- Email: [email protected]
Paclitaxel and gemcitabine versus carboplatin and gemcitabine in patients with advanced non-small-cell lung cancer. A phase III study of the Hellenic Cooperative Oncology Group
original article
Annals of Oncology 19: 115–122, 2008 doi:10.1093/annonc/mdm430 Published online 15 October 2007
Paclitaxel and gemcitabine versus carboplatin and gemcitabine in patients with advanced non-small-cell lung cancer. A phase III study of the Hellenic Cooperative Oncology Group
Department of Medical Oncology, Hygeia Hospital, Athens; 2Division of Oncology, University Hospital of Patras, Patras; 32nd Department of Medical Oncology, Henry Dunant Hospital, Athens; 4Oncology Unit, 3rd Department of Medicine, Athens Medical School, Sotiria General Hospital, University of Athens, Athens; 5Department of Medical Oncology, University of Ioannina School of Medicine, Ioannina; 6Department of Medical Oncology, Metropolitan Hospital, Piraeus; 7Department of Clinical Therapeutics, Alexandra Hospital, University of Athens School of Medicine, Athens; 83rd Department of Medical Oncology, Agii Anargiri Cancer Hospital, University of Ioannina, Athens; 92nd Department of Internal Medicine-Propaedeutic, Oncology Section, University General Hospital Attikon, Athens; 10Department of Medical Oncology, Theagenio Hospital, University of Athens Thessaloniki; 11Department of Medical Oncology, Papageorgiou Hospital, Aristotle University of Thessaloniki School of Medicine, Thessaloniki, Greece
Received 17 May 2007; revised 24 July 2007; accepted 24 July 2007
Background: This phase III study was designed to compare the combination paclitaxel (Taxol)–gemcitabine (PG) versus carboplatin–gemcitabine (CG) in patients with advanced inoperable non-small-cell lung cancer.
Methods: Chemotherapy-naive patients with performance status of zero or one were randomized to gemcitabine 1 gm/m2 on days 1 and 8 plus either paclitaxel 200 mg/m2 on day 1 (arm A) or carboplatin at an area under the concentration–time curve of 6 mg on day 1 (arm B) every 3 weeks. Primary end point was overall survival (OS). Secondary end points included objective response (OR), time to progression and toxicity. Results: A total of 512 patients were enrolled and 452 eligible (arm A, 225; arm B, 227) were analyzed. All characteristics were well balanced with the exception of vena cava obstruction symptoms and lymph node involvement. Median survival was 9.97 months [95% confidence interval (CI) 8.74–12.0] for group A and 10.49 (95% CI 9.04–11.94) for group B. There was no difference in the OS, 1-year survival, OR and TtP. However, statistically significant differences were seen in toxicity. Conclusion: The two regimens are equally active. Myelotoxicity is worse in the CG group whereas alopecia, myalgia and neurotoxicity worse in the PG group. Key words: carboplatin, chemotherapy, gemcitabine, non-small-cell lung cancer, paclitaxel
introduction Non-small-cell lung cancer (NSCLC) remains the leading cause of cancer-related mortality in Western countries and the incidence is increasing among women [1]. Platinum-based chemotherapy offers a survival benefit in advanced NSCLC as compared with best supportive care and represents the standard of care [2, 3]. Nonplatinum combinations have been also tested and are considered alternative regimens for those who cannot tolerate platinum-based chemotherapy [4, 5]. During the past few years, several novel drugs including paclitaxel (Taxol), docetaxel, gemcitabine and vinorelbine have shown significant activity against NSCLC. Doublets of the *Correspondence to: Dr P. A. Kosmidis, Hygeia Hospital, 2 An Tsoha & Vas Sofias Avenue, Athens 11521, Greece. Tel: +30 21-645-8602; Fax: +30 21-645-2616; E-mail: [email protected]
above so-called third-generation compounds or combinations of a platinum compound with a novel drug have been tested in advanced NSCLC, without significant difference in activity among them [6]. Carboplatin is less nephrotoxic and less emetogenic than cisplatin, it lacks neurotoxicity, it is convenient to administer as it does not require hydration and has an overall efficacy that is comparable to cisplatin in NSCLC [7, 8]. Our group has published two phase III studies comparing various doublets with platinum and nonplatinumbased combinations in advanced NSCLC [4]. In a phase III study, a combination of carboplatin at an area under the time–concentration curve (AUC) 6 mg and paclitaxel 200 mg/m2, both given on day 1 every 3 weeks, was found equally active to a nonplatinum combination of paclitaxel 200 mg/m2 on day 1 and gemcitabine 1000 mg/m2 given on days 1 and 8 every 3 weeks [7]. The toxicity profile was similar for both
ª 2007 European Society for Medical Oncology. For Permissions, please email: [email protected]
original article
1
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
P. A. Kosmidis1*, H. P. Kalofonos2, C. Christodoulou3, K. Syrigos4, T. Makatsoris2, D. Skarlos3, C. Bakogiannis1, C. Nicolaides5, D. Bafaloukos6, A. Bamias7, E. Samantas8, N. Xiros9, I. Boukovinas10, G. Fountzilas11 & M. A. Dimopoulos7
original article combinations [4]. Novel-targeted agents especially bevacizumab combined with chemotherapy have proven efficacy and promising results [9]. However, toxicity from chemotherapy remains an unresolved problem. Combinations with better toxicity profile but with equal or improved efficacy are eagerly needed. Therefore, our group decided to compare an easily given combination of carboplatin and gemcitabine with the combination of paclitaxel and gemcitabine in a phase III randomized study in NSCLC patients stage IIIbwet (IIIb with pleural effusion) and IV with good performance status (PS) (0–1).
patients and methods eligibility criteria
treatment plans and dose modifications Eligible patients were centrally randomized at the HeCOG data office to receive either paclitaxel (Taxol; Bristol-Myers Squibb Co, USA) 200 mg/m2 as a 3-h infusion on day 1 in combination with a 30-min infusion of gemcitabine 1 g/m2 on days 1 and 8 (group A) or carboplatin at an AUC of 6 mg according to the Calvert formula, as a 1-h i.v. infusion, in combination with a 30-min infusion of gemcitabine 1 g/m2 on days 1 and 8 (group B). Paclitaxel was always given before gemcitabine with premedication of 20 mg dexamethasone 12 h and 1 h before administration, diphenydramine 50 mg and cimetidine 50 mg i.v., 30 min before paclitaxel were administered. Ondasentron, 16 mg i.v., was given 15 min before chemotherapy, followed by 8 mg bid orally for the next 3 days. Treatment was repeated every 3 weeks until maximum response plus two cycles or unacceptable toxicity. In the situation of stable disease, patients received a maximum of six cycles. Dose modifications were planned according to hematologic and severe non-hematologic toxic effects. Once the doses reduced, they were not reescalated. Patients were treated at four dose levels (0, 21, 22 and 23) as follows. All patients were started at dose level 0, in which they received gemcitabine 1000 mg/m2 combined with paclitaxel 200 mg/m2 (group A) or carboplatin AUC 6 mg (group B). If WHO grade 3 neutropenia and/or thrombocytopenia, grade 2 neurotoxicity and hepatotoxicity, or grade 3 diarrhea occurred, doses were reduced to level 21 (gemcitabine 900 mg/m2 and paclitaxel 175 mg/m2 or carboplatin AUC 5 mg). If grade 4 neutropenia and/or thrombocytopenia or grade 3 hepatotoxicity occurred, the drugs were reduced to level 22 (gemcitabine 800 mg/m2 and paclitaxel 150 mg/m2 or carboplatin AUC 4 mg). If neutropenic fever was observed, doses were further decreased to level 23 (gemcitabine 700 mg/m2 and paclitaxel 140 mg/m2 or carboplatin AUC 4 mg). For patients who did not achieve hematologic recovery on the scheduled day of treatment, a full
116 | Kosmidis et al.
blood count was taken twice weekly until the absolute neutrophil count reached 1500/ll and the platelet count reached 100 000/l. If hematologic recovery was not achieved by day 35 of treatment, the patient was withdrawn from the study. Patients were also withdrawn from the study and discontinued treatment if they had symptomatic arrhythmia or atrioventricular block, grade 3 neurotoxicity or grade 4 hepatoxicity that was not reversible within 15 days.
baseline and follow-up assessments Pretreatment evaluation included a complete medical history, physical examination, complete blood count and standard biochemical profile. Bidimensionally measurable disease was assessed with chest X-rays, computed tomographic scans of chest and abdomen. Bone scanning was carried out as baseline test and then at the discretion of the treating physician. Tumor assessments for response were carried out at least every 6 weeks. Patients receiving at least two cycles of chemotherapy were considered assessable for response. Patients receiving at least one cycle of chemotherapy were assessable for toxicity. Standard WHO criteria were applied for assessment of response. Toxicity was evaluated according to the WHO grading system.
