Original article

Annals of Oncology 16: 1069 1075, 2005 doi:10.1093/annonc/mdi216 Published online 28 April 2005

Randomized phase III trial comparing cisplatin etoposide to carboplatin paclitaxel in advanced or metastatic non-small cell lung cancer
C. P. Belani1*, J. S. Lee2, M. A. Socinski3, F. Robert4, D. Waterhouse5, K. Rowland6, R. Ansari7, R. Lilenbaum8 & R. B. Natale9
1 4

University of Pittsburgh Cancer Institute, Pittsburgh, PA; 2MD Anderson Cancer Center, Houston, TX; 3University of North Carolina, Chapel Hill, NC; University of Alabama at Birmingham, Birmingham, AL; 5Oncology/Hematology Care, Cincinnati, OH; 6Carle Cancer Center, Urbana, IL; 7 Michiana Hematology/Oncology, South Bend, IN; 8Mt. Sinai Comprehensive Cancer Center, Miami Beach, FL; 9Cedars-Sinai Comprehensive Cancer Center, Los Angeles, CA, USA Received 25 August 2004; revised 27 January 2005; accepted 8 February 2005

Downloaded from http://annonc.oxfordjournals.org/ at Library on February 9, 2012

Background: The present study was designed to evaluate the efcacy and safety of the regimen of carboplatin plus paclitaxel (investigational arm) versus the reference regimen of cisplatin plus etoposide for the treatment of advanced or metastatic non-small-cell lung cancer. Patients and methods: A total of 369 patients were enrolled, 179 on arm A (cisplatin 75 mg/m2 and etoposide 100 mg/m2) and 190 on arm B (carboplatin AUC = 6 mg/ml min and paclitaxel 225 mg/m2), with cycles repeated every 3 weeks. The arms were well balanced with respect to age, performance status, weight loss, stage of disease and disease measurability. However, signicantly more women were randomized to arm A than to arm B (P = 0.039). Results: The objective response rate (ORR) was 15% on arm A compared with 23% on arm B (P = 0.061). Median survival time, time to progression and 1-year survival rates for arms A and B were 274 days and 233 days (P = 0.086), 111 days and 121 days (P = 0.877), and 37% and 32%, respectively. The most prevalent toxicities were neutropenia and leukopenia and they occurred at a higher rate in arm A than in arm B. Conclusion: There was no statistically signicant survival advantage for carboplatin paclitaxel compared with cisplatin etoposide. However, there was an overall benet in quality of life with the carboplatin paclitaxel regimen. Key words: carboplatin, cisplatin, combination therapy, etoposide, non-small-cell lung cancer, paclitaxel

Background
Lung cancer remains the leading cause of cancer death for both men and women. In the year 2004, an estimated 173 770 cases of lung cancer will be diagnosed in the USA, 93 110 in men and 80 660 in women [1]. Non-small-cell lung cancer (NSCLC) comprises approximately 80% of all cases of lung cancer [2]. Treatment for NSCLC is based on the clinical stage of the disease. Patients with early stage I or II disease undergo potentially curative surgical resection, and recently adjuvant chemotherapy has been shown to provide a survival benet over surgery alone. Low survival rates fueled the debate over

*Correspondence to: Prof. C. P. Belani, University of Pittsburgh Cancer Institute, UPMC Cancer Pavilion, 150 Centre Avenue, Suite 570, Pittsburgh, PA 15232, USA. Tel: +1 412 648 6619; Fax +1 412 648 6579; E-mail: belanicp@upmc.edu
q 2005 European Society for Medical Oncology

the benet of systemic therapy, especially for patients with advanced but potentially resectable (IIIA, N2) disease. However, in those who present with locally advanced unresectable disease (stage IIIA/B) without supraclavicular/cervical adenopathy or pleural effusion, a multimodality approach is increasingly being used. Randomized multinational trials have demonstrated the efcacy of various combination platinumbased doublets [3 6]. The superiority of a two-drug combination over a single agent was demonstrated by the Cancer and Leukemia Group B (CALGB) 9730 trial [7], and further conrmed in a meta-analysis performed by Delblado et al. [8]. Taken together, these results led to combination doublet chemotherapy becoming an accepted standard of care for stage IV disease. Despite the research and progress, the search for the optimal regimen for advanced NSCLC continues. At the time of initiation of the present study, cisplatin etoposide was the most commonly used regimen for advanced

