The treatment outcome in patients with advanced nonsmall

ORIGINAL ARTICLE Relationship between Response and Survival in More Than 50,000 Patients with Advanced Non-small Cell Lung Cancer Treated with Systemic Chemotherapy in 143 Phase III Trials Katsuyuki Hotta, MD, PhD,* Yoshiro Fujiwara, MD,* Katsuyuki Kiura, MD, PhD,* Nagio Takigawa, MD, PhD,* Masahiro Tabata, MD, PhD,* Hiroshi Ueoka, MD, PhD, and Mitsune Tanimoto, MD, PhD* Background: The association between the objective response to chemotherapy and survival has not yet been fully evaluated using large cohorts in advanced non-small cell lung cancer. Methods: We searched for phase III trials conducted between 1991 and 2006 to investigate the role of systemic chemotherapy for advanced non-small cell lung cancer. Associations were tested by multiple regression analysis. Results: Of the 1255 trials screened, 143 met our criteria, involving 50,569 patients with 309 chemotherapy regimens. In the first-line setting, the median intention-to-treat objective response rate (ORR) and disease control rate (DCR) were 26.4% and 62.5%, respectively (43,551 randomized patients; 290 trials). The median of the median survival time (MST) was 8.5 months in the first-line setting, and both the ORR and DCR were significantly associated with the MST in the multivariate analysis (regression coefficient 0.0788 [p 0.0001] and 0.0794 [p 0.0001], respectively). Subgroup analysis showed no correlation between the ORR and MST in patients receiving chemotherapy containing molecular-targeted agents (p 0.3817). In the second-line or later setting, the median ORR was only 6.8%, whereas the median DCR was 42.4% (4318 randomized patients; 19 trials). The median MST (6.6 months) was not associated with the ORR (p 0.6992), but was associated with the DCR (p 0.0129), despite the small sample size. Conclusions: We found that survival was associated with both the ORR and DCR in the first-line setting, although it should be *Department of Respiratory Medicine, Okayama University Hospital, Okayama, Japan; Department of Respiratory Medicine, National Hospital Organization Sanyo Hospital, Ube, Japan. Katsuyuki Hotta took the lead role, and all other authors actively participated in writing the manuscript. Katsuyuki Hotta and Yoshiro Fujiwara participated in the study design and acquisition of data. Disclosure: The authors declare no conflict of interest. A supplementary Appendix is available via the ArticlePlus feature at www.jto.org. Please go to the April issue and click on the ArticlePlus link posted with the article in the Table of Contents to view this material. Address for correspondence: Katsuyuki Hotta, MD, PhD, Department of Respiratory Medicine, Okayama University Hospital, 2-5-1, Shikata-cho, Okayama, 700-8558, Japan. E-mail: khotta@md.okayama-u.ac.jp Copyright 2007 by the International Association for the Study of Lung Cancer ISSN: 1556-0864/07/0205-0402 402 interpreted cautiously because of the abstracted data based analysis. Regarding chemotherapy regimens containing moleculartargeted agents and salvage chemotherapy regimens, further assessments are warranted to clarify the association between the parameters. Key Words: Lung cancer, Phase III trial, Objective response rate, Disease control rate, Survival. (J Thorac Oncol. 2007;2: 402 407) The treatment outcome in patients with advanced nonsmall cell lung cancer (NSCLC) has recently improved with advances in systemic chemotherapy; newer cytotoxic agents have been clinically introduced since the early 1990s and were proven, in terms of the overall patient survival, to be superior to the older agents in the first-line setting. 1 Some of these agents have also been shown to produce survival prolongation in the second-line setting. 2,3 In addition, molecular-targeted agents, acting on the specific molecule that contributes to tumor proliferation and survival, have also been developed; they were initially designed to be cytostatic agents that are not so much expected to shrink tumors, 4,5 although some of them have already been shown to yield survival advantage in phase III trials. 6 However, despite all these developments, the relation between tumor response and survival duration in solid tumors is still unclear both in clinical trials and in the individual patient. Indeed, a close relationship between the objective response rate (ORR) and survival has only been shown for some advanced cancers including breast and colon cancers by studies reviewing phase III trials. 7,8 In contrast, there have also been few studies addressing this association in patients with advanced NSCLC, using only relatively small cohorts. 9 In this paper, we address this issue for patients of advanced NSCLC by reviewing more recent data sets from larger patient cohorts accumulated from recent randomized phase III trials, where both the surrogate marker (objective response) and the clinical end point of interest (overall survival) have been measured. Journal of Thoracic Oncology Volume 2, Number 5, May 2007