statistical analysis The primary end point of this study is to evaluate the survival between the two arms and secondary are time to disease progression, response and toxicity. For a two-sided test at the 5% level of significance and power of 80%, the number of patients required to detect a 610% difference in survival rate to a baseline rate of 45% at the 1-year time point is 480 patients. The study accrual rate is estimated at 150 patients per year. Taking into consideration a 5% withdrawal rate, 504 patients need to enter the study. The maximum study duration is estimated to be 4.2 years, while assuming that either the null or the alternative hypothesis holds leads to an expected duration of 3.5 and 3.8 years, respectively. An interim analysis based on the O’Brien Fleming (Lan-DeMets alpha spending function) boundary values will be carried out when 50% of the end points (191 deaths) have been reached. The study will end prematurely if either a significant difference is detected or the alternative is rejected at the interim analysis. Survival was estimated from randomization date to the date of last follow-up or until the patient’s death. Time to progression (TtP) was deemed as the time between randomization and disease progression. Patients who died from the disease without having documented progression, or from any cause during the chemotherapy period or after its completion, were considered as events for the estimation of TtP. Time to treatment failure (TTF) was calculated from randomization to the date of treatment discontinuation for any reason, or disease progression or death from any cause during chemotherapy (or after its completion), whichever occurred first. Time to event distributions were estimated using Kaplan–Meier curves and compared using the log-rank test. The Cox proportional hazards models were used to assess the strength of association of overall survival and TtP with various prognostic factors. A backward selection procedure with removal criterion P > 0.10 identified the subclass of significant variables among the following: treatment group (group A versus group B), age categories (£70 versus >70), gender (male versus female), prior Radiation Therapy (no versus yes), prior lung surgery (no versus yes), stage of disease (IIIbwet versus IV), PS (0 versus 1), tumor histology (squamous cell versus adenocarcinoma versus other histology), symptoms (no versus yes), weight loss (>10%) (no versus yes), fever (no versus yes), vena cava obstruction symptoms (no versus yes), nodal involvement (no versus yes), locoregional metastasis (no versus yes), pleura involvement (no versus yes), brain metastasis (no versus yes), bone marrow metastasis (no versus yes), adrenal glands metastasis (no versus yes),
Volume 19 | No. 1 | January 2008
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
Chemotherapy-naive patients, older than 18 years of age, with histologically or cytologically confirmed NSCLC stage IIIbwet or IV, according to the American Joint Committee on Cancer Criteria, were included. An Eastern Cooperative Oncology Group PS of zero or one was required. Additional inclusion criteria were measurable or evaluable disease in nonirradiated fields, a life expectancy of at least 3 months and adequate bone marrow, kidney and liver functions. Patients with stable brain metastases were eligible. Exclusion criteria were a past or current history of other malignant disease, except for basal cell carcinoma of the skin or carcinoma in situ of the cervix, active cardiac disease or active infection and preexisting motor or sensory neuropathy grade >2 according to World Health Organization (WHO). All patients were required to provide written informed consent. The study was approved by the Hellenic Cooperative Oncology Group (HeCOG) Protocol Review Committee and by Institutional Review Boards in participating institutions.
Annals of Oncology
original article
Annals of Oncology
Selective patient and tumor characteristics are presented in Tables 1 and 2. There were no significant differences in major characteristics between the two treatment groups except for vena cava obstruction symptoms and lymph node involvement (P = 0.04 in both cases).
results
treatment characteristics A total of 969 and 995 cycles of chemotherapy were given in groups A and B, respectively. The median number of cycles administered was five in group A (range 1–8) and six (range 1– 10) in group B. Significantly more cycles in group A were administered at full dose (group A: 72% versus group B: 63.5%, P < 0.001) (Table 3). The relative dose intensity of paclitaxel was 0.85 and of gemcitabine 0.82 in group A and 0.77 in group B. The median cumulative dose for carboplatin was 2600 mg. Totally, 108 patients in group A and 116 in group B completed at least six cycles of chemotherapy.
patient and tumor characteristics From October 2000 to March 2004, 512 patients were randomized in this study. Sixty (12%) patients were ineligible. Thirty-two had stage other than IIIb wet or IV and 28 were ineligible for other reasons (mostly because of PS >1, history of prior cancer and no stage data stated at randomization or in their medical records). Furthermore, for four patients (one in group A and three in group B), medical records were incomplete while 10 patients (five in each group) never started protocol treatment. Two patients in each group received the opposite treatment than the one allocated to. The remaining 452 eligible patients were analyzed according to the intention-to-treat principle (N = 452; group A: 225, group B: 227), except treatment and toxicity analyses, which were conducted on the actual treatment arm (N = 438; group A: 219, group B: 219) (Figure 1).
response A total of 80 patients, 42 in group A and 38 in group B, were considered nonassessable for response. Response rate was calculated in the intention-to-treat population (Table 4).
Figure 1. Progress through the various stages of the trial. Survival status update on September 2006.
Volume 19 | No. 1 | January 2008
doi:10.1093/annonc/mdm430 | 117
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
bone metastasis (no versus yes), liver metastasis (no versus yes), skin/soft tissue metastasis (no versus yes) other metastasis (no versus yes) and number of metastatic sites (<3 versus ‡3). All the analysis (except toxicity and treatment characteristics) was conducted according to the ‘intention-to-treat’ principle.
original article
Annals of Oncology
Table 1. Selected patient characteristics
n Age (years) Median Range
Group A (paclitaxel + gemcitabine)
Group B (carboplatin + gemcitabine)
225
227
63 42–82 n
n
%
63 36–83 n
%
194 31
86 14
184 43
81 19
104 121
46 54
106 121
47 53
30 191 4
13 85 2
32 191 4
14 84 2
1 220 4
0.4 98 1.8
8 215 4
3.5 95 1.8
47 178 –
21 79 –
42 184 1
18.5 81 0.4
44 181
20 80
35 192
15 85
*P = 0.04. ECOG, Eastern Cooperative Oncology Group.
Three complete responses (CRs; dissappearance of all evidence of disease for a duration of at least 4 weeks) and 67 partial responses (PRs; a 50% or greater reduction in the product obtained from measurement of each lesion for at least 4 weeks and no appearance of new lesions) were achieved in group A for an overall response rate (ORR) of 31% [95% confidence interval (CI), 25.12–37.60]. One CR and 61 PRs were achieved in group B for an ORR of 27% (95% CI, 21.63–33.60). The ORR, as given by investigators, did not differ significantly between the two treatment groups (P = 0.41).
survival, TtP and TTF At the time of the analysis, after a median follow-up of 44 months (range 0.01–70.56+), 402 patients (89%), 202 in group A and 200 in group B, had died and 418 (92.5%), 209 in group in group A and 209 in group B, had relapsed. Survival, TtP and TTF did not differ significantly between treatment groups. Overall, median survival was 9.97 months (range 0.01–70.56 months, 95% CI 8.74–12.0) for group A and 10.49 months (range 0.01–55.74+, 95% CI 9.04–11.94) for group B (P = 0.99) (Figure 2). The 1-year survival rate was 42% in both groups. TtP was 5.02 months (range 0.01–70.56 months, 95% CI 4.31–5.72) for
118 | Kosmidis et al.
Stage IIIbwet IV Tumor histology Squamous cell Adenocarcinoma Large cell Mixed Undifferentiated Unclassified Unknown Number of metastatic sites <3 ‡3 Unknown Lymph nodes involvement* Yes No Unknown Liver metastasis Yes No Unknown Bone metastasis Yes No Unknown Brain metastasis Yes No Unknown Adrenal glands metastasis Yes No Unknown
Group A (paclitaxel + gemcitabine)
Group B (carboplatin + gemcitabine)
225 n
%
227 n
%
30 195
13 87
31 196
14 86
58 119 10 5 14 14 5
26 53 4 2 6 6 2
67 115 9 4 8 19 5
29.5 51 4 1.8 3.5 8 2
155 66 4
69 29 2
153 71 3
67 31 1
134 87 4
60 39 1.8
157 67 3
69 29.5 1.3
35 186 4
16 83 1.8
33 191 3
14.5 84 1.3
83 138 4
37 61 1.8
98 126 3
43 55.5 1.3
34 187 4
15 83 1.8
31 193 3
14 85 1.3
36 185 4
16 82 1.8
33 191 3
14.5 84 1.3
*P = 0.04.
