adjuvant_cisplatin_vinorelbine_completelyresected by Asociación Latinoamericana de Tórax ALAT

ORIGINAL ARTICLE

Adjuvant Cisplatin and Vinorelbine for Completely Resected Non-small Cell Lung Cancer Subgroup Analysis of the Lung Adjuvant Cisplatin Evaluation Jean-Yves Douillard, MD, PhD,* He´le`ne Tribodet, MSc,† Delphine Aubert, MSc,‡ Frances A. Shepherd, MD,§ Rafael Rosell, MD, PhD,㛳 Keyue Ding, PhD,¶ Anne-Sophie Veillard, MSc,† Lesley Seymour, PhD,¶ Thierry Le Chevalier, MD,# Stephen Spiro, MD,** Richard Stephens,†† Jean Pierre Pignon, MD, PhD,† and on behalf of the LACE Collaborative Group

Background: To evaluate the impact of adjuvant cisplatin-vinorelbine in completely resected non-small cell lung cancer and identify patients likely to benefit from this regimen in the Lung Adjuvant Cisplatin Evaluation (LACE) database. The overall LACE meta-analysis showed survival benefit with cisplatin-based adjuvant chemotherapy (5-year survival benefit of 5.4%, hazard ratio 关HR兴 0.89, p ⫽ 0.004). Subgroup analysis for the cisplatin-vinorelbine regimen was prespecified in the LACE statistical analysis plan. Patients randomized to cisplatin-vinorelbine or observation were the largest subgroup (41%) and the most homogeneous in terms of drug doses and eligibility. Patients and Methods: The LACE-vinorelbine cohort included trials evaluating cisplatin-vinorelbine versus observation. Overall survival was the primary end point. Other studies randomizing patients to other chemotherapy or observation (LACE-other) were also evaluated. Results: The LACE-vinorelbine cohort included 1888 patients from four studies (Adjuvant Navelbine International Trialist Association, Big Lung Trial, International Adjuvant Lung Cancer Trial, and National

*Department of Medical Oncology, Centre R Gauducheau, St-Herblain; †Biostatistics and Epidemiology Department, Institut Gustave Roussy, Paris; ‡Department of Statistics, Institut de Recherche Pierre Fabre, Boulogne, France; §Department of Thoracic Oncology, Princess Margaret Hospital, Toronto, Canada; 㛳Medical Oncology Department of Statistics, Hospital Germans Trias i Pujol, Barcelona, Spain; ¶Department of Biostatistics and Epidemiology, NCIC Clinical Trials Group, Queen’s University, Kingston, Ontario, Canada; #Department of Medicine, Institut Gustave-Roussy, Paris, France; **Department of Thoracic Medicine, The Middlesex Hospital; and ††Cancer Division, MRC Clinical Trials Unit, London, United Kingdom. Disclosure: F.A.S. was paid to speak on behalf of and consulted with Pierre Fabre Oncology and GlaxoSmithKline. L.S.’s employer received partial funding from GlaxoSmithKline to support JBR10. R.S. received an honorarium from Pierre Fabre Oncology to speak at a meeting. J.P.P. received unrestricted grants from Sanofi-Aventis and Pierre Fabre Oncology for this project. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s Web site (www.jto.org). Address for correspondence: Jean-Yves Douillard, MD, PhD, Department of Medical Oncology, Centre R. Gauducheau, Bd J. Monod, St. Herblain 44805, France. E-mail: jy-douillard@nantes.fnclcc.fr Copyright © 2010 by the International Association for the Study of Lung Cancer ISSN: 1556-0864/10/0502-0220

220

Cancer Institute of Canada Clinical Trials Group JBR.10). Baseline characteristics were similar to the LACE-other but had fewer patients with stage IA (2% versus 11%). Survival improvement at 5 years was 8.9% with cisplatin-vinorelbine versus observation (HR 0.80, 95% confidence interval: 0.70 – 0.91, p ⬍0.001). Stage was a significant predictor for survival (test for trend, p ⫽ 0.02; benefit at 5 years: 14.7% 关stage III兴, 11.6% 关stage II兴, and 1.8% 关stage I兴). Similar benefits were seen for disease-free survival (HR 0.75 关0.67– 0.85, p ⬍0.001兴, stage III 关HR 0.62, 0.50 – 0.76兴, stage II 关HR 0.69, 0.57– 0.83兴, and stage I 关HR 0.95, 0.76 –1.19兴). The overall result was statistically superior to LACEother (LACE other HR 0.95, 0.86 –1.05, interaction p ⫽ 0.04). Conclusion: In subgroup analyses, adjuvant cisplatin-vinorelbine provides a superior survival benefit and can be recommended in completely resected stages II and III non-small cell lung cancer. Key Words: Non-small cell lung cancer, Adjuvant cisplatin-vinorelbine, Meta-analysis. (J Thorac Oncol. 2010;5: 220–228)

