863 general the novel agents have a toxicity profile that is superior to that of platinum. This is highly important in this setting in which overall quality of life is a major consideration, along with control of lung cancer symptoms. Combinations of these new drugs in doublets have yielded results comparable with those achievable with cisplatin-containing regimens. In particular, the combination of gemcitabine and vinorelbine proved feasible with myelosuppression being the most frequent toxicity. Most phase II trials reported response rates of 20 40% and median survival duration of 7 11 months [10 13]. The combination of vinorelbine plus paclitaxel showed promising results (32% response) in previously treated patients with small-cell lung cancer [14]. The efficacy and toxicity of combined paclitaxel and gemcitabine was evaluated in a phase II study on chemotherapy-naïve patients with advanced NSCLC [15]. The response rate observed was 36%, the median survival 51 weeks, and the toxicity negligible. Moreover, these results have recently been confirmed in a phase III trial comparing the combination of paclitaxel/ gemcitabine with paclitaxel/carboplatin [16]. With regard to the first experiences with a three drug combination without cisplatin, Masters et al. [17] obtained disappointing results with the combination of paclitaxel, ifosfamide and vinorelbine plus granulocyte colony-stimulating factor (G-CSF) support. In fact, this combination appeared only to exacerbate toxicity, with no increase in efficacy, as compared with the commonly used platinum-containing combinations. In our study we have employed a combination of paclitaxel, gemcitabine and vinorelbine which had never been tested before. The aim of this study was to determine the maximum tolerated dose (MTD) of paclitaxel (T) when combined with a fixed dose of gemcitabine (G) and vinorelbine (V) in the treatment of advanced non-small-cell-lung cancer and subsequently to evaluate in a large phase II study the efficacy of this combination (TGV). Patients and methods Eligibility Patients with histologically proven, measurable, locally advanced (stage IIIB) or metastatic (stage IV) NSCLC who had not previously received chemotherapy or radiotherapy were eligible for the study. The other eligibility criteria were age more than 18 years; Eastern Cooperative Onclogy Group performance status of 0 2; an expected survival duration of 3 months; good bone marrow reserve (absolute granulocyte count >4000/ml, platelet count >100000/ml and hemoglobin level >11 g/dl); adequate hepatic and renal function, including aspartate aminotransferase, alanine aminotransferase, bilirubin, alkaline phosphatase and serum creatinine values <1.25 the upper limit of normal; no severe uncontrolled co-morbidity; and no second malignancies. Written informed consent was obtained from all participating patients before study entry. The pretreatment evaluation included medical history and physical examination with tumor measurements; chest radiograph; complete blood cell count; serum chemistries; liver function tests; and staging studies appropriate to define the extent of metastatic disease, including bronchoscopy, abdominal ultrasound, thoracic and/or abdominal computed tomography, and bone scan, as indicated by clinical or laboratory examination. Treatment plan Escalating doses of paclitaxel were administered intravenously by 3-h infusion followed by a fixed dose of gemcitabine 1000 mg/m 2 over 30 min and vinorelbine 25 mg/m 2 over 20 min on days 1 and 8. The starting dose of paclitaxel was 40 mg/m 2 (dose level 1), and the subsequent dose levels were 50, 60, 70 and 80 mg/m 2 (dose levels 2, 3, 4 and 5, respectively). The chemotherapy cycle was administered every 3 weeks to a maximum total of eight cycles for patients with complete response (CR), partial response (PR) or stable disease (SD). Whenever disease progression occurred, patients were treated with a platinum-containing regimen as second-line