group A and 5.11 months (range 0.01–55.74, 95% CI 4.43– 5.80) for group B (P = 0.35) (Figure3). The respective 1-year progression-free survival rates were 14% and 12%, respectively. TTF was 3.97 months (range 0.01–70.56 months, 95% CI 3.22–4.71) for group A and 4.13 months (range 0.01–55.74, 95% CI 3.15–5.11) for group B (P = 0.89). (Table 5). The Cox multivariate regression analysis revealed that in the presence of treatment group (P = 0.60) and stage of disease (P = 0.26), several prognostic factors including PS [1 versus 0: hazard ratio (HR) = 1.30, 95% CI 1.06–1.60, P = 0.01], histology type (adenocarcinoma versus squamous: HR = 1.37, 95% CI 1.07–1.74, P = 0.01; other versus squamous:
Volume 19 | No. 1 | January 2008
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
Gender Male Female ECOG performance status 0 1 Loss of weight (>10%) Yes No Unknown Vena cava obstruction symptoms* Yes No Unknown Prior radiotherapy Yes No Unknown Prior lung surgery Yes No
Table 2. Selected tumor characteristics
original article
Annals of Oncology
Table 3. Treatment characteristics Group B (carboplatin + gemcitabine)
219 969 5 1–8 72%
219 995 6 1–10 63.5%
57 20.25–68.3
– –
0.85 0.30–1.02
– –
546 182–672
512.5 256–636
0.82 0.27–1.01
0.77 0.38–0.95
– –
2600 430–6560
–
1.0 0.59–1.67
Figure 2. Survival of patients treated with paclitaxel + gemcitabine (red line) or carboplatin + gemcitabine (blue line) (P = 0.99).
a
90% of the dose defined in the protocol (for both drugs). DI, dose intensity; RDI, relative dose intensity.
Table 4. Best response to treatment
n CR PR ORR* SD PD NA Unknown
Group A (paclitaxel + gemcitabine)
Group B (carboplatin + gemcitabine)
225 n (%) 3 (1) 67 (30) 70 (31) 62 (28) 47 (21) 42 (19%) 4 (2)
227 n (%) 1 (0.4) 61 (27) 62 (27) 66 (29) 58 (26) 38 (17) 3 (1)
95% CI 0.27–3.85 23.9–36.21 25.12–37.60 21.82–33.88 15.77–26.79 13.79–24.38 –
95% CI 0.01–2.43 21.22–33.14 21.63–33.60 23.26–35.45 20.0–31.74 12.13–22.24 –
*P = 0.41. CR, complete response; PR, partial response; ORR, overall response rate; SD, stable disease; PD, progression disease; NA, nonassessable.
HR = 1.67, 95% CI 1.23–2.26, P = 0.001), presence of symptoms (yes versus no: HR = 1.47, 95% CI 1.06–2.05, P = 0.02) and number of metastatic sites (‡3 sites versus <3: HR = 1.67, 95% CI 1.32–2.12, P < 0.001) were related to significantly poorer survival. On the other hand, the risk of death was significantly lower for women (female versus male: HR = 0.71, 95% CI 0.53–0.94, P = 0.02), for patients having prior lung surgery
Volume 19 | No. 1 | January 2008
Figure 3. Time to progression of patients treated with paclitaxel + gemcitabine (red line) or carboplatin + gemcitabine (blue line) (P = 0.35).
(yes versus no: HR = 0.73, 95% CI 0.55–0.98, P = 0.04) and patients with lung metastasis at entry (yes versus no: HR = 0.62, 95% CI 0.49–0.79, P < 0.001) (Table 6). The Cox multivariate regression analysis for TtP revealed that in the presence of treatment group (P = 0.49) and stage of disease (P = 0.12), the hazard of disease progression at any time was significantly higher for patients with worse PS (1 versus 0: HR = 1.36, 95% CI 1.12–1.67, P = 0.002) and larger number of metastatic sites (‡3 sites versus <3: HR = 1.87, 95% CI 1.48–2.36, P < 0.001). On the contrary, the risk of disease progression was significantly lower for women (female versus male: HR = 0.79, 95% CI 0.60–1.03, P = 0.08), for patients with lung metastasis at entry (yes versus no: HR = 0.62, 95% CI 0.49–0.78, P < 0.001), for patients with bone marrow metastasis (yes versus no: HR = 0.35, 95% CI 0.11–1.12, P = 0.08) and for patients having prior lung surgery (yes versus no: HR = 0.68, 95% CI 0.52–0.89, P = 0.005) (Table 6).
toxicity One toxic death occurred in group A from pancytopenia after the completion of the first cycle. Additionally, one patient died from pancytopenia after the completion of the fourth
doi:10.1093/annonc/mdm430 | 119
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
n Number of cycles delivered Median Range % of cycles at full dosea DI of paclitaxel delivered Median (mg/m2/week) Range RDI of paclitaxel Median Range DI of gemcitabine delivered Median (mg/m2/week) Range RDI of gemcitabine Median Range Cumulative dose of carboplatin Median Range RDI of carboplatin Median Range
Group A (paclitaxel + gemcitabine)
original article
Annals of Oncology
Table 5. Survival, time to disease progression and TTF Carboplatin + gemcitabine (n = 227)
202/225 0.01–70.56 9.97 8.74–12.0 42%
200/227 0.01–55.74+ 10.49 9.04–11.94 42%
209/225 0.01–70.56 5.02 4.31–5.72 14%
209/227 0.01–55.74 5.11 4.43–5.80 12%
217/225 0.01–70.56 3.97 3.22–4.71
220/227 0.01–55.74 4.13 3.15–5.11
*log-rank P = 0.99; **log-rank P = 0.35; ***log-rank P = 0.89. TTF, time to treatment failure; CI, confidence interval.
cycle and another died from sepsis febrile neutropenia after the completion of the third cycle in group B. Eight patients in group A and 13 in group B developed febrile neutropenia (group A: 4% versus group B: 6%, P = 0.37). Data for hospitalizations, antibiotics, transfusions and supportive care were not available for all patients. Among patients with available data, no significant differences were found between the two groups in terms of hospitalization needs (group A: 42 of 198 = 21% versus group B: 53 of 201 = 26%, P = 0.24). Regarding supportive care, significantly more patients in group B received granulocyte colony-stimulating factor (G-CSF) (group A: 79 of 193 = 41% versus group B: 105 of 187 = 56%, P = 0.004) and Erythropoetin (group A: 46 of 185 = 25% versus group B: 108 of 189 = 57%, P < 0.001). Additionally, red blood cell (RBC) transfusions (group A: 12 of 194 = 6% versus group B: 52 of 195 = 27%, P < 0.001) and platelet transfusions (group A: 2 of 193 = 1% versus group B: 15 of 191 = 8%, P = 0.001) were significantly more frequent in group B. Regarding antibiotics, no significant differences were found between the two groups (group A: 41 of 193 = 21% versus group B: 37 of 193 = 19%, P = 0.70). The rates of severe (grade 3/4) hematologic toxic effects (neutropenia, leucopenia, thrombocytopenia, anemia) were significantly worse in group B, as shown in Table 7. Among grade 3/4 non-hematologic toxic effects, alopecia and neurotoxicity were significantly worse in group A (62.7% versus 6.2%, P < 0.001 and 5.2% versus 0.5%, P = 0.006, respectively). Other severe toxic effects did not differ significantly between the two groups (Table 8).
discussion In this phase III randomized study, we targeted the population of patients with NSCLC, stage IIIbwet and IV and specifically
120 | Kosmidis et al.
Survival Treatment group Paclitaxel + gemcitabine Carboplatin + gemcitabine Stage IIIbwet IV Gender Male Female Prior lung surgery No Yes Symptoms No Yes PS 0 1 Histology type Squamous Adenocarcinoma Other Lung metastasis at entry No Yes Number of metastatic sites <3 ‡3 Time to disease progression Treatment group Paclitaxel + gemcitabine Carboplatin + gemcitabine Stage IIIbwet IV Gender Male Female Prior lung surgery No Yes PS 0 1 Bone marrow metastasis at entry No Yes Lung metastasis at entry No Yes Number of metastatic sites <3 ‡3
HR
95% CI
P value
1 1.06
0.86–1.30
0.60
1 1.20
0.87–1.66
0.26
1 0.71
0.53–0.94
0.02
1 0.73
0.55–0.98
0.04
1 1.47
1.06–2.05
0.02
1 1.30
1.06–1.60
0.01
1 1.37 1.67
1.07–1.74 1.23–2.26
0.01 0.001
1 0.62
0.49–0.79
<0.001
1 1.67
1.32–2.12
<0.001
1 1.07
0.88–1.30
0.49
1 1.27
0.93–1.73
0.12
1 0.79
0.60–1.03
0.08
1 0.68
0.52–0.89
0.005
1 1.36
1.12–1.67
0.002
1 0.35
0.11–1.12
0.08
1 0.62
0.49–0.78
<0.001
1 1.87
1.48–2.36
<0.001
HR, hazard ratio; CI, confidence interval; PS, performance status.