1070 and metastatic disease. A study by the Southwestern Oncology Group (SWOG) established that there was no advantage to increasing the dose or dose intensity of cisplatin [9]. Higher doses of etoposide did not increase the therapeutic index of the cisplatin etoposide combination in a study conducted by the Cancer and Leukemia Group B (CALGB) [10]. The regimen of cisplatin 75 mg/m2 plus etoposide 100 mg/m2 daily for three consecutive days was adopted as the reference arm by the Eastern Cooperative Oncology Group (ECOG) for its 5592 study [11] and was also used in the present study. A recent retrospective analysis of phase III studies conducted between 1973 and 1994 in patients with advanced NSCLC concluded that gains in response and survival were modest at best, even with regimens that showed promise in earlier phase studies [12]. The optimal regimen being sought was not identied in the studies included in that analysis. Nevertheless, options for treatment of NSCLC expanded tremendously throughout the 1990s with the development of new agents such as the taxanes, gemcitabine, CPT-11 and vinorelbine. In addition, carboplatin began to replace cisplatin as the platinum compound of choice in combination doublet regimens as it had equivalent efcacy with lower toxicity compared with cisplatin [1315]. The carboplatin paclitaxel combination had emerged as an encouraging regimen from early phase II studies, with response rates ranging from 27% to 62% [1619], and thus was chosen as the investigational arm for the present study. The present study was designed to evaluate the efcacy and safety of the regimen of carboplatin plus paclitaxel compared with the reference regimen of cisplatin plus etoposide to treat advanced or metastatic NSCLC.
Treatment plan
Patients were randomly assigned to either treatment arm A or treatment arm B (Figure 1). The cycle duration was 3 weeks on both arms. Patients on arm A received cisplatin (75 mg/m2, 1 h infusion) on day 1 plus etoposide (100 mg/m2, 45 min infusion) i.v. on days 1 3. Those on arm B received carboplatin (AUC = 6 mg/ml min, 30 min infusion) on day 1 immediately after paclitaxel (225 mg/m2, 3 h infusion). Actual body weight was used for dose calculations in both arms. Patients on arm B received premedication consisting of oral dexamethasone 20 mg administered  12 and 6 h before paclitaxel, diphenhydramine 50 mg i.v. and cimetidine 300 mg or ranitidine 50 mg i.v. prior to paclitaxel. Treatment was administered every 3 weeks if the absolute granulocyte _ _ count (AGC) and platelet counts were > 1500/mm3 and > 100 000/mm3, respectively. Treatment could be delayed for a maximum of 2 weeks to allow recovery from toxicity. In subsequent cycles, doses were reduced for severe hematologic and non-hematologic toxicities (graded using the NCI Common Toxicity Criteria) according to a specic modication schema. A maximum of three dose reductions per patient was allowed. Dose escalation was prohibited. Patients were not allowed to receive any other anticancer drugs, including immunotherapy, while in the study. Growth factor support, antiemetics, and therapy for sensory neuropathies were permitted at the investigators discretion. Treatment duration was based on patient response. Patients achieving complete or partial responses were recommended to receive two cycles of treatment beyond maximum response. A maximum of 10 cycles of therapy was permitted. Prior to the completion of six cycles, patients could be withdrawn from the study because of excessive toxicity, disease progression or pregnancy, or on request.

Downloaded from http://annonc.oxfordjournals.org/ at Library on February 9, 2012

Assessment of efcacy, toxicity and quality of life

Survival was the primary endpoint of the study and was dened as the interval between the study date and the date of death. The analysis of survival included all randomized patients. Secondary parameters included the objective response rate, time to disease progression, safety prole and quality of life. Response was assessed every 6 weeks with a conrmatory measurement not less than 4 weeks after the initial claim of response. Complete response was dened as the disappearance of all clinical evidence of cancer. Partial response for bidimensionally measurable disease was _ a > 50% decrease in the sum of the products of measured lesions. Partial _ response for unidimensionally measurable disease was a > 30% decrease