chemotherapy. Patients with stage IIIB disease suitable for radiotherapy received the chemotherapy program as a part of a combined treatment that consisted of induction chemotherapy (six cycles) followed by radical radiotherapy (50.4 Gy, given in daily fractions of 1.8 Gy). With regard to dose reduction, no dose modifications were allowed in the phase I study. On the other hand, in the phase II study dose reductions were allowed as follows: full doses of chemotherapy were given if the neutrophil and platelet counts on the day of treatment were >2 10 9 /l and 100 10 9 /l, respectively. If neutropenia or thrombocytopenia of WHO grade 1 or higher occurred on day 1, the administration of chemotherapy was delayed for 1 week. If, after this delay, grade 1 neutropenia or thrombocytopenia persisted, the treatment was given at doses reduced to 75% of that originally planned. If grade 1 neutropenia or thrombocytopenia occurred on day 8, chemotherapy was given without any delay at doses reduced to 75%. In the presence of grade 2 or greater neutropenia or thrombocytopenia on either day 1 or 8, the treatment was omitted. Finally, if grade 4 neutropenia or thrombocytopenia lasted longer than 7 days or if grade 3 or 4 non-hematological toxicity occurred at any time during the treatment, the doses of each drug were resumed at a 75% dose level in all subsequent chemotherapy administrations. The treatment was definitely discontinued if neutropenia or thrombocytopenia, greater than grade 1, or major non-hematological toxicity, persisted for 3 or more weeks after the planned time for the schedule to be recycled. Toxicity and evaluation of response Clinical monitoring was performed twice weekly with a complete blood cell count. Toxicity was scored using WHO criteria [18]. In the phase I part of the study, where no dose reductions were allowed, a minimum of three patients were treated at each dose level. To determine the MTD of this combination, toxicity was assessed by one of the following events: (i) grade 4 thrombocytopenia for 1 day or grade 4 leukopenia for 4 days; (ii) grade 2 renal or liver dysfunction; (iii) any other non-hematological, non-hepatic and non-renal grade 3 toxicity, excluding alopecia; or (iv) inability to deliver full doses of both drugs on day 8 of the cycle of chemotherapy because of grade 2 4 leukopenia or thrombocytopenia. If any of these side-effects developed in one of three patients, one patient was added. If dose-limiting toxicity was observed in two out of three or four patients, a fifth patient was added. If dose-limiting toxicity was observed in none of three patients, one out of four patients, or two out of five patients at a single dose level, the dose of paclitaxel was increased by 10 mg/m 2. If dose-limiting toxicity was observed in three out of three, four or five patients, the MTD was reached, thus the previous dose level was that recommended for the phase II trial (best-of-five schedule). Antitumor activity was evaluated every two courses (i.e. 6 weeks) on all measurable lesions, and all patients were scheduled for at least two cycles to be

1864 eligible for assessment of tumor response. In patients with tumor responses or SD, treatment was continued to a maximum of eight cycles. Tumor response was classified according to WHO criteria [18] and documented by two investigations at least 6 weeks apart. The Kaplan Meier method was used to analyze time to disease progression and overall survival. Confidence intervals for the response rates were calculated using the method described by Simon [19]. P values were calculated by Fisher exact tests. Results From February 1998 to December 2000, 62 patients with a median age of 59 years (range 39 69), were enrolled into this phase I/II study: 21 patients in the phase I study (three patients at dose levels 1 and 2, five at dose levels 3, 4 and 5) and 41 in the subsequent phase II study. Dose escalation phase Twenty-one patients, not previously submitted