ung cancer is the most frequent cause of cancer death worldwide.1 Patients with completely resected stage I nonsmall cell lung cancer (NSCLC) have a 5-year survival of ⬃60%, whereas it falls to ⬃10% for completely resected stage III.2,3 The first meta-analysis published in 19954 showed a nonsignificant 5% survival benefit at 5 years for patients receiving cisplatin combinations (hazard ratio 关HR兴 0.87, p ⫽ 0.08). In the following years, doublet combinations of cisplatin with new agents improved survival in patients with metastatic disease, but only one of these agents, vinorelbine, has been tested prospectively in combination with cisplatin in randomized adjuvant trials. Several meta-analyses including newer cisplatin-based regimens showed a small but consistent effect on survival.5–9 The more recent Lung Adjuvant Cisplatin Evaluation (LACE)10 is a meta-analysis based on individual patient data from the five largest adjuvant trials of cisplatin-based adjuvant chemotherapy published after 1995.11–17 LACE showed a HR of 0.89 for death (95% confidence interval 关CI兴: 0.82– 0.96, p ⫽ 0.004) corresponding to an absolute survival benefit of 5.4% at 5 years, in agreement with the 1995 meta-analysis.10

Journal of Thoracic Oncology • Volume 5, Number 2, February 2010

Patients randomized to cisplatin-vinorelbine or observation constituted the largest subgroup (1888 of 4584 patients, 41%) and the most homogeneous in terms of cisplatin dose. A prospectively planned analysis of chemotherapy effect according to regimen in LACE revealed a statistical trend (interaction p ⫽ 0.11) for superiority of cisplatin-vinorelbine compared with the other cisplatin-based doublets with vindesine, vinblastine, or etoposide and triplet combinations with mitomycin C, ifosfamide, or vinblastine. Although not originally planned, the indirect comparison of cisplatin-vinorelbine to other combinations (pooled) demonstrated a significant interaction (i.e., variation of the effect of chemotherapy compared with no treatment according to the type of chemotherapy, p ⫽ 0.04) in favor of cisplatin-vinorelbine. We report here the results of these subgroup analyses.

PATIENTS AND METHODS Inclusion and Exclusion Criteria The LACE meta-analysis has been described previously.10 Briefly, eligible trials (a) were randomized trials performed after the NSCLC meta-analysis published in 1995, (b) included only patients with completely resected NSCLC, and (c) compared cisplatin-based chemotherapy versus no chemotherapy. Trials were excluded if they (a) used concomitant radiochemotherapy or preoperative chemotherapy; (b) included incompletely resected patients; and (c) were included in the 1995 NSCLC meta-analysis. LACE prospectively planned evaluation of treatment effect according to types of chemotherapy received (cisplatin-vinorelbine, other cisplatin-based doublets, or triplets). Trials included in LACE comparing cisplatin-vinorelbine with observation were included in the LACE-vinorelbine cohort and included Adjuvant Navelbine International Trialist Association (ANITA), Big Lung Trial (BLT), International Adjuvant Lung Cancer Trial (IALT), and JBR.10. All other trials or strata within trials were pooled into a “LACE-other” subgroup and were compared with the LACE-vinorelbine subgroup.

Analyses For LACE-vinorelbine, the primary objective was overall survival. Secondary objectives were (a) interaction between predefined baseline covariates and the effect of cisplatin-vinorelbine; (b) cause-specific mortality; (c) disease-free survival (DFS); and (d) chemotherapy-related severe toxicity. Each trial was analyzed individually, and all data were pooled for the overall analysis. In addition, the results for LACE-vinorelbine were compared with those for the LACE-other subgroup.

Population All patients, including ineligible patients, were analyzed in the group to which they were randomized, independent of the chemotherapy actually received (intent-to-treat analysis).

Data Checks As described previously,10 individual patient data were retrieved for each study. Systematic data checking and reanalysis were performed by the meta-analysis team at Institut Gustave Roussy.

Adjuvant Cisplatin and Vinorelbine

Statistical Analyses Times to events were measured from the date of randomization. Survival was the time until death of any cause or last known alive date. DFS was the time until first recurrence or death of any cause or last disease evaluation date. Second cancers were not considered as an event, and follow-up was not censored after a second cancer. Median follow-up was computed using the reverse Kaplan-Meier method.18 Survival and DFS per treatment group were estimated using Peto curves19 and their comparison summarized by HRs stratified by trial with their 95% CIs. The analysis of survival was based on a log-rank test stratified by trial. Chi-square tests were used to study heterogeneity among trials and groups. I2 statistic was also used.20 Test of heterogeneity used to compare the HR of groups of trials is also called test for interaction. Cause-specific mortality (lung cancer, 关death from other cause after a lung cancer recurrence and death from unknown cause兴 versus nonlung cancer)10 was described by treatment arm and compared using the log-rank test stratified by trial. As proposed by Peto,21 data were censored at the first recurrence for nonlung cancer mortality. The log-rank statistic for nonlung cancer mortality was subtracted from the log-rank statistic for mortality from all causes to obtain the log-rank statistic of lung cancer mortality.21 Variations in treatment effect were described using forest plots of HRs. Multivariate Cox regression analysis was performed to estimate chemotherapy effect adjusted for prespecified factors of age, type of surgery, histology, and stage.