Patients and methods

Eligibility criteria
_ Men and women aged > 18 years with histologically or cytologically conrmed inoperable stage IIIB or IV NSCLC who were not candidates for denitive radiotherapy or combined modality therapy were eligible for this study. At least one bidimensionally measurable lesion or at least one _ unidimensionally measurable lesion > 1 cm had to be denable by radiography or CT scan. Patients could not have received any prior chemotherapy or hormonal therapy. Prior radiation therapy and/or major surgery had to have been completed at least 3 weeks prior to registration with complete recovery. There could be no previous radiation to the only _ site of measurable disease. Karnofsky performance status had to be > 70% _ and life expectancy > 12 weeks. Adequate bone marrow and organ function was required, as well as a negative baseline pregnancy test for female patients. Patients could not have a signicant history of cardiac disease, serious active infection or evidence of current peripheral neuropathy. The protocol was approved by Institutional Review Boards with jurisdiction over the specic sites that registered patients. Each patient gave written informed consent prior to enrollment. The following pretreatment evaluations were performed: medical history, physical examination, vital signs, height and body weight, performance status, measurement of tumor(s), complete blood count with differential and platelet count, hemoglobin and chemistries (serum creatinine, bilirubin, glucose, alkaline phosphatase, electrolytes, serum calcium and serum magnesium).

Ar Arm A: Cisplatin: 75 mg/m2, Day 1 Etoposide: 100 mg/m2, Days 1-3

Randomization

Repeat cycle every 21 days Arm B: Carboplatin: AUC = 6 mg/mL min, Day1 Paclitaxel: 225 mg/m2, Day 1 Repeat cycle every 21 days

Figure 1. Treatment schema.

1071
in linear tumor measurements. There could be no simultaneous progression of any lesion and no appearance of new lesions for claims of partial response, and non-measurable lesions had to remain stable or regress. Stable disease was dened as <50% decrease or <25% increase in the sum of the products of measured lesions, with no appearance of new _ lesions. Progression was dened as an increase of > 25% in measured lesions compared with baseline (stable disease patients) or with maximum response (partial response patients), or the appearance of new lesions. Quality of life was assessed using the Functional Assessment of Cancer Therapy-Lung Scale (version 3) [20]. The FACT-L is made up of ve subscales: physical well-being, social/family well-being, emotional well-being, functional well-being and an additional concerns subscale consisting of questions specic to lung cancer. Quality of life was assessed at baseline, on day 1 of the third and fth cycles of therapy and at week 26.

were as follows: disease progression (90), study completion (30), toxicity (20), refusal to receive further treatment (14), death (7), protocol violation (2), lost to follow-up (1) and other reasons (14). Among the 188 patients in arm B, the reasons for discontinuation included disease progression (94), toxicity (30), study completion (29), death (11), refusal of further treatment (6), protocol violation (1), lost to follow-up (1) and other reasons (16). A greater proportion of patients on arm A than on arm B received second- and/or third-line chemotherapy. Ninety of 178 patients (51%) on arm A received additional

Table 1. Patient characteristics

Statistical methods
Patients were stratied at enrollment based on performance status (Karnofsky 90100% and 70 80%), weight loss in the 6 months prior to _ registration (<5% and > 5%), stage of disease (IIIB and IV) and measurability of disease (bidimensional and unidimensional). All analyses were based on these stratication factors. To ensure a power of 85% with a = 0.05, a total of 162 patients per treatment arm were required to determine a 40% improvement in median survival. All patients enrolled on the study were included in an intent-totreat efcacy analysis. The CochranMantelHaenszel row sum test was used to assess baseline comparability of the treatment arms, while comparability between groups for continuous variables was assessed using the Wilcoxon test. An a level of 5% (two-tailed) was used for statistical signicance. Within-group analyses were performed using paired t-tests for continuous variables and the CochranMantelHaenszel test for categorical variables. The KaplanMeier method [21] was used to analyze overall survival and time to disease progression. An analysis of factors that may predict survival and/or time to disease progression was performed using univariate and multivariate Cox proportional hazards regression analyses. Response rates were analyzed using the x2 test. Changes in quality of life from baseline to cycles 3 and 5 and week 26 were assessed using longitudinal analysis and descriptive statistics. Patients enrolled in study Patients who received therapy Sex Male Female Race White Black Hispanic Oriental Filipino Native American Other Mean age (range) (years) Karnofsky performance status 90100% 7080% Measurable disease Bidimensional

