|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 26, No 3 (January 20), 2008: pp. 463-467 © 2008 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.13.0344 Disease Control Rate at 8 Weeks Predicts Clinical Benefit in Advanced Non–Small-Cell Lung Cancer: Results From Southwest Oncology Group Randomized Trials
From the University of California Davis Cancer Center, Sacramento, CA; Fred Hutchinson Cancer Research Center, Seattle, WA; Cancer Research and Biostatistics, Seattle, WA; University of Colorado, Denver, CO; University of Maryland, Baltimore, MD; and the University of Kansas, Kansas City, KS Corresponding author: Primo N. Lara Jr, MD, University of California Davis Cancer Center, 4501 X St, Sacramento, CA 95817; e-mail: primo.lara{at}ucdmc.ucdavis.edu
Purpose Tumor shrinkage categorized as complete response (CR) or partial response (PR) is a fundamental efficacy measure for new cancer treatments and often considered a surrogate for overall survival. However, for any given treatment, many more patients typically achieve stable disease (SD) or have progressive disease (PD) than achieve response. We hypothesized that PD (or its converse, disease control rate [DCR], consisting of CR, PR, SD) is a stronger predictor of survival than response alone in advanced non–small-cell lung cancer (NSCLC), and that this determination might be assessable early on during therapy. Patients and Methods Data from 984 NSCLC patients entered onto three randomized Southwest Oncology Group trials of platinum-based chemotherapy were pooled and subjected to Landmark survival analysis. Patients were categorized according to proportions alive at weeks 8, 14, and 20 after registration, as well as response status. Elements were fitted into a Cox proportional hazards model. Results Tumor response (CR, PR) was seen in 260 patients (27%). Median time to response, time to progression, and survival time were 2.0, 4.3 and 8.9 months, respectively. Median survival times among patients with CR/PR, SD, or PD were 13.5, 8.4, and 3.1 months, respectively. Of 892 patients alive at week 8, DCR was 62%. Although CR/PR at week 8 was associated with longer survival (hazard ratio [HR] = 0.61; P < .001), DCR was superior in predicting survival (HR = 0.45; P < .0001). Conclusion DCR at week 8 is a more powerful predictor of subsequent survival than is the traditional tumor response rate in advanced NSCLC and provides an early assessment of subsequent outcome.
In advanced non–small-cell lung cancer (NSCLC), tumor shrinkage or "response" has typically been equated with clinical benefit from systemic therapy. Traditionally, physicians and patients have assumed that response results in prolonged survival.1 Conversely, tumor growth has been associated with worse outcomes and early death. Disease stabilization, wherein tumor size fails to meet criteria for response or progression, has often been viewed as an equivocal result and is therefore of unclear clinical value. In clinical trials, such "stable disease" (SD) is often discounted in favor of tumor response, which in turn is widely used as a screen for drug activity. In reality, only a minority of patients with advanced NSCLC experience tumor shrinkage after standard platinum-based chemotherapy. Many more patients experience either SD or progressive disease (PD). Moreover, in some phase III trials, improved response rate of one regimen over another has failed to result in improved survival.2-4 Thus, a clear correlation between response and long-term benefit has not yet been established.5 Methods defining tumor response have evolved during the last few decades. In 1960, it was suggested that systemic therapy had a positive outcome if the total tumor mass decreased in size, with no lesions increasing in size and no new lesions appearing.6 In 1979, the WHO codified this philosophy by establishing bidimensional tumor measurement as a standard.7 A partial response (PR) was arbitrarily defined as a 50% or greater decrease in tumor size, whereas progression was defined as a 25% or greater increase in size or new lesions. The Southwest Oncology Group further modified this system in 1992, employing a volumetric definition for progression.8 In 2000, the Response Evaluation Criteria in Solid Tumors (RECIST) was introduced after a multinational effort to standardize tumor response assessment. RECIST established unidimensional tumor measurement, utilizing the longest diameter as a reproducible and simpler standard, and defined PR as a 30% or greater decrease in the sum of the measurable lesions and PD as a 20% or greater increase or appearance of new lesions.9 At the present time, RECIST is the most widely used method of assessing response to anticancer therapy. There are several limitations in tumor size assessment that restrict the broad applicability of standardized objective criteria. For instance, some patients will have no measurable disease, such as those with a malignant pleural effusion or ill-defined pulmonary densities. There is also high inter- and intraobserver variability in the measurement of NSCLC lesions.10 Importantly, some lesions may radiographically change in appearance under the influence of systemic therapy, but not necessarily regress in size. For example, lesions can "fade" or cavitate, in neither case meeting criteria for response. These changes are coded as SD, implying no benefit of therapy, although in some cases SD has been associated with prolonged survival.11 In light of the reality that many more tumors achieve nonprogression than response, we hypothesized that the rate of nonprogression, also termed the disease control rate (DCR) is a stronger predictor of clinical benefit than traditional tumor response rate (the sum of complete response [CR] and PR) after platinum-based chemotherapy in patients with advanced NSCLC.