to chemotherapy, were enrolled into the phase I portion of the study in five cohorts. Patients received a total of 88 courses of chemotherapy. The median age was 58 years (range 39 67) and most patients were male (86%). Performance status was 0, 1 and 2 in seven (33%), five (24%) and nine (43%) patients, respectively. Fourteen patients (67%) had stage IV disease and seven (33%) stage IIIB. Nine patients (43%) had adenocarcinoma, eight (38%) squamous and four (19%) undifferentiated histological type. Characteristics of patients are listed in Table 1. No dose-limiting toxicity was observed in the first three patients treated at dose levels 1 (paclitaxel 40 mg/m 2 ) and Table 1. Patient characteristics of phase I No. of patients (%) Enrolled 21 Sex Male 18 (86) Female 3 (14) Median age, years (range 39 67) 58 ECOG performance status 0 7 (33) 1 5 (24) 2 9 (43) Histology Squamous 8 (38) Adenocarcinoma 9 (43) Large-cell undifferentiated 4 (19) Stage IIIB 7 (33) IV 14 (67) Weight loss >5% 4 (19) ECOG, Eastern Cooperative Oncology Group. 2 (paclitaxel 50 mg/m 2 ). Two of five patients developed grade 4 neutropenia at dose levels 3 (paclitaxel 60 mg/m 2 ) and 4 (paclitaxel 70 mg/m 2 ). Dose-limiting toxicity occurred at the dosage of paclitaxel 80 mg/m 2 with three of five patients developing grade 4 neutropenia. After the enrolment of 21 patients, the results of the phase I study showed that the MTD of paclitaxel in combination with fixed doses of gemcitabine 1000 mg/m 2 and vinorelbine 25 mg/m 2 was 70 mg/m 2. With regard to toxicity observed other than grade 4 neutropenia, grade 2 and 3 neutropenia were observed in 10 patients (48%). Moreover, grade 2 anemia and grade 2 thrombocytopenia were observed in two (10%) and three (14%) patients, respectively. The most common non-hematological toxicity was grade 2 asthenia complained of by seven patients (33%) (three patients treated with 70 mg/m 2 and two patients treated with 60 and 80 mg/m 2 ). With regard to clinical responses, although not required in phase I evaluation, we observed one CR and ten PRs (52.4%). Responses were observed in all dose levels. Moreover, median duration of response was 24 weeks (range 8 56) and median survival 44 weeks (range 4 63). Phase II Forty-one patients were enrolled into this trial and all were assessable for response and toxicity. Patients received a total of 189 courses of chemotherapy. The median age was 60 years (range 43 69). Male:female ratio was 34:7. Moreover, performance status was 0 or 1 in 31 patients (76%) and 2 in 10 patients (24%). Twenty-five patients (61%) had stage IV disease and 16 (39%) stage IIIB. Twelve patients (29%) had adenocarcinoma, 26 (64%) squamous and three (7%) undifferentiated histological types. Fourteen patients (34%) complained of weight loss >10% in the last 6 months. Patient characteristics are reported in Table 2. With regard to clinical response (Table 3), PR was observed in 23 patients (56%) and CR in four patients (10%). It must be emphasized that 13 of 16 patients (81%) with stage IIIB responded as compared with 14 of 25 (56%) stage IV patients. SD was observed in only three patients (7%). Survival The median time to treatment failure (Figure 1) was 26 weeks (range 3 72 weeks; 32 weeks and 20 weeks for stages IIIB and IV, respectively) and median survival (Figure 2) was 62 weeks (range 4 176 weeks; 72 weeks and 56 weeks for stage IIIB and IV, respectively). One-year survival was 64% for all patients (72% for patients with stage IIIB and 52% for those with stage IV). It is worth noting that median survival was 74 weeks for patients with performance status of 0 or 1 and 28 weeks for patients with a performance status of 2 (P = 0.025). All the 41 assessable patients treated in this study received at least one cycle of therapy with a maximum number of eight cycles. The full dose was given in 23 patients (56%). Relative dose intensity, calculated taking into account delays and dose reductions was 87.5%.