Volume 19 | No. 1 | January 2008
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
Survival Events Range (months) Median* 95% CI 1-year survival rate Time to disease progression Events Range (months) Median** 95% CI 1-year disease-free rate TTF Events Range (months) Median*** 95% CI
Paclitaxel + gemcitabine (n = 225)
Table 6. Estimated HRs and 95% CIs for survival and time to disease progression
original article
Annals of Oncology
Table 7. Hematologic toxicity
Neutropenia* Leucopenia** Thrombocytopenia*** Anemia****
Paclitaxel + gemcitabine (n = 219) Grade 3 Grade 4 No. % No. %
Carboplatin + gemcitabine (n = 219) Grade 3 Grade 4 No. % No. %
25/212 11.8 13/212 13/213 6.1 4/213 3/212 1.4 2/212 9/212 4.2 2/212
32/211 34/212 22/212 23/212
6.1 1.9 0.9 0.9
15.2 16 10.4 10.8
30/211 14.2 7/212 3.3 32/212 15.1 5/212 2.4
*P = 0.009; **P = 0.001; ***P < 0.001; ****P = 0.007. Table 8. Non-hematologic toxicity Carboplatin + gemcitabine (n = 219) Grade 3 Grade 4 No. % No. %
1/211 0.5 0/211 0 5/206 2.4 0/206 0 3/208 1.4 0/208 0 1/207 0.5 2/207 1.0 133/212 62.7 0/212 0 11/211 5.2 0/211 0 2/219 0.9 0/219 0 1/212 0.5 1/212 0.5 4/219 1.8 0/219 0 0/219 0 1/219 0.5
2/210 2/208 1/209 3/208 13/210 1/210 3/219 1/210 3/219 1/219
1.0 1.0 0.5 1.4 6.2 0.5 1.4 0.5 1.4 0.5
0/210 0/208 0/209 0/208 0/210 0/210 0/219 0/210 0/219 0/219
0 0 0 0 0 0 0 0 0 0
HSR, rash, edema, pruritus and allergic reaction; mucositis, stomatitis, diarrhea. *P < 0.001; **P = 0.006.
those with a PS zero to one. We tested a nonplatinum combination of third-generation compounds, namely paclitaxel and gemcitabine, versus a combination of carboplatin and gemcitabine. The overall median survival was similar in the two groups (9.97 versus 10.49 months). Moreover, the 1-year survival rate was 42% for both groups. Median TtP was also similar in the two groups (5.02 versus 5.11 months). Finally, ORRs were 31% for group A and 27% for group B based on an intention-to-treat analysis. Based on the results, we conclude that this study has failed to meet each primary end point. This is in accordance with the results of our previous randomized trial, which compared the nonplatinum combination of paclitaxel and gemcitabine with a combination of carboplatin and paclitaxel [4]. In that study, we had included patients with NSCLC stage IIIa and IIIbnon-wet and also patients with PS two. In a subgroup analysis, we had found that patients with PS two have a significantly worse outcome in both groups [4]. This was also consistent with the results from other randomized studies [10, 11]. The above results indicated that the subgroup of NSCLC with poor PS required a different treatment approach. Therefore, in the current study we did not include patients with PS two. On the other hand, it is known that patients with NSCLC stage IIIa
Volume 19 | No. 1 | January 2008
references 1. Ginsberg RJ, Vokes EE, Rosenzweig K. Non small cell lung cancer. In De Vita Jr, Hellman S, Rosenber SR (eds): Cancer.: Principles and Practice of Oncology, 6th edition. Philadelphia, PA: Lippinicott-Raven 2001; 925–983. 2. 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 randomized clinical trials. BMJ 1995; 311: 899–909. 3. 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. 4. Kosmidis P, Mylonakis N, Nicolaides C et al. Paclitaxel plus carboplatin versus gemcitabine plus paclitaxel in advanced non-small-cell lung cancer: a phase III randomized trial. J Clin Oncol 2002; 20: 3578–3585. 5. Georgoulias V, Papadakis E, Alexopoulos A et al. Platinum-based and nonplatinum-based chemotherapy in advanced non-small cell lung cancer: a randomized multicentre trial. Lancet 2001; 357: 1478–1484.
doi:10.1093/annonc/mdm430 | 121
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
Nausea/vomiting HSR Mucositis Infection Alopecia* Neurotoxicity** Hepatotoxicity Cardiotoxicity Fatigue Nephrotoxicity
Paclitaxel + gemcitabine (n = 219) Grade 3 Grade 4 No. % No. %
and IIIbnonwet require a concurrent chemoradiotherapy approach [12, 13] and as a result we did not include such patients in our current study, unlike most of the studies in advanced NSCLC in the literature published so far. Despite these essential differences in patient selection in the current study, we did not find significant differences in activity between the nonplatinum combination and the combination of carboplatin and gemcitabine. A recent meta-analysis comparing platinum-based versus nonplatinumbased chemotherapy in advanced NSCLC confirmed that there is a marginal 1-year survival benefit with platinumbased regiments which is lost when platinum therapies are compared with third-generation-based combination regimens [14]. Toxicity profile was different between the two groups of our study. Myelotoxicity was significantly more severe in group B. Grade 3/4 anemia, leucopenia, neutropenia and thrombocytopenia were significantly worse in group B. As a result, patients in group B received significantly more RBCs and platelet transfusions and received more often G-CSF and erythropoietin. Febrile neutropenia, need for antibiotics and infections did not differ significantly between the two groups. On the other hand, grade 3 alopecia and grade 3 neurotoxicity were significantly more frequent in group A as expected, due to the use of paclitaxel. In our previous study comparing paclitaxel and gemcitabine with carboplatin and paclitaxel, toxicity was mild and no significant differences were found in toxicity profiles between the two groups [4]. In the recent meta-analysis, mentioned above, comparing platinum-based versus nonplatinum-based chemotherapy in advanced NSCLC, the toxicity of platinum-based regimens was significantly higher for hematologic toxicity [14]. This is consistent with the findings of our study. Most of the targeted novel agents have been combined with carboplatin–paclitaxel or cisplatin–gemicitabine. Both of these combinations are responsible for significant toxic effects such as alopecia and nephrotoxicity, respectively. Based on our data, paclitaxel and gemcitabine is equally active to carboplatin and gemcitabine in NSCLC patients of good PS and stage IIIbwet and IV. These regimes have different toxicity profiles and could be delivered according to the patient’s needs.
original article 6. Schiller JH, Harrington D, Belani C et al. Comparison of four chemotherapy regimens for advanced non-small-cell lung cancer. N Engl J Med 2002; 346: 92–98. 7. Bonomi PD, Finkelstein DM, Ruckdeschel JC et al. Combination chemotherapy versus single agents followed by combination chemotherapy in stage IV nonsmall-cell lung cancer: a study of the Eastern Cooperative Oncology Group. J Clin Oncol 1989; 7: 1602–1613. 8. Klastersky J, Sculier P, La Croix 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. 9. Sandler A, Gray R, Perry M et al. Paclitaxel-carboplatin alone or with bevacizumab for non-small cell lung cancer. N Engl J Med 2005; 355: 2542–2550. 10. Sweeney CJ, Zhu J, Sandler AB et al. Outcome of patients with a performance status of 2 in Eastern Cooperative Oncology Group Study E 1594: a phase III trial in patients with metastatic non-small cell lung carcinoma. Cancer 2001; 92: 2639–2647.
Annals of Oncology
11. Georgoulias V, Ardavanis A, Tsiafaki X et al. Vinorelbine plus cisplatin versus docetaxel plus gemcitabine in advanced non-small cell lung cancer: a phase III randomized trial. J Clin Oncol 2005; 23: 2937–3945. 12. Furuse K, Fukuoka M, Kawahara M et al. Phase III study of concurrent versus sequential thoracic radiotherapy in combination with mitomycin, vindesine and cisplatin in unresectable stage III non-small-cell lung cancer. J Clin Oncol 1999; 17: 2692–2699. 13. Fournel P, Robinet G, Thomas P et al. Randomized phase III trial of sequential chemoradiotherapy compared with concurrent chemoradiotherapy in locally advanced non-small-cell lung cancer: Groupe Lyon-Saint-Etienne d’Oncologie Thoracique-Groupe Francais de Pneumo-Cancerologie NPC 95-01 Study. J Clin Oncol 2005; 23: 5910–5917. 14. D’ Addario G, Pintilie M, Leighl N et al. Platinum-based versus non-platinumbased chemotherapy in advanced non-small-cell lung cancer. A meta-analysis of the published literature. J Clin Oncol 2005; 23: 2926–2936.
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
122 | Kosmidis et al.