Treatment Compliance and Toxicity The actual doses of cisplatin and vinorelbine (available for ANITA, JBR.10, and BLT) and time between surgery and start of chemotherapy were described. Grade 3, 4, or 5 toxicity was collected in all trials except IALT, which collected only grades 4 and 5 toxicity. Toxicities were not available from the BLT trial.

LACE Organization A Steering Committee for LACE was constituted with one clinician and one statistician from each trial, together with the two statisticians from the data center in charge of the pooled analysis (see Appendix). This committee validated the statistical analysis plan, selected the data to be included in the publication, and contributed to the writing of the manuscript. All analyses were performed at the Institut Gustave-Roussy.

Role of the Funding Source The sponsors of the LACE meta-analyses and these subgroup analyses had no role in the study design, data collection, data analysis, data interpretation, or the writing of the report. The corresponding author had full access to all data in the study and had final responsibility for the decision to submit for publication.

RESULTS Four studies from LACE were eligible for the LACEvinorelbine subgroup analyses with a total of 1888 patients (Tables 1 and 2). Two studies evaluated the cisplatin-vinorelbine combination exclusively (ANITA and JBR.10). In the other two (IALT and BLT), the investigator could choose among several regimens; each center used only one regimen,

221

Journal of Thoracic Oncology • Volume 5, Number 2, February 2010

Douillard et al.

TABLE 1. Overview of Studies Included in the LACE-Vinorelbine and LACE-Other Subgroups No. of Patients Evaluated and Follow-Up LACE-Vinorelbine

ANITA IALT JBR.10 BLT ALPI Total

LACE, N

Randomized to Cisplatin Vinorelbine

840 1867 482 307 1088 4584

840 500 482 66 0 1888

407 248 242 37 0 934

LACE-Other Follow-Upa (yr) 5.9 4.1 5.2 4.9 — 5.2

Randomized to Other Chemotherapy

Follow-Upa (yr)

0 1367 0 241 1088 2696

0 684 0 115 548 1347

— 4.9 — 5.2 5.4 5.2

a Median. LACE, Lung Adjuvant Cisplatin Evaluation; ANITA, Adjuvant Navelbine International Trialists Association; IALT, International Adjuvant Lung Trial; BLT, Big Lung Trial.

TABLE 2. Patients Characteristics (%) LACE-Vinorelbine Characteristics Age (yr) Median Range Sex (%) Male Female Unknown Pathological TNM stage (%) IA IB II III Unknown Type of surgery (%) Pneumonectomy Lobectomy/other Unknown Performance status (%) 0 1 2 Unknown Histologic type (%) Squamous cell Adenocarcinoma Other Area of the world (%) Europe North America Other

LACE-Other

Chemotherapy (N ⴝ 934)

Observation (N ⴝ 954)

Chemotherapy (N ⴝ 1347)

Observation (N ⴝ 1349)

59 27–81

59 18–77

60 27–83

61 30–79

79 21 ⬍1

80 20 ⬍1

81 19 0

3 34 37 26 0

2 33 39 26 0

12 26 33 29 ⬍1

11 28 33 28 ⬍1

33 62 4

31 65 3

27 64 9

28 63 9

49 47 3 1

50 47 2 1

31 23 5 41

31 24 5 40

48 42 10

49 41 10

49 39 12

48 38 14

68 26 6

91 1 8

LACE, Lung Adjuvant Cisplatin Evaluation; TNM, tumor, node, metastasis.

and patients were randomized by center. Median follow-up was 5.2 years. The LACE-other subgroup consisted of 2696 patients of whom 1347 received nonvinorelbine but cisplatinbased chemotherapy (Tables 1 and 2).

222

Patients aged ⱖ70 years were slightly more common in LACE-vinorelbine (10% versus 8%), and stage IA were more numerous in LACE-other (11%) than in LACE-vinorelbine (2%). In both subgroups approximately two thirds of patients

Journal of Thoracic Oncology â&#x20AC;˘ Volume 5, Number 2, February 2010

Adjuvant Cisplatin and Vinorelbine

FIGURE 1. A, Forest plot of hazard ratios of overall survival and disease-free survival with cisplatin-vinorelbine versus observation (control). B, Forest plot of hazard ratios of overall survival and disease-free survival with cisplatin-vinorelbine versus observation (control) according to stage.

presented with stages II and III. No firm conclusions can be drawn regarding the distribution of performance status among subgroups because the data were missing for approximately 40% of patients in LACE-other. The majority of patients in LACE were evaluated in Europe, and most patients from North America were in the LACE-vinorelbine subgroup (26% versus 1% in LACE-other) as corresponding mostly to the JBR.10

study. Other baseline variables were similarly distributed between LACE-vinorelbine and LACE-other.