1865 Table 2. Patient characteristics of phase II No. of patients (%) Enrolled 41 Sex Male 34 (83) Female 7 (17) Median age, years (range 43 73) 60 ECOG performance status 0 16 (39) 1 15 (37) 2 10 (24) Histology Squamous 26 (64) Adenocarcinoma 12 (29) Large-cell undifferentiated 3 (7) Stage IIIB 16 (39) IV 25 (61) Weight loss >5% 14 (34) ECOG, Eastern Cooperative Oncology Group. Table 3. Results of phase II Response No. of patients (%) Complete response 4 (10) Partial response 23 (56) Stable disease 3 (7) Progression 11 (27) Total 41 (100) Hematological toxicity With regard to toxicity, neutropenia was the most common hematological toxicity observed, with grade 3 and 4 occurring in 11 (27%) and seven (17%) of cases, respectively. Febrile leukopenia was observed in two out of seven patients with grade 4 leukopenia. The administration of G-CSF was not planned in the protocol. However, in order to deliver drugs on time, G-CSF was administered to the 11 patients with grade 3 and to one patient with grade 4 neutropenia. Moreover, grade 3 and 4 thrombocytopenia were observed only in two patients (5%) and one patient (2%), respectively. Grade 3 anemia was observed in only four patients (10%). Non-hematological toxicity For the non-hematological toxicity, asthenia was observed in 12 patients (29%) (grade 1 or 2 in nine patients and grade 3 in three patients) (Table 4). Alopecia was almost universal Figure 1. Time to treatment failure. Figure 2. Overall survival. whereas nausea and vomiting virtually absent. One patient had a hypersensitivity reaction after the first paclitaxel administration and continued to be treated without this drug. Symptom control Quality of life was not evaluated in this trial. However, particular attention was paid to the cancer-related symptoms: cough, dyspnea, pain, loss of appetite, fatigue and performance status. Overall, more than 70% of patients showed an objective improvement of one or more of these parameters with negligible side-effects. Second-line therapy With regard to second-line therapy, all patients primarily refractory to TGV were submitted to a platinum-containing chemotherapy if, after two cycles, at first disease re-evaluation, progressive disease was detected. On the other hand, out of 27 patients responding to TGV, only in 12 was it possible to administer a platinum-containing therapy. In fact, most of these patients progressing after a long therapy-free interval refused second-line therapy, or were in poor performance status condition, not allowing platinum therapy. It is also worth noting that no patient primarily refractory to TGV responded to a platinum-containing combination administered as second-line therapy (generally a combination of cisplatin, ifosfamide and etoposide or cisplatin, vinblastine and mitomycin). Conversely, out of 12 patients who had responded to TGV, eight responded again when treated with

1866 Table 4. Toxicity in 41 patients in phase II cisplatin combinations at the time of relapse, prolonging their survival by 6 12 months. Discussion Grade 1 2 3 4 No. % No. % No. % No. % Neutropenia 0 0 0 0 11 27 7 17 Thrombocytopenia 0 0 0 0 2 5 1 2 Anemia 0 0 0 0 4 10 0 0 Neurological toxicity 0 0 1 2 0 0 0 0 Stomatitis 2 5 0 0 0 0 0 0 Transaminase 2 5 1 2 0 0 0 0 Asthenia 5 12 4 10 3 7 0 0 Vomiting 2 5 0 0 0 0 0 0 In patients with advanced NSCLC, systemic combination chemotherapy has offered a modest but consistent prolongation in survival compared with supportive care [4 7]. The observed improvement has been primarily attributed to cisplatin, which remains the recommended standard of care for patients with NSCLC [20]. However, in advanced metastatic disease in which palliation is the goal, a balance among activity, toxicity and quality of life must be achieved. In this regard, the unfavorable side-effect profile of cisplatin often makes the administration of chemotherapy unacceptable, difficult or sometimes impossible. In recent years, a number of new agents