Data from 984 patients with stage IV (metastatic) or IIIB (malignant pleural effusion) NSCLC accrued onto three randomized Southwest Oncology Group (SWOG) trials (S9509, S9806, and S0003) of platinum-based chemotherapy were pooled. S9509 randomly assigned 415 patients to either carboplatin/paclitaxel or cisplatin/vinorelbine.12 S9806 randomly assigned 193 patients to either cisplatin/vinorelbine followed by docetaxel or carboplatin/gemcitabine followed by paclitaxel.13 S0003 randomly assigned 376 patients to carboplatin/paclitaxel with or without tirapazamine.14 S9509 and S9806 employed the SWOG (modified WHO) tumor response criteria, whereas S0003 used RECIST. In S9509 and S0003, patients underwent disease assessments after cycle 2 (approximately week 6) and every two cycles thereafter during receipt of protocol treatment until disease progression and/or completion of protocol treatment. In S9806, reassessment occurred after the third (approximately week 8) and sixth cycles of therapy. None of the trials showed superiority of one treatment arm over the other. Landmark analyses were performed to assess the association of the intermediate outcomes with overall survival.15 Three separate analyses were performed at weeks 8, 14, and 20 after registration. Each analysis included only patients alive at each of the time points. Disease control status at each time point was defined as the "best status to date," specifically if patients had CR, PR, or SD. The association of clinical prognostic factors with response, disease control, and survival status at the three time points was assessed using logistic regression. A Cox proportional hazards model was used to assess the associations between disease status at the landmark times and to adjust for prognostic factors. The data analysis for this article was generated using SAS/STAT software, Version 9.2 of the SAS System for PC (SAS Inc, Cary, NC).
Patient Characteristics Table 1 summarizes the characteristics of patients included in this analysis. Overall, the median age of patients was 62 years, with 41% over the age of 65 years. There were 647 males (66%). The vast majority of patients (84%) had stage IV disease. Four hundred four patients (42%) reported weight loss of at least 5% or more. Zubrod performance status (PS): 36% had PS = 0 whereas 63% had PS = 1. There was no difference in the overall survival across the three trials (Fig 1). In the pooled data set, overall response rate was 27% with a median time to response among responders of 2 months. This time point corresponded to the initial tumor response assessment after two or three cycles of platinum-based chemotherapy. Median time to progression was 4.3 months, whereas median overall survival time was 8.9 months. These data are summarized in Table 2.
Prognostic Factors The association between survival and potential prognostic factors such as sex, age, weight loss, PS, and stage was assessed. Performance status and weight loss of 5% or greater were significantly associated with worse survival times (P < .0001 for both) whereas male sex, age, and stage of disease were not significantly associated with survival time (P = .07, 0.36, and 0.13, respectively). These clinical factors were then analyzed with respect to response and PD at weeks 8, 14, and 20. Response by week 8 was not associated with any of the prognostic factors (P > .05, for all) while a performance status more than 0 and stage IV disease were associated with a reduced odds of disease control by week 8 (odds ratio [OR] = 0.60; 95% CI, 0.45 to 0.79; and OR = 0.63; 95% CI, 0.42 to 0.95, respectively). Response by week 14 was not significantly associated with any of the clinical factors, whereas performance status and stage IV disease remained negatively associated with DCR at weeks 14 and 20. Performance status of more than 0 was also associated with a reduced chance of response and by week 20. Male sex, age, and weight loss were not associated with response or DCR at any of the time points.