Effect of Cisplatin-Vinorelbine on Survival For the LACE-vinorelbine subgroup, 45.3% of cisplatinvinorelbine patients had died (versus 53.7% observation) with a median follow-up of 5.2 years. Patients randomized to adjuvant

223

Douillard et al.

Journal of Thoracic Oncology • Volume 5, Number 2, February 2010

FIGURE 3. Overall survival curves by stage for the cisplatinvinorelbine versus the observation (no chemotherapy) groups.

years and ⫹11.6% at 5 years) and stage III (⫹11.5% at 3 years and ⫹14.7% at 5 years) (Figure 3). The effect on survival was significantly higher (test of interaction p ⫽ 0.04) in LACE-vinorelbine (HR 0.80, mentioned above) compared with LACE-other (HR 0.95, 95% CI: 0.86 – 1.05, p ⫽ 0.33). There was no significant interaction with stage in LACE-other. In LACE-vinorelbine, no significant interaction was observed between chemotherapy and covariates of sex, performance status, age, histology, type of surgery, and planned radiotherapy, and chemotherapy effect was not different with planned doses of cisplatin or vinorelbine.

Effect of Cisplatin-Vinorelbine on DFS FIGURE 2. A, Survival curves for the cisplatin-vinorelbine and the observation (no chemotherapy) groups. Absolute benefit (⫾1 standard deviation). B, Lung cancer-specific and nonlung cancer-related survival for the cisplatin-vinorelbine and the observation (no chemotherapy) groups.

cisplatin-vinorelbine had a 20% reduction in the risk of death (HR 0.80, 95% CI: 0.70 – 0.91, p ⬍0.001). The test of heterogeneity among trials was not significant (p ⫽ 0.52) (Figure 1A). The absolute benefit was 6.8% at 3 years (64.3% versus 57.5%) and 8.9% at 5 years (55.1% versus 46.2%) (Figure 2A). In the multivariate Cox model chemotherapy (HR 0.76), squamous histology (HR 0.69), pneumonectomy (HR 1.33), age (51– 60 years, HR 1.26, 61– 69, HR 1.40, ⱖ70, HR 1.90), and stage (stage II, HR 1.72; stage III, HR 2.61) were significant prognostic factors. Sex showed a trend for better survival in females (p ⫽ 0.07; see Web Table 1, Supplemental Digital Content 1, http://links.lww.com/JTO/A13). There was a significant interaction of cisplatin-vinorelbine effect and stage (test for trend p ⫽ 0.02) (Figure 1B): patients with stage I had no benefit from chemotherapy (⫺1.2% at 3 years and ⫹1.8% at 5 years), whereas the effect was significant for stage II (⫹10.5% at 3

224

A significant difference in DFS was observed for patients receiving cisplatin-vinorelbine with a HR 0.75 (CI: 0.67– 0.85, p ⬍ 0.001) (Figure 1A). The absolute benefit at 3 years was 10.0% and at 5 years 9.2%. A clear relation between chemotherapy effect and stage was also observed on DFS for LACE-vinorelbine, the effect was significant for stage II: HR 0.69 (CI: 0.57– 0.83) and for stage III: HR 0.62 (CI: 0.50 – 0.76) but not for stage I: HR 0.95 (CI: 0.76 –1.19), test for trend, p ⫽ 0.008 (Figure 1B). The effect on DFS was significantly higher (test for interaction p ⫽ 0.02) in LACE-vinorelbine compared with LACE-other (HR 0.90 关0.82– 0.99], p ⫽ 0.04). For LACEother, the benefit at 5 years was 3.8%. No interaction with stage was demonstrated (data not shown).

Effect of Cisplatin-Vinorelbine on Compliance and Toxicity In the cisplatin-vinorelbine subgroup, 7% of patients did not receive chemotherapy and 0.6% of patients in the observation group did. The majority of patients were planned to receive a total cisplatin dose ⱖ400 mg/m2 (Table 3). The median dose delivered was 303 mg/m2 for cisplatin and 236 mg/m2 for vinorelbine, and the median relative dose intensities were 95% and 69%, respectively. There were 371 patients (43%) who did not complete chemotherapy because of

— — — —

69 (48–100) 69 (50–101) NA NA 69 (49–100)

Adjuvant Cisplatin and Vinorelbine

TABLE 4. Incidence of Grades 3, 4, and 5 Toxicity and Therapy-Related Deaths in LACE Other and LACE-Vinorelbine Vinorelbine