Journal of Thoracic Oncology Volume 2, Number 5, May 2007 Phase III Trial and Survival in Lung Cancer In general, in advanced NSCLC, the ORRs to chemotherapeutic agents in the second-line or later setting are lower than those in the first-line setting in cases of advanced NSCLC. 10,11 Indeed, single-agent docetaxel, considered as the standard regimen for second-line therapy, yielded an ORR of only 6%, even though it was proven to yield a significant survival advantage. 2 In contrast, 43% of patients achieved stable disease. These findings might indicate that not only tumor shrinkage, but also disease stabilization, which is not reflected in the traditional ORR, could contribute to survival benefit in the second-line or later settings. In this situation, one might doubt whether tumor shrinkage would have the same value in the second-line or later settings as in the first-line setting. Thus, in this study, we evaluated the association between the ORR and patient survival separately for first-line chemotherapy and second-line or later chemotherapy. Additionally, we assessed the influence of both tumor shrinkage and disease stabilization on patient survival. PATIENTS AND METHODS Search for Trials We conducted a search for trials reported between 1991 and early 2006. To avoid publication bias, both published and unpublished trials were identified through a computer-based search of the PubMed database and abstracts from the 10 past conferences of the American Society of Clinical Oncology (the abstracts reported before 1996 were not accessible on Web site). We conducted the search using the following search terms: lung neoplasm, carcinoma, non-small cell, chemotherapy, and randomized controlled trial. The search was also guided by thorough examination of reference lists of original articles, review articles, and relevant books, and of the Physician Data Query registry of clinical trials. Selection of Trials Phase III trials were eligible if they compared systemic chemotherapy regimens that contained cytotoxic agents or molecular-targeted agents for advanced NSCLC. Trials that investigated only immunotherapy regimens were excluded from our analysis. Trials initially designed to assess combined treatment modalities including radiotherapy and surgery were also considered to be ineligible. Data Abstraction To avoid bias in the data abstraction process, two observers (K.H. and Y.F.) independently abstracted the data from the trials and subsequently compared the results. Of the trials that met the aforementioned criteria, data were abstracted from only chemotherapy regimens (but not the best supportive care alone arm) for this study. The following information was obtained from each report (per treatment regimen): number of patients enrolled and randomized, median age, proportion with poor performance status ( 2), female gender and stage IV disease, chemotherapy regimen, number of patients with objective response and stable disease (no change), and median survival time (MST). Docetaxel, paclitaxel, vinorelbine, gemcitabine, and irinotecan were defined as new chemotherapeutic agents. 12 Drugs believed to act on known molecular targets, such as tyrosine kinase inhibitors, antibodies to specific receptors, matrix metalloproteinase inhibitors, and antisense oligonucleotide, were defined as molecular-targeted agents. 13 Additionally, disease control was defined as the best tumor response of complete response, partial response, or stable disease (no change). 14 All the data were checked for internal consistency, and disagreements were resolved by discussion among the investigators. Quantitative Data Synthesis The association between the ORR and disease control rate (DCR) and MST was assessed using simple regression analysis. Multiple stepwise regression analysis (with stepping method criteria: probability of F to enter 0.05 and remove 0.10) was also performed to determine whether the ORR and DCR each independently influenced the survival of the patients with advanced NSCLC enrolled in the phase III studies, after adjustment for the following possible confounders: number of randomly assigned patients, median age of the patients, proportion with poor performance status, female gender, and stage IV disease. All p values corresponded with two-sided