Survival As a Function of Tumor Response
Response and Survival: weeks 8, 14, and 20 Eight hundred ninety-two patients (91%) enrolled were alive at 8 weeks after registration. Of these patients, 111 (12%) had a response, 448 (50%) had SD, and 333 (37%) had neither a response nor SD by week 8. Therefore, the DCR among patients alive at 8 weeks was 62%. The median survival from registration was 14.7, 12.0, and 6.4 months among responders, those with SD, and those without a response or SD, respectively. Both response and SD at 8 weeks were associated with longer survival times by comparison with patients who never achieved a response or SD. Response (v nonresponse) at week 8 was associated with longer survival times with an estimated hazard ratio equal to 0.61 (95% CI, 0.50 to 0.76; P < .001). Additionally, disease control at week 8 was also associated with longer survival, with an HR of 0.45 (95% CI, 0.39 to 0.52; P < .0001). There was a survival advantage among patients who had achieved a response compared with those with SD, and this difference was statistically significant (P = .03). However, patients with SD had significantly longer survival times than did those with PD (P < .0001). Figure 2 presents Kaplan-Meier survival curves as a function of tumor response at week 8 of protocol therapy, demonstrating how survival after SD tracks closer to response than to PD. Finally, performance status and weight loss were also found to be significantly associated with survival, with HR of 1.34 for PS greater than 0 and HR of 1.27 for weight loss of 5% or greater.
A similar analysis was performed for weeks 14 and 20 of protocol therapy. In general, no substantial new findings were seen compared with week 8 of therapy, as summarized in Tables 4 and 5. In all analyses, the survival outcomes for those patients with SD were significantly different from those of patients who never achieved a response or SD. Although there was also a significant difference between responders and achievement of SD, for the earlier time points, survival outcomes for those with SD were more similar to those of responders than to those who never achieved a response or SD.
In this Landmark survival analysis of a large SWOG database of patients with advanced NSCLC, we found that DCR is a stronger predictor of subsequent outcome after platinum-based chemotherapy than is the traditional response rate of CR + PR. The reduction in the risk of death was substantially more significant for DCR (P < .0001) than it was for standard CR + PR rate alone. Thus, DCR may serve as a surrogate for survival after systemic therapy in this disease. It is worth noting that an ideal surrogate end point must be associated with clinical outcome (such as survival) and fully capture the net effect of a treatment on that outcome. In the absence of a large data set from randomized controlled trials that shows a statistically significant effect of treatment on survival, we believe that the approach taken in this analysis serves as a reasonable alternative. In the era of biologic "targeted" therapies that may increase the proportion of SD, a DCR metric more closely mirrors treatment effect than the traditional response rate. If prospectively validated, application of the concept of DCR at 8 weeks might also provide clinical investigators an early-look clinical tool to assess the value of systemic therapy in this setting, allowing in-progress alterations in study design or sample size. These data also support the relative value of disease stabilization as a "positive outcome" for advanced NSCLC. Recent prospective data have shown that SD is clinically relevant. Specifically, in a randomized placebo-controlled trial of erlotinib in the second- and third-line settings, patients with SD seemed to have derived clinical benefit from the then-experimental epidermal growth factor receptor–targeted therapy.11 Our results can potentially guide clinical practice in that oncologists can appropriately counsel NSCLC patients who are receiving platinum-based chemotherapy about the relative value of tumor shrinkage, SD, and PD using evidence-based data rather than anecdotal experience. Clinicians would thus have such data on which to base the logical conclusion that patients with nonprogression at the first radiographic assessment have a survival benefit over those with PD. Employing DCR has practical implications for clinical investigations in which the assessment of PD is, in some cases, less equivocal than assessment of response. For example, the development of new lesions on physical examination or radiographic studies is less equivocal than measuring gradations of tumor shrinkage. If DCR were selected as a primary trial end point rather than response (which is often used in phase II efficacy assessment), then measurable disease would not be required at baseline. This change would broaden eligibility for phase II trials, increasing general applicability, and may speed trial completion. In conclusion, our data suggest that DCR is a more powerful predictor of subsequent survival than is the traditional response rate (CR + PR) in patients receiving platinum-based chemotherapy for advanced stage NSCLC. Employing DCR instead of response rate may enhance clinical trial interpretation and guide clinical practice. If validated, an early-look measure such as DCR at week 8 could also enhance trial design and conduct in advanced NSCLC. These findings may have broad implications for future trial design and will be prospectively tested in SWOG.