Arm A (%) 179 178

Arm B (%) 190 188

Downloaded from http://annonc.oxfordjournals.org/ at Library on February 9, 2012

100 79 145 19 5 4 3 2 1 60.7 (3579)

126 64 151 19 11 6 2 1 61.3 (2880)

102 (57%) 77 (43%)

107 (56%) 83 (44%)

169 (94%) 10 (6%)

176 (93%) 14 (7%)

Results
Patient characteristics
Between 19 May 1995 and 17 July 1996, 369 patients were enrolled on the study at 42 sites. Patient characteristics are summarized by treatment group in Table 1. The two treatment arms differed signicantly in terms of gender. Specically, there were more women in arm A and more men in arm B (P = 0.039). The treatment arms were well balanced with respect to the stratication factors, including performance status, weight loss, stage of disease and measurability of disease. A combined total of 1508 treatment cycles were administered, 696 in arm A and 812 in arm B. The median number of cycles per patient was three in arm A (range 1 10) and four in arm B (range 110). A greater proportion of patients remained on treatment and at full dose on arm B than in arm A. One patient on arm A and one patient on arm B refused therapy. One patient was removed from the study because of deep vein thrombosis prior to treatment. Reasons for discontinuation for the 178 patients in arm A receiving treatment

Unidimensional Stage of disease IIIb IV

37 (21%) 142 (79%)

44 (23%) 146 (77%)

Weight loss in 6 months prior to enrolment <5% > 5% _ Reason for removal from study Study completed Progressive disease Refused further treatment Toxicity Lost to follow-up Protocol violation Death Other Total no. of cycles 30 90 15 20 1 2 7 14 696 29 94 7 30 1 1 11 17 812 117 (65%) 62 (35%) 121 (64%) 69 (36%)

1072 chemotherapy. Fifty-nine of these patients (33%) received a taxane-containing regimen [carboplatin paclitaxel or paclitaxel alone (n = 46) or docetaxel alone n = 13]. In contrast, 73 of 188 patients (39%) on arm B received additional chemotherapy, with 23 (12%) of these receiving a taxane regimen [carboplatin paclitaxel or paclitaxel alone (n = 23) and docetaxel alone (n = 5)]. patients available for response on arm A, there were four complete responders and 22 partial responders. Of the 177 patients available for response on arm B, there was one complete responder and 42 partial responders. The rate of stable disease was 45% and 44% for arms A and B, respectively, while the rate of progressive disease was 32% and 26%, respectively. Compliance for the FACT-L quality-of-life questionnaire was 180 patients (79 arm A and 101 arm B) at baseline and cycle 3, 107 patients (49 arm A and 58 arm B) at baseline and cycle 5, and 65 patients (33 arm A and 32 arm B) at baseline and week 26. The difference in total score between baseline and cycle 3 was signicantly better (P = 0.038 based on the Wilcoxon rank sum test) for arm B than for arm A. Because of the decline in the number of quality-of-life questionnaires available for analysis at cycle 5 (P = 0.950) and week 26 (P = 0.475), the differences in scores for the two treatment arms did not reach statistical signicance.
Downloaded from http://annonc.oxfordjournals.org/ at Library on February 9, 2012

Efcacy results
The overall survival for patients in each of the treatment arms is shown in Figure 2. Patients in arm A (cisplatin etoposide) had a median survival time of 274 days, whereas those on arm B (carboplatin paclitaxel) had a median survival time of 233 days. The difference was not statistically signicant (P = 0.086). One-year survival rates were 37% for arm A and 32% for arm B. Median time to progression (Figure 3) was 111 days for patients on arm A and 121 days for patients on arm B (P = 0.877). Univariate regression analyses identied Karnofsky performance status and prior weight loss as signicant predictors of both overall survival and time to progression, while gender only predicted overall survival. The female gender is a signicant prognostic factor by univariate analysis (P = 0.028) but not by multivariate analysis (P = 0.129). In general, poor performance status and greater weight loss were predictors of decreased overall survival and time to progression; men had worse overall survival than women. Complete analyses of prognostic factors for overall survival and time to progression are provided in Table 2. The response analysis was performed on all patients who had uni- or bidimensionally measurable disease and had received at least two cycles of treatment. A summary of response is provided in Table 3. Overall response was 15% for arm A and 23% for arm B (P = 0.061). Among 164