Calculated as the ratio between the total dose received per patient and the number of received cycles expressed in weeks. Ratio between the actual dose intensity and the projected dose intensity. Actual duration of treatment not available for the BLT and the IALT and received dose of vinorelbine not available for the IALT. LACE, Lung Adjuvant Cisplatin Evaluation; ANITA, Adjuvant Navelbine International Trialists Association; IALT, International Adjuvant Lung Trial; BLT, Big Lung Trial; Na, not applicable. c

— — — — — — — — 30 (16–34) NA NA 30 (16–34) 39 52 8 100 300 300–400 150–240 NA

291 (95–308) 300 (100–400) 150 (50–160) 291 (95–308)

91 (49–103) NA NA 91 (49–103)

— — — —

21 (14–30) 17 (13–25) NA NA 20 (13–30) 24 (16–25) 23 (15–25) NA NA 24 (15–25) 42 27 27 4 100 400 400 300–400 240 NA

304 (99–402) 302 (50–406) 300 (100–400) 180 (80–240) 303 (59–403)

96 (62–101) 92 (62–101) NA NA 95 (62–101)

480 400 270–390 180 NA

42 27 27 4 100

270 (30–473) 202 (25–350) NA 150 (30–180) 236 (30–450)

Toxicity

LACE-vinorelbine ANITA JBR.10 IALTc BLTc Overall LACE-other ALPI IALTc BLTc Overall

Study

Median Dose Median Dose Median Relative Planned Total Median Dose Median Dose Relative Dose Received Intensitya Dose Intensityb Received Intensitya Intensityb Planned Total Vinorelbine Percentage (90th Percentile) (90th Percentile) (90th Percentile) Cisplatin Dose Percentage (90th Percentile) (90th Percentile) (90th Percentile) Dose of Patients of Patients (mg/m2) (mg/m2) (mg/m2) (mg/m2/wk) (%) (mg/m2) (mg/m2/wk) (%)

TABLE 3. Planned and Received Doses of Cisplatin and Vinorelbine

Journal of Thoracic Oncology • Volume 5, Number 2, February 2010

N Neutropenia Febrile neutropenia Thrombocytopenia Neuropathy Renal impairment Nausea and vomiting Infection Constipation

Other

Grades 3–5 (%)a

Grades 4–5 (%)b

Grades 3–5 (%)a

Grades 4–5 (%)b

599 80 9 3 3 1 20 8 4

837 55 0 1 0 0 3 2 1

460 27 2 5 3 2 16 0 1

1077 10 0 2 0 0 3 0 0

All studies except BLT and IALT. All studies except BLT. IALT, International Adjuvant Lung Trial; BLT, Big Lung Trial. b

patient refusal (42%), toxicity (30%), or unknown reasons (10%). The median time between surgery and initiation of chemotherapy was 41 days with 6% of patients initiating chemotherapy more than 2 months after surgery. Twenty percent of patients in ANITA, 17% in JBR.10, 11% in IALT, and 41% in BLT received only one cycle of chemotherapy. For LACE-other, 9% of patients randomized to chemotherapy did not receive it, whereas 2% randomized to observation did. Planned doses of cisplatin were lower than for the LACE-vinorelbine subgroup (Table 3). The incidence of grade ⱖ3 toxicity was collected in two of the trials in LACE-vinorelbine (ANITA and JBR.10), whereas only grades 4 and 5 toxicity were in IALT. ALPI (LACE-other) collected only toxicity ⱖ grade 3. No data were available for BLT. Neutropenia ⱖ grade 4 was observed in 55% in LACE-vinorelbine and 10% of patients in LACE-other, but life threatening or fatal febrile neutropenia (ⱖ grade 4) was 0 in LACE-vinorelbine and ⬍1% in LACE-other. There was no difference in peripheral neurotoxicity grade ⱖ3. Overall, grades 4 ⫹ 5 toxicities were more frequent in LACE-vinorelbine (59% versus 16% in LACE-other, Table 4).

Effect of Cisplatin-Vinorelbine on CancerRelated and Noncancer-Related Deaths As shown in Table 5, there were 13 (1.4%) reported drug-related deaths in the cisplatin-vinorelbine arm, mainly because of septic shock. The drug-related death rate for LACE-other was 0.4. When considering the overall follow-up period, in the LACE-vinorelbine subgroup, noncancer-related death was not significantly different between the two arms (HR 1.31, 关0.91–1.88兴 p ⫽ 0.15), whereas there was a highly statistically significant difference in lung cancer-specific death (HR 0.74 关0.65– 0.85兴, p ⬍0.001) (Figure 2B). For LACE-other, there was a significant increase (p ⫽ 0.01) in nonlung cancer deaths (HR 1.40 关1.08 –1.82兴). A significant interaction over time was observed both for LACE-vinorelbine (p ⫽ 0.002) and LACE-other (p ⫽

225

Journal of Thoracic Oncology • Volume 5, Number 2, February 2010

Douillard et al.