The author(s) indicated no potential conflicts of interest.
Conception and design: Primo N. Lara Jr, Mary W. Redman Administrative support: Mary W. Redman, John J. Crowley Provision of study materials or patients: Primo N. Lara Jr, Karen Kelly, Martin J. Edelman, Stephen K. Williamson, David R. Gandara Collection and assembly of data: Primo N. Lara Jr, Mary W. Redman, John J. Crowley Data analysis and interpretation: Primo N. Lara Jr, Mary W. Redman, Karen Kelly, Martin J. Edelman, Stephen K. Williamson, John J. Crowley, David R. Gandara Manuscript writing: Primo N. Lara Jr, Mary W. Redman, David R. Gandara Final approval of manuscript: Primo N. Lara Jr, Mary W. Redman, Karen Kelly, Martin J. Edelman, Stephen K. Williamson, John J. Crowley, David R. Gandara
Presented in part at the 42nd Annual Meeting of the American Society of Clinical Oncology, June 2-6, 2006, Atlanta, GA. Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Therasse P: Measuring the clinical response: What does it mean? Eur J Cancer 38:1817-1823, 2002[CrossRef][Medline] 2. Gatzemeier U, von 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 18:3390-3399, 2000 3. Klastersky J, Sculier JPO, Laccroiz H, et al: A randomized study comparing cisplatin or carboplatin with etoposide in patients with advanced non-small cell lung cancer. J Clin Oncol 8:1556-1562, 1990[Abstract] 4. Gandara DR, Crowley J, Livingston R, 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 11:873-878, 1993 5. Weiss GB, Bunce H III, Hokanson JA: Comparing survival of responders and non-responders after treatment: A potential source of confusion in interpreting cancer clinical trials. Controlled Clin Trials 4:43-52, 1983[Medline] 6. Zubrod CG, Schneiderman M, Frei E III, et al: Appraisal of methods for the study of chemotherapy of cancer in man: Comparative therapeutic trial of nitrogen mustard and triethylene thiophosphoramide. J Chronic Dis 11:7-33, 1960 7. Miller AB, Hoogstraten B, Staquet M, et al: Reporting results of cancer treatment. Cancer 47:207-214, 1981[CrossRef][Medline] 8. Green S, Weiss GR: Southwest Oncology Group standard response criteria, endpoint definitions, and toxicity criteria. Invest New Drugs 10:259-253, 1992 9. Therasse P, Arbuck SG, Eisenhauer EA, et al: New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 92:205-216, 2000 10. Erasmus JJ, Gladish GW, Broemeling L, et al: Interobserver and intraobserver variability in measurement of non-small cell lung cancer lesions: Implications for assessment of tumor response. J Clin Oncol 21:2574-2582, 2003 11. Shepherd FA, Rodrigues Pereira J, Ciuleanu T, et al: Erlotinib in lung cancer. N Engl J Med 353:1739-1741, 2005 12. Kelly K, Crowley J, Bunn P, 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 19:3210-3218, 2001 13. Edelman M, Clark JI, Chanski K, et al: Randomized phase II trial of sequential chemotherapy in advanced non-small cell lung cancer (SWOG 9806): carboplatin/gemcitabine followed by paclitaxel or cisplatin/vinorelbine followed by docetaxel. Clin Cancer Res 10:5022-5026, 2004 14. Williamson SK, Crowley J, Lara PN, et al: Phase III trial of paclitaxel plus carboplatin with or without tirapazamine in advanced non–small-cell lung cancer: Southwest Oncology Group trial S0003. J Clin Oncol 23:9097-9104, 2005 15. Anderson JR, Cain KC, Gelber RD: Analysis of survival by tumor response. J Clin Oncol 1:710-719, 1983[Abstract] Submitted June 12, 2007; accepted October 16, 2007.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2008 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