Safety
Adverse events reported during the study and graded according to the NCI Common Toxicity Criteria were subsequently converted to preferred terms for analysis using a modied COSTART translation dictionary. Of 178 patients receiving therapy in arm A (cisplatin etoposide), 177 (99.4%) experienced at least one adverse event. All 188 patients treated in arm B (carboplatin paclitaxel) experienced at least one adverse event. A summary of grade 3 and 4 toxicities is presented in Table 4. The most prevalent toxicities were hematalogic in nature, including neutropenia and leukopenia, and occurred at a slightly higher rate in arm A (75.8% and 44.9%, respectively) than in arm B (64.9% and 30.3%, respectively). Nausea (13.5%) and vomiting (11.8%) were the most common non-hematologic toxicities for arm A, whereas arm B had

Figure 2. Median survival time.

Figure 3. Median time to progression.

1073
Table 2. Analysis of predictors of overall survival (OS) and time to progression (TTP) Prognostic factor Univariate analysis (P-value) OS Treatment Gender Karnofsky performance status Prior weight loss Disease stage Disease measurability Age Race Previous radiation treatment Duration of disease before entering study Histology 0.086 0.028 <0.001 <0.001 0.686 0.754 0.127 0.621 0.470 0.257 0.325 TTP 0.867 0.489 <0.001 0.016 0.399 0.681 0.108 0.271 0.274 0.377 0.859 Multivariate analysis (P-value) OS 0.309 0.129 <0.001 0.015 0.784 0.412 0.281 0.802 0.661 0.429 0.886 TTP 0.969 0.552 <0.001 0.148 0.295 0.438 0.064 0.345 0.566 0.196 0.345 Table 3. Overall response to treatment Arm A Total no. of patients Overall response (CR + PR) CR PR SD PD Unevaluable or not assessed 179 26 (15%) 4 (2%) 22 (12%) 81 (45%) 57 (32%) 15 (8%) Arm B 190 43 (23%) 1 (0.5%) 42 (22%) 84 (44%) 50 (26%) 13 (7%) 0.061 0.203 0.014 0.917 0.253 P-valuea

CR, complete responders; PR, partial responders; SD, stable disease; PD, progressive disease. a Based on Fishers two-tail exact test.

a higher prevalence of paresthesia (10.1%), hypesthesia (5.3%) and neuropathy (5.3%). Sixteen patients (9.0%) in arm A and 19 patients (10.1%) in arm B experienced cardiovascular adverse events; one patient in each arm developed congestive heart failure. There were four deaths, due to causes other than disease progression, which occurred within 30 days of drug administration and were deemed either probably or possibly related to treatment. Three of these patients were on arm A and included one case each of suicide, cardiac arrest, and pneumonia. One patient on arm B died as a result of pulmonary arrest (attributed to chronic obstructive pulmonary disease, asthma and advanced lung cancer).

study. Subsequent to this study, ECOG 5592 demonstrated superior survival with cisplatin paclitaxel compared with cisplatin etoposide [11]. The experimental arm of the present study used carboplatin paclitaxel based on promising efcacy in two previous phase III trials with a lack of signicant cumulative myelosuppression [27, 28]. More recently, a metaanalysis of randomized clinical trials comparing cisplatin with carboplatin in patients with advanced NSCLC failed to demonstrate any survival difference between the two platinum compounds [29]. A longer median survival time was achieved with cisplatin paclitaxel than with carboplatin paclitaxel in a European study using the two regimens for the treatment of advanced NSCLC [30]. Despite this observation, the authors conclude that paclitaxel carboplatin is a viable regimen for treatment of this disease based on comparable response rate and ease of administration.
Table 4. Grade 3 and 4 hematologic and non-hematologic toxicities Adverse event Arm A Arm B