have demonstrated high single-agent activity and low-toxicity profiles in the treatment of advanced NSCLC. The next step, consisting of the combination of these new non-platinum agents together was the natural evolution of the treatment process. Gemcitabine and vinorelbine, because of the same administration schedule and compatible toxicity, were combined in a number of phase II studies. These trials reported response rates of 20 40%, myelosuppression as the major toxicity and median survival duration of 7 11 months [10 13]. The efficacy and toxicity of combined paclitaxel and gemcitabine have recently been evaluated in 54 chemotherapynaïve patients with metastatic NSCLC [15]. Gemcitabine 1000 mg/m 2 was administered on days 1 and 8 and paclitaxel 200 mg/m 2 as a continuous 3-h infusion on day 1. Treatment was repeated every 21 days. Two patients (4.3%) achieved a CR and 15 (31.9%) achieved a PR, giving an overall response rate of 36.2%. The median survival time was 51 weeks, with a 1-year survival probability of 0.48. Grade 3/4 neutropenia and febrile neutropenia occurred in 15.2 and 2.2% of courses, respectively. Grade 3/4 thrombocytopenia was rare (1.8% of courses). Among non-hematological toxicities, grade 1 or 2 peripheral neurotoxicity developed in 25 patients (47.2%) and grade 1 or 2 arthalgia/myalgia was observed in 30 patients (56.6%). Similar results were confirmed by Isla et al. [21] with a fortnightly regimen. Moreover, the combination of paclitaxel/gemcitabine was directly compared with paclitaxel/ carboplatin, a platinum-containing regimen considered as the gold standard in the USA. In this randomized study of 329 patients, no significant efficacy difference between the two arms was demonstrated, despite a trend toward a slightly higher response rate for the gemcitabine arm when compared with the carboplatin arm (36.5 versus 28.7, respectively; P = 0.17). The median survival for the gemcitabine arm was 12.3 months compared with 10.7 months for the carboplatin arm. Grade 3 or 4 neutropenia was also comparable in both arms, with a 10.5% incidence for gemcitabine/paclitaxel and 9.6% incidence for carboplatin/paclitaxel [16]. Our data show that paclitaxel, gemcitabine and vinorelbine can be delivered together in chemotherapy-naïve NSCLC patients with manageable hematological and non-hematological toxicity. The overall response rate of 66% achieved in the phase II study in 41 assessable patients is superior to the results achievable with any cisplatin-containing doublet and substantially comparable with the response reported with a three-drug platinum-containing combination [22]. The survival data also seem encouraging. In fact, a 14.5 month median survival has been observed in our series. No response to second-line platinum-containing chemotherapy was achieved in the 14 patients (three SD and 11 progressive disease) not responding to TGV, whereas all the responses to second-line platinum were recorded in patients who had already responded to TGV. This would suggest that the addition of a platinum compound, in front-line treatment, would not have been useful in this subgroup of patients. It is also worth noting that with TGV there was evidence of clear improvement of cancer-related symptoms and performance status in more than 70% of our patients. In fact, the only remarkable side-effect observed was asthenia in 29% of cases, but only in three patients (7%) as grade 3. However, a similar toxicity profile