tests, and significance was set at a p value of less than 0.05. RESULTS Trial Flow and Characteristics of the Eligible Trials and Chemotherapy Regimens The flow chart of our study is shown in Figure 1. There were 148 potentially appropriate trials, of which five were excluded from our analysis because we were unable to obtain any relevant data from them. Finally, 143 trials involving 50,569 enrolled patients and 309 chemotherapy regimens fulfilled the eligibility criteria for inclusion in our analysis. Of the total number of patients, 47,869 were randomly allocated to chemotherapy regimens (43,551 patients to 290 regimens and 4,318 patients to 19 regimens in the first-line setting and FIGURE 1. A flow chart of the progress of trials through the review. Copyright 2007 by the International Association for the Study of Lung Cancer 403

Hotta et al. Journal of Thoracic Oncology Volume 2, Number 5, May 2007 TABLE 1. Demographics of the 143 Trials (per Trial) First-line Second-line or Later Total Total no. of trials 133 10 143 Total no. of randomized patients 43,551 4,318 47,869 Median no. of randomly assigned patients (range) 247 (24 1218) 316 (104 1129) 259 (24 1218) Publication type (full text/abstract form only) 117/16 8/2 125/18 No. of chemotherapy regimens ( 2/3/ 4) 105/22/6 9/1/0 114/23/6 % of patients with poor PS (median, range) 19 (0 100) 33 (0 35) 19 (0 100) % of female patients, median (range) 24 (3 80) 28 (9 35) 24 (3 80) % of pts with stage IV disease, median (range) 67 (27 100) 77 (48 89) 67 (27 100) PS, performance status. second-line or later setting, respectively). The baseline characteristics of the 143 trials are shown in Table 1. Most trials had two treatment regimens and low proportions of patients with poor performance status and female patients. The demographics of the 309 chemotherapy regimens are also listed in Table 2. In the first-line setting, cisplatin-based chemotherapy was the most frequently investigated (167 regimens; 58%). Molecular-targeted agents were assessed in 10 regimens (3%), all of them in combination with cytotoxic chemotherapeutic agents. Conversely, in the second-line or later settings, cisplatin-based chemotherapy was not investigated at all and newer agent monotherapy was the most frequently evaluated (13 regimens; 68%). Associations between the ORR and DCR and MST in Patients Receiving First-Line Chemotherapy Regimens In the first-line setting, based on intention-to-treat analysis using all randomized patients, the median ORR and DCR were 26.4% (range, 0 59.0%) and 62.5% (range, 14.% 94.3%), respectively. Cisplatin-based chemotherapy yielded a median ORR of 28.0%, which rose slightly to 30.4% when cisplatin was combined with newer agents. The overall median MST of patients receiving first-line chemotherapy was 8.5 months (range, 3.8 14.8 months), and the ORR was significantly associated with the MST (p 0.0001; Figure 2A). Subgroup analysis showed the close association between these parameters for the 61 combination chemotherapy regimens of cisplatin plus newer agents, one of the current standard treatments (p 0.0436). However, there was no significant correlation between the ORR and MST for molecular-targeted agents (3838 randomized patients with 10 regimens; p 0.3817), even though all these agents were combined with platinum-containing regimens. Regarding the DCR, we found a significant correlation between its rate and the MST in the first-line chemotherapy regimens (p 0.0001; Figure 2B). Multiple linear regression analysis revealed that the ORR and DCR were each independently associated with the MST, even after adjustment for the various possible confounders (regression coefficients 0.0788; p 0.0001 and 0.0794; p 0.0001, respectively). In this model, the MST was also affected by both proportion of patients with poor performance status (p 0.0001). Associations between the ORR and DCR and MST in Patients Receiving Second-Line or Later Chemotherapy Regimens (Salvage Regimens) In the second-line or later setting (salvage setting), data on the ORR were available for all 19 regimens. The median intention-to-treat ORR was only 6.8% (range, 0.8% 12.2%), whereas the median DCR was 42.4% (range, 30.9% 58.5%). The median MST was 6.6 months (range, 5.4 10.2 months) and showed no correlation with the ORR (p 0.6992, Figure 3A). In contrast, the DCR was better associated with the MST despite the small sample size (p 0.0129, Figure 3B). DISCUSSION We found a positive association between the ORR and MST based on the abstracted data obtained from clinical trials of advanced NSCLC evaluating newer chemotherapeutic agents in the first-line setting. A similar study was also conducted in 1998 using a smaller cohort. It demonstrated a significant correlation between the ORR and MST, although only single-agent chemotherapy regimens were assessed, and both first-line chemotherapy and second-line chemotherapy were investigated together. 