Downloaded from http://annonc.oxfordjournals.org/ at Library on February 9, 2012

Conclusions
In this phase III randomized clinical trial, the hypothesis that carboplatin paclitaxel was superior to cisplatin etoposide in the treatment of patients with advanced or metastatic NSCLC, based on statistical considerations, was not conrmed. Carboplatin-paclitaxel produced a higher objective response rate (23% compared with 15%) and a similar time to progression (121 days compared with 111 days) than cisplatin etoposide. However, neither difference was statistically signicant or sufcient to have a positive effect on survival. Despite these results, the overall benet in quality of life with carboplatin paclitaxel cannot be discounted. When this trial was being developed in 1995, cisplatin was widely regarded as the most active and most important agent in combination regimens for advanced or metastatic NSCLC [2225]. In a randomized phase III study conducted by ECOG to evaluate seven regimens in the treatment of metastatic NSCLC, the cisplatin etoposide combination had the highest proportion of 1-year survivors (25%) [26]; thus this combination was used as the standard arm in the present

Hematologic toxicity Neutropenia Leukopenia Anemia Thrombocytopenia Non-hematologic toxicity Nausea Vomiting Anorexia Dehydration Hyperglycemia Dyspnea Pneumonia Asthenia Pain Neuropathy Paresthesia Hypesthesia 24 (13.5%) 21 (11.8%) 9 (5.1%) 12 (6.7%) 4 (2.2%) 12 (6.7%) 8 (4.5%) 17 (9.6%) 11 (6.2%) 0 1 (0.6%) 1 (0.6%) 9 (4.8%) 8 (4.3%) 4 (2.1%) 7 (3.7%) 17 (9.0%) 20 (10.6%) 10 (5.3%) 26 (13.8%) 6 (3.2%) 10 (5.3%) 19 (10.1%) 10 (5.3%) 135 (75.8%) 80 (44.9%) 20 (11.2%) 14 (7.9%) 122 (64.9%) 57 (30.3%) 26 (13.8%) 15 (8.0%)

1074 There were several factors that may have confounded the outcome of this trial, including a gender imbalance in the number of women entered in the study, which favored the cisplatin etoposide arm. Female gender is a known positive contributor for improved survival, and analyses in this study showed that gender was a predictor of survival with men faring worse than women [31]. It is also notable that a substantially greater proportion of patients on arm A than patients on arm B received second- and/or third-line chemotherapy. Furthermore, 33% of patients on arm A received a taxane-containing regimen compared with only 12% on arm B. Thus treatment with taxanes in the second-line setting in the control arm may have had an impact on survival. This study was the rst randomized comparison utilizing the combination carboplatin paclitaxel as the investigational regimen. It contributed to the establishment of carboplatin paclitaxel as the preferred treatment regimen in the USA for advanced and metastatic NSCLC [32]. Subsequent phase III studies have been reported that have utilized a similar regimen (e.g. SWOG 9509, ECOG 1594) and have conrmed these results. The schedule of paclitaxel administration is a major determinant of the pattern of toxicity when administered in combination regimens [5]. Myelosuppression is the major toxicity when it is administered as a prolonged infusion [5, 10], while short infusions (1 3 h) are associated with a higher incidence of neuropathy [4 5]. The incidence of grade 2 and 3 neuropathy for the carboplatin paclitaxel arm in the present study was 10%. Carboplatin is better tolerated than cisplatin when either is combined with paclitaxel [5]. Similar observations were made in the present study as overall quality of life was signicantly better (P = 0.038) between baseline and cycle 3 on the carboplatin paclitaxel arm than on the control arm containing cisplatin. The SWOG 9509 trial, which compared carboplatin paclitaxel with cisplatin vinorelbine for untreated advanced NSCLC [4], also demonstrated equal efcacy on both arms, while safety and tolerability favored the carboplatin paclitaxel regimen. Fewer patients discontinued treatment because of toxicity on the carboplatin paclitaxel arm (15%) than on the cisplatin vinorelbine arm (28%). SWOG and ECOG subsequently adopted the carboplatin paclitaxel combination as the reference regimen for all of their ongoing randomized trials. More recently, CALGB have reported the results of their phase III trial 9730, which compared therapy with paclitaxel alone with the carboplatin paclitaxel combination for advanced NSCLC [7]. The response rates for the single-agent and combination arms were 17% and 29%, respectively _ (P < 0.0001). The median survival (8.8 months compared with 6.7 months, P = 0.0125, Wilcoxon) and progression-free survival (4.6 months compared with 2.5 months, P = 0.0001, Wilcoxon) also favored the carboplatin paclitaxel arm. These results conrm the use of the carboplatin paclitaxel doublet for the treatment of advanced NSCLC. In the present study, response rate, median survival and time to progression were similar in the two treatment arms, but tolerability seemed to favor the paclitaxel and carboplatin combination. Carboplatinpaclitaxel has become the foundation for adding novel targeted agents based on the favorable toxicity prole seen across multiple randomized studies.