TABLE 5. Causes of Death LACE-Vinorelbine Cause of Death

LACE-Other

Chemotherapy (N ⴝ 934)

Observation (N ⴝ 954)

Chemotherapy (N ⴝ 1347)

Observation (N ⴝ 1349)

423 (45.3) 354 (83.7) 69 (16.3) 13 (1.4)

512 (53.7) 464 (90.6) 48 (9.4)

713 (52.9) 579 (81.2) 134 (18.8) 6 (0.4)

742 (55.0) 651 (87.7) 91 (12.3)

Total no. of deathsa No. of lung cancer deaths (%)b No. of nonlung cancer deaths (%)b No. of drug- related deaths (%)c a

Percentage of the total number of patient randomized in the corresponding arm. Percentage of the total number of deaths in the corresponding arm. c Included in the nonlung cancer deaths. LACE, lung adjuvant cisplatin evaluation. b

0.03) with a significant increase of nonlung cancer death in the first 6 months in both groups: 2.92 (1.54 –5.51, p ⬍0.001) for LACE-vinorelbine (29 versus 9 deaths) and 2.17 (1.33– 3.53, p ⫽ 0.002) for LACE-other (45 versus 20 deaths). In LACE-vinorelbine, among the 29 early nonlung cancer deaths observed in the cisplatin-vinorelbine arm, 13 (45%) were drug related and 12 (41%) related to cardiovascular or pulmonary disease. There were six cardiovascular or pulmonary deaths in the observation arm of nine early deaths. In LACE-other, there were six (13%) drug-related deaths and 28 (62%) cardiovascular or pulmonary deaths among the 45 early deaths in the chemotherapy arm. In the corresponding observation arm, there were 15 cardiovascular or pulmonary deaths of 20 early deaths.

Use of Adjuvant Postoperative Radiation Therapy The impact of postoperative radiation therapy (PORT) was not assessable in either the LACE-vinorelbine or the LACE-other analysis, because the use and regimens of PORT varied among trials and was not mandatory or randomized. PORT was planned for a similar percentage of patients in LACE-vinorelbine and LACE-other: 33% and 31% for the observation groups and 30% and 32% in the chemotherapy groups, respectively.

DISCUSSION We have evaluated the effect of cisplatin-vinorelbine on survival in patients with completely resected stages I to III NSCLC within the LACE database. Baseline characteristics of the patients in the trials evaluating this combination were similar to those included in the other LACE trials, except for a slightly lower incidence of stage IA and a predominance of patients from North America. Survival of patients receiving cisplatin-vinorelbine was statistically superior to observation (absolute benefit 8.9% at 5 years) and was also statistically superior to the benefit observed in LACE-other (overall test for interaction p ⫽ 004). Baseline characteristics were similar and could not account for these results. In LACE-vinorelbine, stage was a strong predictor. Patients with stage IIIA benefited the most, closely followed by patients with stage II. Too few patients with stage IA were treated with cisplatin-vinorelbine to make a definitive interpretation,

226

whereas in stage IB (approximately 34% of the total group), cisplatin-vinorelbine did not affect outcomes, when compared with observation. In LACE-other, there was no variation of treatment effect with stage, although a significant effect was observed in stage IB for DFS (HR 0.75, 95% CI: 0.60 – 0.93) and a borderline effect for survival (HR 0.81, 95% CI: 0.64 –1.02). These results are in agreement with the Cancer and Leukemia Group B 9633 trial using paclitaxel-carboplatin in stage IB showing a significant benefit on failure-free survival but no significant difference on overall survival.22 However, in the updated NSCLC meta-analysis of 30 trials and 8147 patients, no interaction with stage was observed.23 Thus, the absence of an effect of chemotherapy in stage I with cisplatin-vinorelbine reported here should be interpreted with caution. Although the planned cisplatin dose was higher in LACE-vinorelbine than LACE-other, the median dose of cisplatin delivered in LACE-vinorelbine was lower than intended and close to the median dose delivered with other cisplatin doublets but higher than that delivered with cisplatin triplets in LACE.10 Toxicity observed with the cisplatin-vinorelbine regimen was as expected for the combination, and was higher in LACE-vinorelbine than in LACE-other, in particular for grade 4 neutropenia (80% grade ⱖ3). The incidence of febrile neutropenia grade 4 was similar in LACE-vinorelbine and LACE-other. Overall reported grade ⱖ3 toxicity was observed in 90% of patients receiving cisplatin-vinorelbine versus 49% in LACE-other. There seemed to be a small increase in chemotherapy-related deaths in LACE-vinorelbine compared with LACE-other (1.4% versus 0.4%). It should be noted that these data are difficult to interpret because ANITA and JBR.10 collected laboratory and toxicity data in a more rigorous fashion than other trials. A quantification of differences in tolerability between cisplatin-vinorelbine and other regimens evaluated in LACE was not an objective of the meta-analysis and cannot be addressed adequately. However, in the IALT study, vinorelbine was associated with a higher grades 4 to 5 toxicity rate than other associated drugs with an odds ratio adjusted on region of the world and received dose of cisplatin at first cycle of 2.5 (1.7–3.7) for vinorelbine versus other (A. Dunant et al., personal communication).