Volume 19 | No. 1 | January 2008
Annals of Oncology 19: 115–122, 2008 doi:10.1093/annonc/mdm430 Published online 15 October 2007
Paclitaxel and gemcitabine versus carboplatin and gemcitabine in patients with advanced non-small-cell lung cancer. A phase III study of the Hellenic Cooperative Oncology Group
Department of Medical Oncology, Hygeia Hospital, Athens; 2Division of Oncology, University Hospital of Patras, Patras; 32nd Department of Medical Oncology, Henry Dunant Hospital, Athens; 4Oncology Unit, 3rd Department of Medicine, Athens Medical School, Sotiria General Hospital, University of Athens, Athens; 5Department of Medical Oncology, University of Ioannina School of Medicine, Ioannina; 6Department of Medical Oncology, Metropolitan Hospital, Piraeus; 7Department of Clinical Therapeutics, Alexandra Hospital, University of Athens School of Medicine, Athens; 83rd Department of Medical Oncology, Agii Anargiri Cancer Hospital, University of Ioannina, Athens; 92nd Department of Internal Medicine-Propaedeutic, Oncology Section, University General Hospital Attikon, Athens; 10Department of Medical Oncology, Theagenio Hospital, University of Athens Thessaloniki; 11Department of Medical Oncology, Papageorgiou Hospital, Aristotle University of Thessaloniki School of Medicine, Thessaloniki, Greece
Received 17 May 2007; revised 24 July 2007; accepted 24 July 2007
Background: This phase III study was designed to compare the combination paclitaxel (Taxol)–gemcitabine (PG) versus carboplatin–gemcitabine (CG) in patients with advanced inoperable non-small-cell lung cancer.
Methods: Chemotherapy-naive patients with performance status of zero or one were randomized to gemcitabine 1 gm/m2 on days 1 and 8 plus either paclitaxel 200 mg/m2 on day 1 (arm A) or carboplatin at an area under the concentration–time curve of 6 mg on day 1 (arm B) every 3 weeks. Primary end point was overall survival (OS). Secondary end points included objective response (OR), time to progression and toxicity. Results: A total of 512 patients were enrolled and 452 eligible (arm A, 225; arm B, 227) were analyzed. All characteristics were well balanced with the exception of vena cava obstruction symptoms and lymph node involvement. Median survival was 9.97 months [95% confidence interval (CI) 8.74–12.0] for group A and 10.49 (95% CI 9.04–11.94) for group B. There was no difference in the OS, 1-year survival, OR and TtP. However, statistically significant differences were seen in toxicity. Conclusion: The two regimens are equally active. Myelotoxicity is worse in the CG group whereas alopecia, myalgia and neurotoxicity worse in the PG group. Key words: carboplatin, chemotherapy, gemcitabine, non-small-cell lung cancer, paclitaxel
introduction Non-small-cell lung cancer (NSCLC) remains the leading cause of cancer-related mortality in Western countries and the incidence is increasing among women [1]. Platinum-based chemotherapy offers a survival benefit in advanced NSCLC as compared with best supportive care and represents the standard of care [2, 3]. Nonplatinum combinations have been also tested and are considered alternative regimens for those who cannot tolerate platinum-based chemotherapy [4, 5]. During the past few years, several novel drugs including paclitaxel (Taxol), docetaxel, gemcitabine and vinorelbine have shown significant activity against NSCLC. Doublets of the *Correspondence to: Dr P. A. Kosmidis, Hygeia Hospital, 2 An Tsoha & Vas Sofias Avenue, Athens 11521, Greece. Tel: +30 21-645-8602; Fax: +30 21-645-2616; E-mail: [email protected]
above so-called third-generation compounds or combinations of a platinum compound with a novel drug have been tested in advanced NSCLC, without significant difference in activity among them [6]. Carboplatin is less nephrotoxic and less emetogenic than cisplatin, it lacks neurotoxicity, it is convenient to administer as it does not require hydration and has an overall efficacy that is comparable to cisplatin in NSCLC [7, 8]. Our group has published two phase III studies comparing various doublets with platinum and nonplatinumbased combinations in advanced NSCLC [4]. In a phase III study, a combination of carboplatin at an area under the time–concentration curve (AUC) 6 mg and paclitaxel 200 mg/m2, both given on day 1 every 3 weeks, was found equally active to a nonplatinum combination of paclitaxel 200 mg/m2 on day 1 and gemcitabine 1000 mg/m2 given on days 1 and 8 every 3 weeks [7]. The toxicity profile was similar for both
ª 2007 European Society for Medical Oncology. For Permissions, please email: [email protected]
original article
1
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
P. A. Kosmidis1*, H. P. Kalofonos2, C. Christodoulou3, K. Syrigos4, T. Makatsoris2, D. Skarlos3, C. Bakogiannis1, C. Nicolaides5, D. Bafaloukos6, A. Bamias7, E. Samantas8, N. Xiros9, I. Boukovinas10, G. Fountzilas11 & M. A. Dimopoulos7
original article combinations [4]. Novel-targeted agents especially bevacizumab combined with chemotherapy have proven efficacy and promising results [9]. However, toxicity from chemotherapy remains an unresolved problem. Combinations with better toxicity profile but with equal or improved efficacy are eagerly needed. Therefore, our group decided to compare an easily given combination of carboplatin and gemcitabine with the combination of paclitaxel and gemcitabine in a phase III randomized study in NSCLC patients stage IIIbwet (IIIb with pleural effusion) and IV with good performance status (PS) (0–1).
patients and methods eligibility criteria
treatment plans and dose modifications Eligible patients were centrally randomized at the HeCOG data office to receive either paclitaxel (Taxol; Bristol-Myers Squibb Co, USA) 200 mg/m2 as a 3-h infusion on day 1 in combination with a 30-min infusion of gemcitabine 1 g/m2 on days 1 and 8 (group A) or carboplatin at an AUC of 6 mg according to the Calvert formula, as a 1-h i.v. infusion, in combination with a 30-min infusion of gemcitabine 1 g/m2 on days 1 and 8 (group B). Paclitaxel was always given before gemcitabine with premedication of 20 mg dexamethasone 12 h and 1 h before administration, diphenydramine 50 mg and cimetidine 50 mg i.v., 30 min before paclitaxel were administered. Ondasentron, 16 mg i.v., was given 15 min before chemotherapy, followed by 8 mg bid orally for the next 3 days. Treatment was repeated every 3 weeks until maximum response plus two cycles or unacceptable toxicity. In the situation of stable disease, patients received a maximum of six cycles. Dose modifications were planned according to hematologic and severe non-hematologic toxic effects. Once the doses reduced, they were not reescalated. Patients were treated at four dose levels (0, 21, 22 and 23) as follows. All patients were started at dose level 0, in which they received gemcitabine 1000 mg/m2 combined with paclitaxel 200 mg/m2 (group A) or carboplatin AUC 6 mg (group B). If WHO grade 3 neutropenia and/or thrombocytopenia, grade 2 neurotoxicity and hepatotoxicity, or grade 3 diarrhea occurred, doses were reduced to level 21 (gemcitabine 900 mg/m2 and paclitaxel 175 mg/m2 or carboplatin AUC 5 mg). If grade 4 neutropenia and/or thrombocytopenia or grade 3 hepatotoxicity occurred, the drugs were reduced to level 22 (gemcitabine 800 mg/m2 and paclitaxel 150 mg/m2 or carboplatin AUC 4 mg). If neutropenic fever was observed, doses were further decreased to level 23 (gemcitabine 700 mg/m2 and paclitaxel 140 mg/m2 or carboplatin AUC 4 mg). For patients who did not achieve hematologic recovery on the scheduled day of treatment, a full
116 | Kosmidis et al.
blood count was taken twice weekly until the absolute neutrophil count reached 1500/ll and the platelet count reached 100 000/l. If hematologic recovery was not achieved by day 35 of treatment, the patient was withdrawn from the study. Patients were also withdrawn from the study and discontinued treatment if they had symptomatic arrhythmia or atrioventricular block, grade 3 neurotoxicity or grade 4 hepatoxicity that was not reversible within 15 days.
baseline and follow-up assessments Pretreatment evaluation included a complete medical history, physical examination, complete blood count and standard biochemical profile. Bidimensionally measurable disease was assessed with chest X-rays, computed tomographic scans of chest and abdomen. Bone scanning was carried out as baseline test and then at the discretion of the treating physician. Tumor assessments for response were carried out at least every 6 weeks. Patients receiving at least two cycles of chemotherapy were considered assessable for response. Patients receiving at least one cycle of chemotherapy were assessable for toxicity. Standard WHO criteria were applied for assessment of response. Toxicity was evaluated according to the WHO grading system.