Acknowledgements
The authors wish to acknowledge the contributions of the late Kasi Sridhar, formerly of the Sylvester Comprehensive Cancer Center, to the conduct of this study. This work was supported in part by a grant from the Bristol Myers Squibb Company.

References
1. Cancer Facts and Figures 2004. Atlanta, GA: American Cancer Society 2004. 2. Khuri FR, McKenna RJ Jr, Movsas B. Non-small cell lung cancer and mesothelioma. In Pazdur R, Coia LR, Hosins WJ, Wagman LD (eds): Cancer Management: A Multidisciplinary Approach, 4th edn. NY: PRR: Melville 2000; 91 124. 3. Schiller JH, Harrington D, Belani CP et al. Comparison of four chemotherapy regimens for advanced non-small cell lung cancer. N Engl J Med 2002; 346: 9298. 4. Kelly K, Crowley J, Bunn PA et al. Randomized phase III trial of paclitaxel plus carboplatin versus vinorelbine plus cisplatin in the treatment of patients with advanced non-small cell lung cancer: a Southwest Oncology Group Trial. J Clin Oncol 2001; 19: 32103218. 5. Gatzemeier U, Pawel J, Gottfried M et al. Phase III comparative study of high-dose cisplatin versus a combination of paclitaxel and cisplatin in patients with advanced non-small cell lung cancer. J Clin Oncol 2000; 18: 33903399. 6. Kubota K, Nishiwaki Y, Ohashi N et al. The four-arm cooperative study (FACS) for advanced non-small cell lung cancer (NSCLC). Proc Am Soc Clin Oncol 2004; 23: 616 abstr 7006. 7. Lilenbaum RC, Herndon J, List M et al. Single-agent (SA) versus combination chemotherapy (CC) in advanced non-small cell lung cancer (NSCLC): a CALGB randomized trial of efcacy, quality of life (QOL), and cost-effectiveness. Proc Am Soc Clin Oncol 2002; 21: 1a, abstr 2. 8. Delbaldo C, Michiels S, Syz N et al. Benets of adding a drug to a single-agent or a 2-agent chemotherapy regimen in advanced nonsmall cell lung cancer. JAMA 2004; 292: 470484. 9. Gandara DR, Crowley J, Livingston RB et al. Evaluation of cisplatin intensity in metastatic non-small cell lung cancer: a phase III study of the Southwest Oncology Group. J Clin Oncol 1993; 11: 873878. 10. Muscato JJ, Cirrincione C, Clamon G et al. Etoposide (VP-16) and cisplatin at maximum tolerated dose in non-small cell lung carcinoma: a Cancer and Leukemia Group B study. Lung Cancer 1995; 13: 285294. 11. Bonomi P, Kim KM, Fairclough D et al. Comparison of survival and quality of life in advanced non-small cell lung cancer patients treated with two dose levels of paclitaxel combined with cisplatin versus etoposide with cisplatin: results of an Eastern Cooperative Oncology Group trial. J Clin Oncol 2000; 18: 623631. 12. Breathnach OS, Freidlin B, Conley B et al. Twenty-two years of phase III trial for patients with advanced non-small cell lung cancer: sobering results. J Clin Oncol 2001; 19: 1734 1742. 13. Klastersky J, Sculier JP, Lacroix H et al. A randomized study comparing cisplatin or carboplatin with etoposide in patients with advanced non-small cell lung cancer: European Organization for Research