9 Their results and ours might support the hypothesis that the achievement of an objective response is associated with a true survival benefit even in patients of advanced NSCLC that is only moderately chemosensitive. In this study, we used data collected in phase III trials not phase II trials because phase III trials involve very large data sets of patients. These trials also accurately measure the key clinical endpoint, namely, the overall survival, in addition to evaluation of the major surrogate marker, that is, the objective response. 15 However, our results must be interpreted with caution because the ORRs in phase III trials are not always similar to those in phase II trials, in which tumor shrinkage is exactly assessed as a surrogate marker. 16 One can also argue that the ORR data from phase III trials are less reliable than those of phase II trials because of the annotation with the major response criteria that confirmation of changes in tumor size may not be part of the standard practice in protocols where overall survival is evaluated as the key endpoint. 17 However, it was recently reported that confirmation of responses did not seem to add any value to response 404 Copyright 2007 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology Volume 2, Number 5, May 2007 Phase III Trial and Survival in Lung Cancer TABLE 2. Chemotherapy Regimens and Outcomes (per Chemotherapy Regimen) No. of Treatment Arms No. of Randomized pts. CR Rate (%), PR Rate (%), OR Rate (%), DC Rate (%), MST (mo), First-line chemotherapy Total no. of chemotherapy 290 43,551 0.9 (0 13.6) 24.3 (0 53.8) 26.4 (0 59.0) 62.5 (14.3 94.3) 8.5 (3.8 14.8) regimens Platinum-based regimens Cisplatin-based Cisplatin monotherapy 7 1194 0.4 (0 2.8) 11.1 (6.3 12.4) 11.3 (6.7 13.9) 46.4 (31.9 85.0) 6.9 (5.3 9.1) Cisplatin new agents 61 10,685 1.3 (0 5.8) 30.2 (12.1 45.4) 30.4 (13.0 47.2) 64.3 (37.0 79.8) 9.6 (6.3 14.8) combinations Cisplatin other (older) 99 9820 1.0 (0 9.3) 25.0 (5.4 53.8) 27.1 (7.6 59.0) 68.3 (33.1 94.3) 8.0 (4.9 12.0) agents Other platinum-based Carboplatin-based 43 9616 1.2 (0 6.1) 27.1 (7.9 43.4) 28.3 (7.9 45.0) 67.4 (38.8 84.8) 9.0 (6.6 12.3) combinations Others 1 70 0 11.4 11.4 70.0 8.9 Nonplatinum regimens Regimens containing newer agents Monotherapy 24 3101 0.3 (0 2.2) 14.5 (5.7 41.1) 15.5 (5.7 41.1) 47.5 (14.3 90.4) 7.1 (4.2 14.3) Combination therapy 22 3806 1.0 (0 4.3) 27.6 (9.2 37.2) 27.4 (9.2 41.0) 57.2 (31.2 72.6) 8.7 (6.7 11.5) Regimens containing 23 1421 0 (0 13.6) 17.6 (0 37.9) 17.6 (0 51.5) 53.6 (28.1 75.7) 6.8 (3.8 10.2) other (older) agents Molecular-targeted agents (with cytotoxic chemotherapy) EGFR-TKI containing 6 2494 1.9 (0.6 3.0) 30.0 (27.8 44.1) 31.0 (21.5 47.1) 56.1 9.9 (8.7 10.6) Other combinations 4 1344 1.1 (0.9 1.2) 22.8 (21.2 23.8) 24.3 (22.4 25.8) 58.4 (53.2 63.5) 10.8 (8.6 12.5) Second-line chemotherapy or later settings Total no. of chemotherapy 19 4318 0.1 (0 6.1) 6.2 (0.8 10.4) 6.8 (0.8 12.2) 42.4 (30.9 58.5) 6.6 (5.4 10.2) regimens Nonplatinum regimens Regimens containing newer agents Monotherapy 13 1816 0.5 (0 6.1) 6.1 (2.7 10.4) 8.3 (2.7 12.2) 42.7 (31.3 58.2) 7.1 (5.4 10.1) Combination therapy 1 65 0 6.2 6.2 58.5 10.2 Regimens containing other (older) agents Monotherapy 3 820 0 (0 0) 2.7 (0.8 4.6) 4.6 (0.8 8.5) 31.1 (30.9 31.4) 6.5 (5.6 8.3) Molecular-targeted agents (with cytotoxic chemotherapy) EGFR-TKI-containing a 2 1617 0.35 b 7.0 b 7.3 b 36.5 b 6.2 b a Monotherapy with tyrosine kinase inhibitor. b Mean scores. CR, complete response; PR, partial response; OR, objective response; DC, disease control; MST, median survival time; EGFR-TKI, epidermal growth factor receptor tyrosine kinase inhibitor. assessment based on the first best response observed. 18 Thus, this issue might only minimally affect our results, although further studies may be required. In contrast to the case for first-line chemotherapy, we failed to find any clear association between the ORR and survival after second-line or later chemotherapy for relapsed NSCLC (p 0.6992). In contrast, the MST correlated with the DCR (p 0.0129). In this setting, chemotherapeutic agents could generally be expected to bring about disease stabilization rather than tumor shrinkage. It has been proposed that the proportion of patients whose best response is progressive disease may be a surrogate marker that seems to reflect the DCR, complete response, partial response, or stable disease (no change). 