This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ajani, J. A. Search for Related Content PubMed PubMed Citation Articles by Ajani, J. A. Social Bookmarking                            What's this?

Standard of Care for Gastric Cancer Based on Meta-Analysis? Treading On Thin Ice or It Is Very Nice!

Department of Gastrointestinal Medical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, TX

To the Editor:

The article by Wagner et al¹ discusses important research issues for patients with advanced gastric cancer that should concern all investigators and treating physicians alike. We definitely need sustained momentum for conducting many prospective randomized trials to improve survival and quality of life of our patients with this disease. Meta-analysis is not a substitute for a well-designed and well-conducted clinical trial. It is also not a substitute when such trials have not been conducted. What elements might constitute an exemplary clinical trial? Each trial must ask one simple question to determine the contribution of a potentially valuable agent. The trial must be meticulously designed and executed with minimum flaws in clinical trial methodologies. The trials must encompass an overall survival question as the primary end point and have a sufficient sample size (I estimate about 600 or more patients) to discern realistic differences. Now we have access to many interesting new agents and therefore, we should not waste our precious patient resource on noninferiority questions. Just a digression, what may be clinically relevant survival advantage to one group of investigators may not be relevant to others because of differences in perceptions, cultural habits, and health care systems. Most of us are trained to look at differences at the median value and pay much less attention to the area between the curves that is best represented by a log-rank P value. Nevertheless, a 3-month median survival advantage might be debated by a few but a difference of 2 weeks or 4 weeks will be debated by many (again without much emphasis on what the area between the curves might tell us).

The trial designs must also include relevant stratifications. The interim analyses and subgroup analyses must be stipulated before the trial accrues patients. An independent data monitoring committee must oversee the conduct of the trial throughout and review data periodically. Circumstances that might result in subsequent imbalances (for example, inclusion of patients who might end up with curative operations) that might jeopardize the primary or other coveted end points must be avoided. In addition, we must now take advantage of the technological frontier that has been opened up for conducting translational research to further improve our understanding of the agents, cancer, and patients.

Should we not use these approaches to make advances and set new standards? Yes. Is there a substitute? No, but obviously these authors might be suggesting that there is. Should we consider their recommendation of using meta-analysis to set a new standard of care by analyzing a total of less than 525 patients (Fig 3 of Wagner et al¹)? No. Should I recommend an anthracycline-containing regimen, such as epirubicin, cisplatin, and fluorouracil (ECF), to my next patient with advanced gastric cancer who does not wish to participate in a clinical trial? Should my colleagues in Japan, Korea, China, and South America do the same? I hardly think so. There has never been a trial to suggest that an anthracycline, by itself, provides a survival advantage to this group of patients. Figure 4 in the Wagner et al¹ article casts substantial doubts on the authors' assumption on the contribution of anthracyclines. It is disappointing that the authors failed to suggest a trial of epirubicin + CF (this is just one example of a relevant combination) versus CF, but I am not surprised.

What about developing one global standard for advanced gastric cancer? It would be naïve to assume that this is truly possible. Even for common malignancies such as breast, lung and colon cancers, such globally accepted standards are not in place. The reasons for these are beyond the data emerging from properly conducted trials and span from availability of agents, cultural differences in approach to patients, physician/patient habits/expectations, and reimbursement issues. In the future oncology worlds of individualization of therapy, the "standards" will disappear anyway.

The authors have successfully taken us back to the darker ages by using the meta-analysis as a tool to suggest a standard of care. A perfect meta-analysis would utilize individual patient data from large well-conceived and well-conducted trials that are identified prospectively. Any shortcomings to that approach (as evident in this article) results in uninterpretable conclusions. It is also subject to substantial publication bias. Wagner et al¹ managed to collect a handful of studies conducted over a span of 25 years using substantially heterogeneous (and often suboptimal) clinical methodologies in design and conduct. They have included randomized studies of even less than 50 patients (for example, their reference 25 that substantially skews the results in their Fig 3)² and developed an untenable message. May I remind them that individual trials with more than 1,000 patients have been completed or are being completed?³

We should focus our energies on exploiting the emerging frontier of biochemotherapy, molecular biology, and patient genetics to develop efficacious therapy for our patients and forget about spending our efforts on ill-conceived retrospective analyses that are nothing but depressing, inaccurate, present a view that is disappointing, and weaken our rigor for conducting good prospective clinical research.

Author's Disclosures of Potential Conflicts of Interest

The author indicated no potential conflicts of interest.

ACKNOWLEDGMENTS

I thank Donald Berry, PhD, Department of Biostatistics, The University of Texas M.D. Anderson Cancer Center, Houston, TX, for his helpful comments.

REFERENCES

1. Wagner AD, Grothe W, Haerting J, et al: Chemotherapy in advanced gastric cancer: A systematic review and analysis on aggregate data. J Clin Oncol 24:2903-2909, 2006[Abstract/Free Full Text]

2. Kyoto Research Group for Chemotherapy of Gastric Cancer: A randomized, comparative study of combination chemotherapies in advanced gastric cancer: 5-fluorouracil and cisplatin (FP) versus 5-fluorouracil, cisplatin, and 4'-epirubicin (FPEPIR). Anticancer Res 12:1983-1988, 1992[Medline]

3. Cunningham D, Rao S, Starling T, et al: Randomized multicenter phase III study comparing capecitabine with fluorouracil and oxaliplatin with cisplatin in patients with advanced esophagogastric cancer: The REAL 2 trial. J Clin Oncol 24:18S, 2006 (suppl; abstr LBA4017)[CrossRef]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				Y.-K. Kang, W.-K. Kang, D.-B. Shin, J. Chen, J. Xiong, J. Wang, M. Lichinitser, Z. Guan, R. Khasanov, L. Zheng, et al. Capecitabine/cisplatin versus 5-fluorouracil/cisplatin as first-line therapy in patients with advanced gastric cancer: a randomised phase III noninferiority trial Ann. Onc., April 1, 2009; 20(4): 666 - 673. [Abstract] [Full Text] [PDF]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ajani, J. A. Search for Related Content PubMed PubMed Citation Articles by Ajani, J. A. Social Bookmarking                            What's this?