Downloaded from http://annonc.oxfordjournals.org/ at Library on February 9, 2012

1075
and Treatment of Cancer Protocol 07861. J Clin Oncol 1990; 8: 15561562. Jelic S, Radosavjelic D, Elezar E et al. Survival advantage for carboplatin 500 m/gm2 substituting cisplatin 120 mg/m2 in combination with vindesine and mitomycin C in patients with stage IIIB and IV squamous-cell bronchogenic carcinoma: a randomized phase III study in 221 patients. Lung Cancer 1997; 18: 1415. Helsing M, Bergman B. Chemotherapy with carboplatin and etoposide improves quality of life and survival. Lung Cancer 1997; 18: 8 9. Johnson DH, Paul DM, Hande KR et al. Paclitaxel plus carboplatin in advanced non-small cell lung cancer: a Phase II trial. J Clin Oncol 1996; 14: 2054 2060. Langer CJ, Leighton JC, Comis RL et al. Paclitaxel and carboplatin in combination in the treatment of advanced non-small cell lung cancer: a phase II toxicity, response, and survival. J Clin Oncol 1995; 13: 18601870. Belani CP, Kearns CM, Zuhowski EG et al. A phase I trial, including pharmacokinetic and pharmacodynamic correlations, of combination paclitaxel and carboplatin in patients with metastatic non-small cell lung cancer. J Clin Oncol 1999; 17: 676 684. Muggia FM, Vafai D, Natale R et al. Paclitaxel 3-hour infusion given alone and combined with carboplatin: preliminary results of doseescalation trials. Semin Oncol 1995; 22: 6366. Cella DF, Tulsky DS, Gray G et al. The functional assessment of cancer therapy scale: development and validation of the general measure. J Clin Oncol 1993; 11: 570 579. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457481. Gralla RJ, Kriss MG. Chemotherapy in non-small cell lung cancer: results of recent trials. Semin Oncol 1988; 15 (Suppl 4): 205. 23. Casper ES, Gralla RJ, Kelsen DP et al. Phase II study of high-dose cis-dichlorodiammine-platinum (II) in the treatment of non-small cell lung cancer. Cancer Treat Rep 1979; 63: 2107 2198. 24. De Jager R, Longeval I, Klastersky J. High-dose cisplatin with uid and mannitol-induced diuresis in advanced lung cancer: a phase II clinical trial of EORTC lung cancer working party (Belgium). Cancer Treat Rep 1980; 64: 13411346. 25. Bunn PA. The expanding role of cisplatin in the treatment of nonsmall cell lung cancer. Semin Oncol 1989; 16: 10 21. 26. Finkelstein DM, Ettinger DS, Ruckdeschel JC. Long-term survivors in metastatic non-small cell lung cancer: an Eastern Cooperative Oncology Group Study. J Clin Oncol 1986; 4: 702 709. 27. Israel VK, Zarestsky S, Natale RB. Phase I/II trial of combination carboplatin and Taxol in advanced non-small cell lung cancer (NSCLC). Proc Am Soc Clin Oncol 1994; 13: 351. 28. Vafai D, Israel V, Zaretsky S et al. Phase I/II trial of combination carboplatin and Taxol in advanced non-small cell lung cancer. Proc Am Soc Clin Oncol 1995; 14: 1067. 29. Katsuyuki H, Matsuo K, Ueoka H et al. Meta-analysis of randomized clinical trials comparing cisplatin to carboplatin in patients with advanced non-small cell lung cancer. J Clin Oncol 2004; 22: 38523859. 30. Rosell R, Gatzemeier U, Betticher DC et al. Phase III randomized trial comparing paclitaxel/carboplatin with paclitaxel/cisplatin in patients with advanced non-small cell lung cancer: a cooperative multinational trial. Ann Oncol 2002; 13: 15391549. 31. Albain KS, Crowley JJ, LeBlanc M et al. Survival determinants in extensive-stage non-small cell ling cancer: the Southwest Oncology Group experience. J Clin Oncol 1991; 9: 16181626. 32. Shyr Y, Choy H, Mohr. P et al. International pattern of practice survey: non-small cell lung cancer (NSCLC). Proc Am Soc Clin Oncol 1999; 478a: abstr 1843.

14.

15.

16.

17.

18.

Downloaded from http://annonc.oxfordjournals.org/ at Library on February 9, 2012

19.

20.

21. 22.