14,19 The association between this progressive disease rate and the MST was also investigated in 42 first- and second-line single-agent phase II trials of NSCLC, and the progressive disease rate indeed correlated closely with the MST. 20 Similar to this impact of progressive disease rate on patient survival, we also found an association between the DCR and MST with the larger data set. These suggest that the DCR might be a good candidate for one of the surrogate markers for novel types of anticancer agents both in the first-line and second-line or later settings, although the second-line data were more heterogeneous and smaller. In regard to molecular-targeted agents, the majority usually do not produce any significant tumor shrinkage. 13 Copyright 2007 by the International Association for the Study of Lung Cancer 405

Hotta et al. Journal of Thoracic Oncology Volume 2, Number 5, May 2007 FIGURE 2. A: Association between objective response rate and median survival time in the first-line chemotherapy setting (p 0.0001). B: Association between disease control rate and median survival time in the first-line chemotherapy setting (p 0.0001). FIGURE 3. A: Association between objective response rate and median survival time in the secondline or later chemotherapy setting (p 0.6992). B: Association between disease control rate and median survival time in the secondline or later chemotherapy setting p 0.0129). Many investigators agree that objective response can no longer be considered as an accurate surrogate marker for molecular-targeted agents. 21 Rather, the efficacy data obtained when these agents are combined with other chemotherapeutic agents have generally been considered to be helpful for defining the potential efficacy of molecular-targeted agents. 13 However, we failed to find that the objective response to such combination regimens was associated with patient survival, which suggests that this methodology might not be sufficiently sensitive for accurate determination of the subtle effects of these compounds. The limitation in our study was that the number of trials involving molecular-targeted agents that were included in the study was obviously limited. We also assessed the association between the ORR and patient survival in the first-line setting for all moleculartargeted agents together, although these agents have diverse mechanisms of action. 4 Thus, further investigation including more relevant trials are warranted to confirm whether data from combination chemotherapy would be useful. Several major limitations should be discussed; we dealt with mostly negative and very heterogeneous trials, limiting our conclusions. Additionally, although the MST is the most frequently reported indicator as survival data, this parameter has not yet proven to represent the entire benefit from chemotherapy. All our analyses were based on abstracted data and not individual patient data (IPD). An IPD-based analysis could be a more appropriate statistical design to assess the relationship between the ORR and DCR and survival and must be strongly considered in a future study. Furthermore, the response criteria varied among the trials reviewed by us, mainly between World Health Organization criteria and Response Evaluation Criteria in Solid Tumors. However, this factor seems unlikely to have significantly affected our key results because there are few differences between these two sets of criteria in terms of validity in relation to tumor volume and intercriteria reproducibility in NSCLC patients. 22 In conclusion, our data might support the hypothesis that the achievement of an objective response to first-line chemotherapy in advanced NSCLC patients is associated with a true survival benefit, although an IPD-based metaanalysis would be strongly needed to confirm it. Despite the several limitations of our study, our data reveal critical points that must be evaluated in future studies. Conversely, for treatment situations that are less expected to bring about tumor shrinkage, including assessment of molecular-targeted agents or chemotherapy in the second-line or later settings, it might be desirable to develop alternative surrogate markers. Further assessments are warranted to clarify the associations between the ORR and DCR and patient outcome, especially in such regimens. ACKNOWLEDGMENTS The authors thank Drs. Paul Lorigan, Paolo Lissoni, Thierry Le Chevalier, Cesare Gridelli, Francesco Perrone, Massimo Di Maio, Sumitra Thongprasert, Edith A. Perez, Silvia Novello, Corey Langer, Christian Manegold, John D. Hainsworth, Joan B. Erland, and Nasser Hanna and Gianfranco Buccheri for their support, provision of data, or comments on our analyses. REFERENCES 1. Bunn PA. Chemotherapy for advanced non-small-cell lung cancer: who, what, when, why? J Clin Oncol 2002;20(18 Suppl):23S 33S. 2. Shepherd FA, Dancey J, Ramlau R, et al. Prospective randomized trial of docetaxel versus best supportive care in patients with non-small-cell 406 Copyright 2007 by the International Association for the Study of Lung Cancer