Journal of Thoracic Oncology • Volume 5, Number 2, February 2010

The excess of early nonlung cancer deaths described in LACE10 was found in both LACE-vinorelbine and LACEother. Possible explanations for this unexpected effect have been discussed previously and were mainly related to chemotherapy toxicity and lethal cardiopulmonary events.10 Addition of PORT to adjuvant chemotherapy with cisplatin-vinorelbine could not be assessed. A descriptive analysis in the ANITA trial suggested a possible benefit in patients with N2 disease, whereas patients with N1 disease experienced a potential detrimental effect.24 The role of PORT in N2 disease is presently being evaluated in the randomized Lung Adjuvant Radiation Therapy trial. Cisplatin-vinorelbine is the only third generation drug in combination with cisplatin tested so far to demonstrate a significant survival benefit with a reduction in the risk of death of 20%

APPENDIX.

Adjuvant Cisplatin and Vinorelbine

versus observation, despite an increase in toxicity. Indirect comparison with other cisplatin-based combinations (3– 8% reduction) is in favor of cisplatin-vinorelbine. Based on these results, the evidence supports the selection of cisplatin-vinorelbine as the adjuvant chemotherapy of choice for patients with completely resected stages II and IIIA.

ACKNOWLEDGMENTS Supported by Institut Gustave-Roussy, Programme Hospitalier de Recherche Clinique, Ligue Nationale Contre le Cancer, Sanofi-Aventis (unrestricted grants), and Pierre Fabre Oncology (unrestricted grants). The contribution of NCIC CTG is supported by the Canadian Cancer Society. JBR.10 was an Intergroup study, led by NCIC CTG and to which CALGB, NCCTG, and SWOG contributed.

Members of the Steering Committee

Trials

Clinician

Statistician

ANITA

Jean-Yves Douillard, Department of Medical Oncology, Centre R Gauducheau, Nantes Thierry Le Chevalier, Department of Medicine, Institut Gustave-Roussy, Paris Frances Shepherd, Princess Margaret Hospital, Toronto Stephen Spiro, Department of Thoracic Medicine, The Middlesex Hospital, London

Patricia HIS, Institut de Recherche Pierre Fabre Boulogne, France Jean-Pierre Pignon, Biostatistics and Epidemiology Department, Institut Gustave-Roussy, Paris Keyue Ding, NCIC Clinical Trials Group, Queen’s University, Kingston Richard Stephens, Cancer Division, MRC Clinical trials Unit, London Hélène Tribodet, Biostatistics and Epidemiology Department, Institut Gustave-Roussy, Paris

IALT JBR.10 BLT LACE Secretariat

REFERENCES 1. Ferlay J, Bray F, Pisani P, et al. GLOBOCAN 2000: Cancer Incidence, Mortality and Prevalence Worldwide. Version 1.0. Lyon: IARC Press, 2001. 2. Ponn RB, Lo Cicero J III, Daly BD. Surgical treatment of non-small cell lung cancer. In TW Shields, J Lo Cicero III, R Ponn, et al. (Eds.), General Thoracic Surgery, 6th Ed. Philadelphia: Lippincott Williams & Wilkins, 2005. Pp. 1548 –1587. 3. Luzzi L, Voltolini L, Campione A, et al. Pneumonectomy vs lobectomy in the treatment of pathologic N1 NSCLC: could the type of surgical resection dictate survival? J Cardiovasc Surg (Torino) 2003;44:119 – 123. 4. Chemotherapy in non-small cell lung cancer: a meta-analysis using updated data on individual patients from 52 randomized clinical trials. Non-small Cell Lung Cancer Collaborative Group. BMJ 1995;311:899 – 909. 5. George S, Schell MJ, Detterbeck FC, et al. Adjuvant chemotherapy for resected non-small-cell carcinoma of the lung: why we still don’t know. Oncologist 1998;3:35– 44. 6. Hotta K, Matsuo K, Ueoka H, et al. Role of adjuvant chemotherapy in patients with resected non-small-cell lung cancer: reappraisal with a meta-analysis of randomized controlled trials. J Clin Oncol 2004;22: 3860 –3867. 7. Sedrakyan A, van Der Meulen J, O’Byrne K, et al. Postoperative chemotherapy for non-small cell lung cancer: a systematic review and meta-analysis. J Thorac Cardiovasc Surg 2004;128:414 – 419. 8. Alam N, Darling G, Shepherd FA, et al; The Lung Cancer Disease Site Group of Cancer Care Ontario’s Program in Evidence-Based Care.