statistical analysis The primary end point of this study is to evaluate the survival between the two arms and secondary are time to disease progression, response and toxicity. For a two-sided test at the 5% level of significance and power of 80%, the number of patients required to detect a 610% difference in survival rate to a baseline rate of 45% at the 1-year time point is 480 patients. The study accrual rate is estimated at 150 patients per year. Taking into consideration a 5% withdrawal rate, 504 patients need to enter the study. The maximum study duration is estimated to be 4.2 years, while assuming that either the null or the alternative hypothesis holds leads to an expected duration of 3.5 and 3.8 years, respectively. An interim analysis based on the O’Brien Fleming (Lan-DeMets alpha spending function) boundary values will be carried out when 50% of the end points (191 deaths) have been reached. The study will end prematurely if either a significant difference is detected or the alternative is rejected at the interim analysis. Survival was estimated from randomization date to the date of last follow-up or until the patient’s death. Time to progression (TtP) was deemed as the time between randomization and disease progression. Patients who died from the disease without having documented progression, or from any cause during the chemotherapy period or after its completion, were considered as events for the estimation of TtP. Time to treatment failure (TTF) was calculated from randomization to the date of treatment discontinuation for any reason, or disease progression or death from any cause during chemotherapy (or after its completion), whichever occurred first. Time to event distributions were estimated using Kaplan–Meier curves and compared using the log-rank test. The Cox proportional hazards models were used to assess the strength of association of overall survival and TtP with various prognostic factors. A backward selection procedure with removal criterion P > 0.10 identified the subclass of significant variables among the following: treatment group (group A versus group B), age categories (£70 versus >70), gender (male versus female), prior Radiation Therapy (no versus yes), prior lung surgery (no versus yes), stage of disease (IIIbwet versus IV), PS (0 versus 1), tumor histology (squamous cell versus adenocarcinoma versus other histology), symptoms (no versus yes), weight loss (>10%) (no versus yes), fever (no versus yes), vena cava obstruction symptoms (no versus yes), nodal involvement (no versus yes), locoregional metastasis (no versus yes), pleura involvement (no versus yes), brain metastasis (no versus yes), bone marrow metastasis (no versus yes), adrenal glands metastasis (no versus yes),
Volume 19 | No. 1 | January 2008
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
Chemotherapy-naive patients, older than 18 years of age, with histologically or cytologically confirmed NSCLC stage IIIbwet or IV, according to the American Joint Committee on Cancer Criteria, were included. An Eastern Cooperative Oncology Group PS of zero or one was required. Additional inclusion criteria were measurable or evaluable disease in nonirradiated fields, a life expectancy of at least 3 months and adequate bone marrow, kidney and liver functions. Patients with stable brain metastases were eligible. Exclusion criteria were a past or current history of other malignant disease, except for basal cell carcinoma of the skin or carcinoma in situ of the cervix, active cardiac disease or active infection and preexisting motor or sensory neuropathy grade >2 according to World Health Organization (WHO). All patients were required to provide written informed consent. The study was approved by the Hellenic Cooperative Oncology Group (HeCOG) Protocol Review Committee and by Institutional Review Boards in participating institutions.
Annals of Oncology
original article
Annals of Oncology
Selective patient and tumor characteristics are presented in Tables 1 and 2. There were no significant differences in major characteristics between the two treatment groups except for vena cava obstruction symptoms and lymph node involvement (P = 0.04 in both cases).
results
treatment characteristics A total of 969 and 995 cycles of chemotherapy were given in groups A and B, respectively. The median number of cycles administered was five in group A (range 1–8) and six (range 1– 10) in group B. Significantly more cycles in group A were administered at full dose (group A: 72% versus group B: 63.5%, P < 0.001) (Table 3). The relative dose intensity of paclitaxel was 0.85 and of gemcitabine 0.82 in group A and 0.77 in group B. The median cumulative dose for carboplatin was 2600 mg. Totally, 108 patients in group A and 116 in group B completed at least six cycles of chemotherapy.
patient and tumor characteristics From October 2000 to March 2004, 512 patients were randomized in this study. Sixty (12%) patients were ineligible. Thirty-two had stage other than IIIb wet or IV and 28 were ineligible for other reasons (mostly because of PS >1, history of prior cancer and no stage data stated at randomization or in their medical records). Furthermore, for four patients (one in group A and three in group B), medical records were incomplete while 10 patients (five in each group) never started protocol treatment. Two patients in each group received the opposite treatment than the one allocated to. The remaining 452 eligible patients were analyzed according to the intention-to-treat principle (N = 452; group A: 225, group B: 227), except treatment and toxicity analyses, which were conducted on the actual treatment arm (N = 438; group A: 219, group B: 219) (Figure 1).
response A total of 80 patients, 42 in group A and 38 in group B, were considered nonassessable for response. Response rate was calculated in the intention-to-treat population (Table 4).
Figure 1. Progress through the various stages of the trial. Survival status update on September 2006.
Volume 19 | No. 1 | January 2008
doi:10.1093/annonc/mdm430 | 117
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
bone metastasis (no versus yes), liver metastasis (no versus yes), skin/soft tissue metastasis (no versus yes) other metastasis (no versus yes) and number of metastatic sites (<3 versus ‡3). All the analysis (except toxicity and treatment characteristics) was conducted according to the ‘intention-to-treat’ principle.
original article
Annals of Oncology
Table 1. Selected patient characteristics
n Age (years) Median Range
Group A (paclitaxel + gemcitabine)
Group B (carboplatin + gemcitabine)
225
227
63 42–82 n
n
%
63 36–83 n
%
194 31
86 14
184 43
81 19
104 121
46 54
106 121
47 53
30 191 4
13 85 2
32 191 4
14 84 2
1 220 4
0.4 98 1.8
8 215 4
3.5 95 1.8
47 178 –
21 79 –
42 184 1
18.5 81 0.4
44 181
20 80
35 192
15 85
*P = 0.04. ECOG, Eastern Cooperative Oncology Group.
Three complete responses (CRs; dissappearance of all evidence of disease for a duration of at least 4 weeks) and 67 partial responses (PRs; a 50% or greater reduction in the product obtained from measurement of each lesion for at least 4 weeks and no appearance of new lesions) were achieved in group A for an overall response rate (ORR) of 31% [95% confidence interval (CI), 25.12–37.60]. One CR and 61 PRs were achieved in group B for an ORR of 27% (95% CI, 21.63–33.60). The ORR, as given by investigators, did not differ significantly between the two treatment groups (P = 0.41).
survival, TtP and TTF At the time of the analysis, after a median follow-up of 44 months (range 0.01–70.56+), 402 patients (89%), 202 in group A and 200 in group B, had died and 418 (92.5%), 209 in group in group A and 209 in group B, had relapsed. Survival, TtP and TTF did not differ significantly between treatment groups. Overall, median survival was 9.97 months (range 0.01–70.56 months, 95% CI 8.74–12.0) for group A and 10.49 months (range 0.01–55.74+, 95% CI 9.04–11.94) for group B (P = 0.99) (Figure 2). The 1-year survival rate was 42% in both groups. TtP was 5.02 months (range 0.01–70.56 months, 95% CI 4.31–5.72) for
118 | Kosmidis et al.
Stage IIIbwet IV Tumor histology Squamous cell Adenocarcinoma Large cell Mixed Undifferentiated Unclassified Unknown Number of metastatic sites <3 ‡3 Unknown Lymph nodes involvement* Yes No Unknown Liver metastasis Yes No Unknown Bone metastasis Yes No Unknown Brain metastasis Yes No Unknown Adrenal glands metastasis Yes No Unknown
Group A (paclitaxel + gemcitabine)
Group B (carboplatin + gemcitabine)
225 n
%
227 n
%
30 195
13 87
31 196
14 86
58 119 10 5 14 14 5
26 53 4 2 6 6 2
67 115 9 4 8 19 5
29.5 51 4 1.8 3.5 8 2
155 66 4
69 29 2
153 71 3
67 31 1
134 87 4
60 39 1.8
157 67 3
69 29.5 1.3
35 186 4
16 83 1.8
33 191 3
14.5 84 1.3
83 138 4
37 61 1.8
98 126 3
43 55.5 1.3
34 187 4
15 83 1.8
31 193 3
14 85 1.3
36 185 4
16 82 1.8
33 191 3
14.5 84 1.3
*P = 0.04.