Journal of Thoracic Oncology Volume 2, Number 5, May 2007 Phase III Trial and Survival in Lung Cancer lung cancer previously treated with platinum-based chemotherapy. J Clin Oncol 2000;18:2095 2103. 3. Fossella FV, DeVore R, Kerr RN, et al. Randomized phase III trial of docetaxel versus vinorelbine or ifosfamide in patients with advanced non-small-cell lung cancer previously treated with platinum-containing chemotherapy regimens. The TAX 320 Non-Small Cell Lung Cancer Study Group. J Clin Oncol 2000;18:2354 2362. 4. Kohn EC, Liotta LA. Molecular insights into cancer invasion: strategies for prevention and intervention. Cancer Res 1995;55:1856 1862. 5. Shepherd FA, Rodrigues Pereira J, Ciuleanu T, et al. Erlotinib in previously treated non-small-cell lung cancer. N Engl J Med 2005;353: 123 132. 6. Sandler AB, Brahmer RGJ, Dowlati A, Schiller JH, Perry MC, Johnson DH. Randomized phase II/III Trial of paclitaxel (P) plus carboplatin (C) with or without bevacizumab (NSC # 704865) in patients with advanced non-squamous non-small cell lung cancer (NSCLC): An Eastern Cooperative Oncology Group (ECOG) Trial - E4599. Proc Am Soc Clin Oncol 2005;23(abstr 4). 7. Bruzzi P, Del Mastro L, Sormani MP, et al. Objective response to chemotherapy as a potential surrogate end point of survival in metastatic breast cancer patients. J Clin Oncol 2005;23:5117 5125. 8. Buyse M, Thirion P, Carlson RW, et al. Relation between tumor response to first-line chemotherapy and survival in advanced colorectal cancer: a meta-analysis Meta-Analysis Group in Cancer. Lancet 2000; 356:373 378. 9. Sekine I, Kubota K, Nishiwaki Y, Sasaki Y, Tamura T, Saijo N. Response rate as an endpoint for evaluating new cytotoxic agents in phase II trials of non-small-cell lung cancer. Ann Oncol 1998;9:1079 1084. 10. Huisman C, Smit EF, Giaccone G, Postmus PE. Second-line chemotherapy in relapsing or refractory non-small-cell lung cancer: a review. J Clin Oncol 2000;18:3722 3730. 11. Hanna N, Shepherd FA, Fossella FV, et al. Randomized phase III trial of pemetrexed versus docetaxel in patients with non-small-cell lung cancer previously treated with chemotherapy. J Clin Oncol 2004;22:1589 1597. 12. Pfister DG, Johnson DH, Azzoli CG, et al. American Society of Clinical Oncology treatment of unresectable non-small-cell lung cancer guideline: update 2003. J Clin Oncol 2004;22:330 353. 13. Yamanaka T, Okamoto T, Ichinose Y, Oda S, Maehara Y. Methodological aspects of current problems in target-based anticancer drug development. Int J Clin Oncol 2006;11:167 175. 14. Fukuoka M, Yano S, Giaccone G, et al. Multi-institutional randomized phase II trial of gefitinib for previously treated patients with advanced non-small-cell lung cancer (The IDEAL 1 Trial). J Clin Oncol 2003;21: 2237 2246. 15. Schilsky RL. End points in cancer clinical trials and the drug approval process. Clin Cancer Res 2002;8:935 938. 16. Shanafelt TD, Loprinzi C, Marks R, Novotny P, Sloan J. Are chemotherapy response rates related to treatment-induced survival prolongations in patients with advanced cancer? J Clin Oncol 2004;22:1966 1974. 17. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 2000;92:205 216. 18. Perez-Gracia JL, Munoz M, Williams G, et al. Assessment of the value of confirming responses in clinical trials in oncology. Eur J Cancer 2005;41:1528 1532. 19. Eisenhauer EA. Phase I and II trials of novel anti-cancer agents: endpoints, efficacy and existentialism. The Michel Clavel Lecture, held at the 10th NCI-EORTC Conference on New Drugs in Cancer Therapy, Amsterdam, 16 19 June 1998. Ann Oncol 1998;9:1047 52. 20. Sekine I, Tamura T, Kunitoh H, et al. Progressive disease rate as a surrogate endpoint of phase II trials for non-small-cell lung cancer. Ann Oncol 1999;10:731 733. 21. Morabito A, Di Maio M, De Maio E, Normanno N, Perrone F. Methodology of clinical trials with new molecular-targeted agents: where do we stand? Ann Oncol 2006;17(Suppl 7):vii128 vii131. 22. Watanabe H, Yamamoto S, Kunitoh H, et al. Tumor response to chemotherapy: the validity and reproducibility of RECIST guidelines in NSCLC patients. Cancer Sci. 2003;94:1015 1020. Copyright 2007 by the International Association for the Study of Lung Cancer 407