10.

11.

12.

13.

14.

15.

Postoperative chemotherapy in nonsmall cell lung cancer: a systematic review. Ann Thorac Surg 2006;81:1926 –1936. Berghmans T, Paesmans M, Meert A, et al. Survival improvement in resectable non-small cell lung cancer with (neo)adjuvant chemotherapy: results of a meta-analysis of the literature. Lung cancer 2005; 49:13–23. Pignon JP, Tribodet H, Scagliotti GV, et al; LACE Collaborative Group. Lung Adjuvant Cisplatin Evaluation (LACE): a pooled analysis by the LACE Collaborative Group. J Clin Oncol 2008;26:3552–3559. Winton T, Livingston R, Johnson D, et al; National Cancer Institute of Canada Clinical Trials Group; National Cancer Institute of the United States Intergroup JBR.10 Trial Investigators. Vinorelbine plus cisplatin vs. observation in resected non-small-cell lung cancer. N Engl J Med 2005;352:2589 –2597. Waller D, Peake S, Milroy G, et al. Chemotherapy for patients with non-small cell lung cancer: the surgical setting of the big lung trial. Eur J Cardiothorac Surg 2004;26:173–182. Dunant A, Pignon J, Le Chevalier T; on behalf of the International Adjuvant Lung Cancer Trial Collaborative Group. Adjuvant chemotherapy for non-small cell lung cancer: contribution of the International Adjuvant Lung Trial. Clin Cancer Res 2005;11(Suppl 13): 5017s–5021s. Arriagada R, Bergman B, Dunant A, et al; International Adjuvant Lung Trial Collaborative Group. Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small-cell lung cancer. N Engl J Med 2004;350:351–360. Douillard JY, Rosell R, De Lena M, et al. Adjuvant vinorelbine plus cisplatin versus observation in patients with completely resected stage IB–IIIA non-small cell lung cancer (Adjuvant Navelbine International

227

Douillard et al.

16.

17.

18. 19.

20.

Journal of Thoracic Oncology • Volume 5, Number 2, February 2010

Trialist Association 关ANITA兴): a randomised controlled trial. Lancet Oncol 2006;7:719 –727. Scagliotti G, Fossati R, Torri V, et al; Adjuvant Lung Project Italy/ European Organisation for Research Treatment of Cancer-Lung Cancer Cooperative Group Investigators. Randomized study of adjuvant chemotherapy for completly resected stage I, II, or IIIA non-small-cell lung cancer. J Natl Cancer Inst 2003;95:1453–1461. Scagliotti G; on behalf of the Adjuvant Lung Cancer Project Italy/ European Organization for Research and Treatment of Cancer Lung Cancer Cooperative Group. The ALPI Trial: the Italian/European experience with adjuvant chemotherapy in resectable non-small lung cancer. Clin Cancer Res 2005;11(Suppl 13):5011s–5015s. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control clin trials 1996;16:343–346. Early Breast Cancer Trialists’ Collaborative Group. Effects of radiotherapy and surgery in early breast cancer. An overview of randomized trials. Early Breast Cancer Trialists’ Collaborative Group. N Engl J Med 1995;333:1444 –1455. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539 –1558.

228

21. Systemic treatment of early breast cancer by hormonal, cytotoxic or immune therapy. 133 randomised trials involving 31,000 recurrences and 24,000 death among 75,000 women. Early Breast Cancer Trialists’ Collaborative Group. Lancet 1992;339:1–15. 22. Strauss GM, Herndon JE, Maddaus MA, et al. Adjuvant paclitaxel plus carboplatin compared with observation in stage IB non-small-cell lung cancer: CALGB 9633 with the Cancer and Leukemia Group B, Radiation Therapy Oncology Group, and North Central Cancer Treatment Group Study Groups. J Clin Oncol 2008;26:5043–5051. 23. Stewart LA, Burdett S, Tierney JF, et al. Surgery and adjuvant chemotherapy (CT) versus surgery alone in non-small cell lung cancer (NSCLC): a meta-analysis using individual patient data (IPD) from randomised clinical trials (RCTs). J Clin Oncol 2007;25(Suppl 397) (abstract 7552). 24. Douillard JY, Rosell R, De Lena M, et al; Adjuvant Navelbine International Trialist Association. Impact of post-operative radiation therapy on survival in patients with complete resection and stage I, II, IIIA nonsmall-cell lung cancer treated with adjuvant chemotherapy: the Adjuvant Navelbine International Trialist Association (ANITA) randomized trial. Int J Radiat Oncol Biol Phys 2008;72:695–701.