1867 had already been observed with the combination of paclitaxel and gemcitabine [15, 16]. Thus the addition of vinorelbine to these two agents did not seem to worsen neurotoxicity. In conclusion, in view of the favorable safety profile of paclitaxel, gemcitabine and vinorelbine combination, coupled with encouraging response and survival rates, further comparative randomized trials are warranted in order to define the possible advantage of treating patients in the first-line with a non-platinum triplet, keeping platinum, as salvage therapy, in reserve. Acknowledgements The authors are grateful to Dr Silvana Valerio for her assistance in the preparation of this manuscript. References 1. Haskell CM. Chemotherapy and survival in patients with non-small cell lung cancer. A contrary view. Chest 1991; 99: 1325 1326. 2. Grilli R, Oxman AD, Julian JA. Chemotherapy for advanced nonsmall cell lung cancer: how much benefit is enough? J Clin Oncol 1993; 11: 1866 1872. 3. Soquet PJ, Chauvin F, Boissel JP et al. Polichemotherapy in advanced non-small cell lung cancer: a meta-analysis. Lancet 1993; 342: 19 21. 4. Non-Small-Cell Lung Cancer Collaborative Group. Chemotherapy in non-small-cell lung cancer: meta-analysis using updated data on individual patients from 52 randomised clinical trials. Br Med J 1995; 311: 899 909. 5. Cullen MH, Billingham LJ, Woodroffe CM et al. Mytomicin, ifosfamide, and cisplatin, in unresectable non-small cell lung cancer: effects on survival and quality. J Clin Oncol 1999; 17: 3188 3194. 6. Ellis PA, Smith IE, Hardy JR et al. Symptom relief with MVP (mitomycin C, vinblastine and cisplatin) chemotherapy in advanced non-small cell lung cancer. Br J Cancer 1995; 71: 366 370. 7. Cellerino R, Tummarello D, Piga A. Chemotherapy or not in advanced non-small cell lung cancer? Lung Cancer 1990; 6: 99 109. 8. Bakker W, van Oosterom AT, Aaronson NK et al. Vindesine, cisplatin and bleomycin combination chemotherapy in non-small cell lung cancer: survival and quality of life. Eur J Cancer Clin Oncol 1986; 22: 963 970. 9. Giaccone G. New drugs for the management of lung cancer. Br J Hosp Med 1996; 55: 634 638. 10. Lorusso V, Carpagnano F, Frasci G et al. Phase I/II study of gemcitabine plus vinorelbine as first-line chemotherapy of non-small-cell lung cancer. J Clin Oncol 2000; 18: 405 411. 11. Chen YM, Perng RP, Yang KY et al. A multicenter phase II trial of vinorelbine plus gemcitabine in previously untreated inoperable (stage IIIB/IV) non-small cell lung cancer. Chest 2000; 117: 1583 1589. 12. Krajnik G, Mohn-Staudner A, Thaler J et al. Vinorelbine gemcitabine in advanced non-small cell lung cancer (NSCLC): an AASLC phase II trial. Ann Oncol 2000; 11: 993 998. 13. Laack E, Mende T, Benk J et al. Gemcitabine and vinorelbine as first-line chemotherapy for advanced non-small cell lung cancer, a phase II trial. Eur J Cancer 2001; 37: 583 590. 14. Iaffaioli RV, Facchini G, Tortoriello A et al. Phase I study of vinorelbine and paclitaxel in small cell lung cancer. Cancer Chemother Pharmacol 1997; 41: 86 90. 15. Douillard JY, Lerouge D, Monnier A et al. Combined paclitaxel and gemcitabine as first-line treatment in metastatic non-small cell lung cancer: a multicentre phase II study. Br J Cancer 2001; 84: 1179 1184. 16. Kosmidis P. Interim results of a phase III trial. Paclitaxel/carboplatin vs paclitaxel/gemcitabine in advanced non small cell lung cancer. Oncology 2000; 7 (Suppl 4): 41 48. 17. Masters G, Mauer A, Hoffman PH et al. A phase I/II study of paclitaxel ifosfamide and vinorelbine, with G-CSF support in advanced non-small-cell lung cancer. Ann Oncol 1998; 9: 677 680. 18. WHO Handbook for Reporting Results of Cancer Treatment. Geneva, Switzerland: World Health Organization 1979. 19. Simon R. Confidence interval for reporting results of clinical trials. Ann Intern Med 1986; 105: 429 435. 20. The American Society of Clinical Oncology. Clinical practice guidelines for the treatment of unresectable non-small cell lung cancer. Adopted on May 16, 1997 by the American Society of Clinical Oncology. J Clin Oncol 1997; 15: 2996 3018. 21. Isla D, Rosell R, Sanchez JJ et al. Phase II trial of paclitaxel plus gemcitabine in patients with locally advanced or metastatic nonsmall-cell lung cancer. J Clin Oncol 2001; 19: 1071 1077. 22. Frasci G, Panza N, Comella P et al. Cisplatin, gemcitabine and paclitaxel in locally advanced or metastatic non-small cell lung cancer: a phase I II study. J Clin Oncol 1999; 17: 2316 2325.