group A and 5.11 months (range 0.01–55.74, 95% CI 4.43– 5.80) for group B (P = 0.35) (Figure3). The respective 1-year progression-free survival rates were 14% and 12%, respectively. TTF was 3.97 months (range 0.01–70.56 months, 95% CI 3.22–4.71) for group A and 4.13 months (range 0.01–55.74, 95% CI 3.15–5.11) for group B (P = 0.89). (Table 5). The Cox multivariate regression analysis revealed that in the presence of treatment group (P = 0.60) and stage of disease (P = 0.26), several prognostic factors including PS [1 versus 0: hazard ratio (HR) = 1.30, 95% CI 1.06–1.60, P = 0.01], histology type (adenocarcinoma versus squamous: HR = 1.37, 95% CI 1.07–1.74, P = 0.01; other versus squamous:
Volume 19 | No. 1 | January 2008
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
Gender Male Female ECOG performance status 0 1 Loss of weight (>10%) Yes No Unknown Vena cava obstruction symptoms* Yes No Unknown Prior radiotherapy Yes No Unknown Prior lung surgery Yes No
Table 2. Selected tumor characteristics
original article
Annals of Oncology
Table 3. Treatment characteristics Group B (carboplatin + gemcitabine)
219 969 5 1–8 72%
219 995 6 1–10 63.5%
57 20.25–68.3
– –
0.85 0.30–1.02
– –
546 182–672
512.5 256–636
0.82 0.27–1.01
0.77 0.38–0.95
– –
2600 430–6560
–
1.0 0.59–1.67
Figure 2. Survival of patients treated with paclitaxel + gemcitabine (red line) or carboplatin + gemcitabine (blue line) (P = 0.99).
a
90% of the dose defined in the protocol (for both drugs). DI, dose intensity; RDI, relative dose intensity.
Table 4. Best response to treatment
n CR PR ORR* SD PD NA Unknown
Group A (paclitaxel + gemcitabine)
Group B (carboplatin + gemcitabine)
225 n (%) 3 (1) 67 (30) 70 (31) 62 (28) 47 (21) 42 (19%) 4 (2)
227 n (%) 1 (0.4) 61 (27) 62 (27) 66 (29) 58 (26) 38 (17) 3 (1)
95% CI 0.27–3.85 23.9–36.21 25.12–37.60 21.82–33.88 15.77–26.79 13.79–24.38 –
95% CI 0.01–2.43 21.22–33.14 21.63–33.60 23.26–35.45 20.0–31.74 12.13–22.24 –
*P = 0.41. CR, complete response; PR, partial response; ORR, overall response rate; SD, stable disease; PD, progression disease; NA, nonassessable.
HR = 1.67, 95% CI 1.23–2.26, P = 0.001), presence of symptoms (yes versus no: HR = 1.47, 95% CI 1.06–2.05, P = 0.02) and number of metastatic sites (‡3 sites versus <3: HR = 1.67, 95% CI 1.32–2.12, P < 0.001) were related to significantly poorer survival. On the other hand, the risk of death was significantly lower for women (female versus male: HR = 0.71, 95% CI 0.53–0.94, P = 0.02), for patients having prior lung surgery
Volume 19 | No. 1 | January 2008
Figure 3. Time to progression of patients treated with paclitaxel + gemcitabine (red line) or carboplatin + gemcitabine (blue line) (P = 0.35).
(yes versus no: HR = 0.73, 95% CI 0.55–0.98, P = 0.04) and patients with lung metastasis at entry (yes versus no: HR = 0.62, 95% CI 0.49–0.79, P < 0.001) (Table 6). The Cox multivariate regression analysis for TtP revealed that in the presence of treatment group (P = 0.49) and stage of disease (P = 0.12), the hazard of disease progression at any time was significantly higher for patients with worse PS (1 versus 0: HR = 1.36, 95% CI 1.12–1.67, P = 0.002) and larger number of metastatic sites (‡3 sites versus <3: HR = 1.87, 95% CI 1.48–2.36, P < 0.001). On the contrary, the risk of disease progression was significantly lower for women (female versus male: HR = 0.79, 95% CI 0.60–1.03, P = 0.08), for patients with lung metastasis at entry (yes versus no: HR = 0.62, 95% CI 0.49–0.78, P < 0.001), for patients with bone marrow metastasis (yes versus no: HR = 0.35, 95% CI 0.11–1.12, P = 0.08) and for patients having prior lung surgery (yes versus no: HR = 0.68, 95% CI 0.52–0.89, P = 0.005) (Table 6).
toxicity One toxic death occurred in group A from pancytopenia after the completion of the first cycle. Additionally, one patient died from pancytopenia after the completion of the fourth
doi:10.1093/annonc/mdm430 | 119
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
n Number of cycles delivered Median Range % of cycles at full dosea DI of paclitaxel delivered Median (mg/m2/week) Range RDI of paclitaxel Median Range DI of gemcitabine delivered Median (mg/m2/week) Range RDI of gemcitabine Median Range Cumulative dose of carboplatin Median Range RDI of carboplatin Median Range
Group A (paclitaxel + gemcitabine)
original article
Annals of Oncology
Table 5. Survival, time to disease progression and TTF Carboplatin + gemcitabine (n = 227)
202/225 0.01–70.56 9.97 8.74–12.0 42%
200/227 0.01–55.74+ 10.49 9.04–11.94 42%
209/225 0.01–70.56 5.02 4.31–5.72 14%
209/227 0.01–55.74 5.11 4.43–5.80 12%
217/225 0.01–70.56 3.97 3.22–4.71
220/227 0.01–55.74 4.13 3.15–5.11
*log-rank P = 0.99; **log-rank P = 0.35; ***log-rank P = 0.89. TTF, time to treatment failure; CI, confidence interval.
cycle and another died from sepsis febrile neutropenia after the completion of the third cycle in group B. Eight patients in group A and 13 in group B developed febrile neutropenia (group A: 4% versus group B: 6%, P = 0.37). Data for hospitalizations, antibiotics, transfusions and supportive care were not available for all patients. Among patients with available data, no significant differences were found between the two groups in terms of hospitalization needs (group A: 42 of 198 = 21% versus group B: 53 of 201 = 26%, P = 0.24). Regarding supportive care, significantly more patients in group B received granulocyte colony-stimulating factor (G-CSF) (group A: 79 of 193 = 41% versus group B: 105 of 187 = 56%, P = 0.004) and Erythropoetin (group A: 46 of 185 = 25% versus group B: 108 of 189 = 57%, P < 0.001). Additionally, red blood cell (RBC) transfusions (group A: 12 of 194 = 6% versus group B: 52 of 195 = 27%, P < 0.001) and platelet transfusions (group A: 2 of 193 = 1% versus group B: 15 of 191 = 8%, P = 0.001) were significantly more frequent in group B. Regarding antibiotics, no significant differences were found between the two groups (group A: 41 of 193 = 21% versus group B: 37 of 193 = 19%, P = 0.70). The rates of severe (grade 3/4) hematologic toxic effects (neutropenia, leucopenia, thrombocytopenia, anemia) were significantly worse in group B, as shown in Table 7. Among grade 3/4 non-hematologic toxic effects, alopecia and neurotoxicity were significantly worse in group A (62.7% versus 6.2%, P < 0.001 and 5.2% versus 0.5%, P = 0.006, respectively). Other severe toxic effects did not differ significantly between the two groups (Table 8).
discussion In this phase III randomized study, we targeted the population of patients with NSCLC, stage IIIbwet and IV and specifically
120 | Kosmidis et al.
Survival Treatment group Paclitaxel + gemcitabine Carboplatin + gemcitabine Stage IIIbwet IV Gender Male Female Prior lung surgery No Yes Symptoms No Yes PS 0 1 Histology type Squamous Adenocarcinoma Other Lung metastasis at entry No Yes Number of metastatic sites <3 ‡3 Time to disease progression Treatment group Paclitaxel + gemcitabine Carboplatin + gemcitabine Stage IIIbwet IV Gender Male Female Prior lung surgery No Yes PS 0 1 Bone marrow metastasis at entry No Yes Lung metastasis at entry No Yes Number of metastatic sites <3 ‡3
HR
95% CI
P value
1 1.06
0.86–1.30
0.60
1 1.20
0.87–1.66
0.26
1 0.71
0.53–0.94
0.02
1 0.73
0.55–0.98
0.04
1 1.47
1.06–2.05
0.02
1 1.30
1.06–1.60
0.01
1 1.37 1.67
1.07–1.74 1.23–2.26
0.01 0.001
1 0.62
0.49–0.79
<0.001
1 1.67
1.32–2.12
<0.001
1 1.07
0.88–1.30
0.49
1 1.27
0.93–1.73
0.12
1 0.79
0.60–1.03
0.08
1 0.68
0.52–0.89
0.005
1 1.36
1.12–1.67
0.002
1 0.35
0.11–1.12
0.08
1 0.62
0.49–0.78
<0.001
1 1.87
1.48–2.36
<0.001
HR, hazard ratio; CI, confidence interval; PS, performance status.
Volume 19 | No. 1 | January 2008
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
Survival Events Range (months) Median* 95% CI 1-year survival rate Time to disease progression Events Range (months) Median** 95% CI 1-year disease-free rate TTF Events Range (months) Median*** 95% CI
Paclitaxel + gemcitabine (n = 225)
Table 6. Estimated HRs and 95% CIs for survival and time to disease progression
original article
Annals of Oncology
Table 7. Hematologic toxicity
Neutropenia* Leucopenia** Thrombocytopenia*** Anemia****
Paclitaxel + gemcitabine (n = 219) Grade 3 Grade 4 No. % No. %
Carboplatin + gemcitabine (n = 219) Grade 3 Grade 4 No. % No. %
25/212 11.8 13/212 13/213 6.1 4/213 3/212 1.4 2/212 9/212 4.2 2/212
32/211 34/212 22/212 23/212
6.1 1.9 0.9 0.9
15.2 16 10.4 10.8
30/211 14.2 7/212 3.3 32/212 15.1 5/212 2.4
*P = 0.009; **P = 0.001; ***P < 0.001; ****P = 0.007. Table 8. Non-hematologic toxicity Carboplatin + gemcitabine (n = 219) Grade 3 Grade 4 No. % No. %
1/211 0.5 0/211 0 5/206 2.4 0/206 0 3/208 1.4 0/208 0 1/207 0.5 2/207 1.0 133/212 62.7 0/212 0 11/211 5.2 0/211 0 2/219 0.9 0/219 0 1/212 0.5 1/212 0.5 4/219 1.8 0/219 0 0/219 0 1/219 0.5
2/210 2/208 1/209 3/208 13/210 1/210 3/219 1/210 3/219 1/219
1.0 1.0 0.5 1.4 6.2 0.5 1.4 0.5 1.4 0.5
0/210 0/208 0/209 0/208 0/210 0/210 0/219 0/210 0/219 0/219
0 0 0 0 0 0 0 0 0 0
HSR, rash, edema, pruritus and allergic reaction; mucositis, stomatitis, diarrhea. *P < 0.001; **P = 0.006.
those with a PS zero to one. We tested a nonplatinum combination of third-generation compounds, namely paclitaxel and gemcitabine, versus a combination of carboplatin and gemcitabine. The overall median survival was similar in the two groups (9.97 versus 10.49 months). Moreover, the 1-year survival rate was 42% for both groups. Median TtP was also similar in the two groups (5.02 versus 5.11 months). Finally, ORRs were 31% for group A and 27% for group B based on an intention-to-treat analysis. Based on the results, we conclude that this study has failed to meet each primary end point. This is in accordance with the results of our previous randomized trial, which compared the nonplatinum combination of paclitaxel and gemcitabine with a combination of carboplatin and paclitaxel [4]. In that study, we had included patients with NSCLC stage IIIa and IIIbnon-wet and also patients with PS two. In a subgroup analysis, we had found that patients with PS two have a significantly worse outcome in both groups [4]. This was also consistent with the results from other randomized studies [10, 11]. The above results indicated that the subgroup of NSCLC with poor PS required a different treatment approach. Therefore, in the current study we did not include patients with PS two. On the other hand, it is known that patients with NSCLC stage IIIa
Volume 19 | No. 1 | January 2008
references 1. Ginsberg RJ, Vokes EE, Rosenzweig K. Non small cell lung cancer. In De Vita Jr, Hellman S, Rosenber SR (eds): Cancer.: Principles and Practice of Oncology, 6th edition. Philadelphia, PA: Lippinicott-Raven 2001; 925–983. 2. 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 randomized clinical trials. BMJ 1995; 311: 899–909. 3. 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. 4. Kosmidis P, Mylonakis N, Nicolaides C et al. Paclitaxel plus carboplatin versus gemcitabine plus paclitaxel in advanced non-small-cell lung cancer: a phase III randomized trial. J Clin Oncol 2002; 20: 3578–3585. 5. Georgoulias V, Papadakis E, Alexopoulos A et al. Platinum-based and nonplatinum-based chemotherapy in advanced non-small cell lung cancer: a randomized multicentre trial. Lancet 2001; 357: 1478–1484.
doi:10.1093/annonc/mdm430 | 121
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
Nausea/vomiting HSR Mucositis Infection Alopecia* Neurotoxicity** Hepatotoxicity Cardiotoxicity Fatigue Nephrotoxicity
Paclitaxel + gemcitabine (n = 219) Grade 3 Grade 4 No. % No. %
and IIIbnonwet require a concurrent chemoradiotherapy approach [12, 13] and as a result we did not include such patients in our current study, unlike most of the studies in advanced NSCLC in the literature published so far. Despite these essential differences in patient selection in the current study, we did not find significant differences in activity between the nonplatinum combination and the combination of carboplatin and gemcitabine. A recent meta-analysis comparing platinum-based versus nonplatinumbased chemotherapy in advanced NSCLC confirmed that there is a marginal 1-year survival benefit with platinumbased regiments which is lost when platinum therapies are compared with third-generation-based combination regimens [14]. Toxicity profile was different between the two groups of our study. Myelotoxicity was significantly more severe in group B. Grade 3/4 anemia, leucopenia, neutropenia and thrombocytopenia were significantly worse in group B. As a result, patients in group B received significantly more RBCs and platelet transfusions and received more often G-CSF and erythropoietin. Febrile neutropenia, need for antibiotics and infections did not differ significantly between the two groups. On the other hand, grade 3 alopecia and grade 3 neurotoxicity were significantly more frequent in group A as expected, due to the use of paclitaxel. In our previous study comparing paclitaxel and gemcitabine with carboplatin and paclitaxel, toxicity was mild and no significant differences were found in toxicity profiles between the two groups [4]. In the recent meta-analysis, mentioned above, comparing platinum-based versus nonplatinum-based chemotherapy in advanced NSCLC, the toxicity of platinum-based regimens was significantly higher for hematologic toxicity [14]. This is consistent with the findings of our study. Most of the targeted novel agents have been combined with carboplatin–paclitaxel or cisplatin–gemicitabine. Both of these combinations are responsible for significant toxic effects such as alopecia and nephrotoxicity, respectively. Based on our data, paclitaxel and gemcitabine is equally active to carboplatin and gemcitabine in NSCLC patients of good PS and stage IIIbwet and IV. These regimes have different toxicity profiles and could be delivered according to the patient’s needs.
original article 6. Schiller JH, Harrington D, Belani C et al. Comparison of four chemotherapy regimens for advanced non-small-cell lung cancer. N Engl J Med 2002; 346: 92–98. 7. Bonomi PD, Finkelstein DM, Ruckdeschel JC et al. Combination chemotherapy versus single agents followed by combination chemotherapy in stage IV nonsmall-cell lung cancer: a study of the Eastern Cooperative Oncology Group. J Clin Oncol 1989; 7: 1602–1613. 8. Klastersky J, Sculier P, La Croix 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. 9. Sandler A, Gray R, Perry M et al. Paclitaxel-carboplatin alone or with bevacizumab for non-small cell lung cancer. N Engl J Med 2005; 355: 2542–2550. 10. Sweeney CJ, Zhu J, Sandler AB et al. Outcome of patients with a performance status of 2 in Eastern Cooperative Oncology Group Study E 1594: a phase III trial in patients with metastatic non-small cell lung carcinoma. Cancer 2001; 92: 2639–2647.
Annals of Oncology
11. Georgoulias V, Ardavanis A, Tsiafaki X et al. Vinorelbine plus cisplatin versus docetaxel plus gemcitabine in advanced non-small cell lung cancer: a phase III randomized trial. J Clin Oncol 2005; 23: 2937–3945. 12. Furuse K, Fukuoka M, Kawahara M et al. Phase III study of concurrent versus sequential thoracic radiotherapy in combination with mitomycin, vindesine and cisplatin in unresectable stage III non-small-cell lung cancer. J Clin Oncol 1999; 17: 2692–2699. 13. Fournel P, Robinet G, Thomas P et al. Randomized phase III trial of sequential chemoradiotherapy compared with concurrent chemoradiotherapy in locally advanced non-small-cell lung cancer: Groupe Lyon-Saint-Etienne d’Oncologie Thoracique-Groupe Francais de Pneumo-Cancerologie NPC 95-01 Study. J Clin Oncol 2005; 23: 5910–5917. 14. D’ Addario G, Pintilie M, Leighl N et al. Platinum-based versus non-platinumbased chemotherapy in advanced non-small-cell lung cancer. A meta-analysis of the published literature. J Clin Oncol 2005; 23: 2926–2936.
Downloaded from http://annonc.oxfordjournals.org/ at Florida International University on June 18, 2015
122 | Kosmidis et al.
Volume 19 | No. 1 | January 2008