Correspondence
Adjuvant Chemoradiotherapy for Gastric Cancer
N Engl J Med 2002; 346:210-211January 17, 2002
- Article
To the Editor:
The study by Macdonald and colleagues (Sept. 6 issue)1 is commendable for its duration, detailed analysis, and quality control. However, we believe that the authors should have randomly assigned the patients to one of the two study groups before surgery. The definitions of resectability and stratification according to tumor stage are meaningless because the surgical approach was not uniform. Moreover, the conclusion that postoperative chemoradiotherapy should be considered for all patients at high risk for recurrence does not correlate with the aims of the study. There appears to have been a substantial number of major side effects associated with chemoradiotherapy (Table 3 of the article), and only 64 percent of the patients assigned to chemoradiotherapy could complete the treatment. One wonders whether it is really worth giving chemoradiotherapy to these patients postoperatively, for a nine-month median survival benefit. In addition, we would be interested to know the number of patients who were excluded before randomization because they opted not to enter the trial.
1 ReferencesIype Abraham, M.D.
Puneet Dhar, M.Ch.
Raaj K. Praseedom, F.R.C.S.
Amrita Institute of Medical Sciences and Research Centre, Cochin 682026, India
gisurgery@aimshospital.org 1
Macdonald JS ,Smalley SR ,Benedetti J , et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med 2001;345:725-730
Full Text | Web of Science | Medline
To the Editor:
Macdonald et al. report the results of a randomized trial of postoperative adjuvant chemoradiotherapy for gastric carcinoma. Two important questions are whether the distribution of patients undergoing D0, D1, or D2 resection (resection involving incomplete, complete, or extensive lymph-node dissection, respectively) was well balanced between the treatment groups and whether adjuvant treatment had a beneficial effect in patients with early-stage disease as well as in those with late-stage disease. Although we may accept the authors' statement that D2 resection has not been proven to improve survival, we must not forget the importance of type II errors: the negative studies they cite should not be taken as proof that the extent of resection has no effect on survival.1 Furthermore, we may question whether D0 resection results in poorer survival than D1 or D2 resection. Since some clinicians will no doubt question whether the inclusion of node-negative, stage T1 and T2 tumors was clinically appropriate, a subgroup analysis according to stage would be relevant.
The authors also provide confidence intervals for hazard ratios on the basis of a Cox regression analysis. However, the survival curves in Figure 1 of the article cross, indicating that the hazards are not proportional and thus violating a basic requirement of the Cox model.2 Since the last patient was registered on July 15, 1998, we may assume that follow-up information was available for virtually all the patients at three years. Readers should therefore note that the 95 percent confidence interval for the 9 percent difference between the groups in the reported overall survival rate at three years can be calculated to be 1 to 17 percent.3
3 ReferencesCarl D. Atkins, M.D.
South Shore Hematology–Oncology Associates, Rockville Centre, NY 11570
c3atkins@optonline.net 1
Freiman JA, Chalmers TC, Smith H Jr, Keubler RR. The importance of beta, the type II error, and sample size in the design and interpretation of the randomized controlled trial: survey of two sets of “negative” trials. In: Bailar JC III, Mosteller F, eds. Medical uses of statistics. 2nd ed. Boston: NEJM Books, 1992:357-73.
2
Concato J ,Feinstein AR ,Holford TR . The risk of determining risk with multivariable models. Ann Intern Med 1993;118:201-210
Web of Science | Medline3
Armitage P, Berry G. Statistical methods in medical research. 3rd ed. Oxford: Blackwell Scientific, 1994.
To the Editor:
Macdonald et al. found a greater median survival with surgery plus chemoradiotherapy than with surgery alone in patients with resectable gastric cancer. They concluded that chemoradiotherapy should be considered for all patients who have a high risk of recurrence. The authors recommend extensive (D2) resection, but only 10 percent of the patients in their study underwent such extensive dissection; a D0 resection was performed in more than half the patients. Two European randomized trials1,2 did not find any survival benefit with a formal D2 lymph-node dissection, as compared with a less invasive, D1 resection.
Macdonald et al. found a three-year relapse-free survival rate of 31 percent in their surgery-only group, as compared with about 60 percent in patients who underwent D1 or D2 resection in a Dutch study.1 Furthermore, the five-year survival rates achieved in the two European trials after D1 or D2 procedures were 35 to 47 percent, whereas Macdonald et al. found a three-year overall survival rate of 41 percent in their surgery-only group. It has clearly been demonstrated that examination of 15 or more nodes is necessary for accurate pathological staging.3 Such a number simply cannot be achieved with a D0 resection. These results strongly suggest that D0 lymphadenectomy is an inadequate oncologic procedure in terms of local control and survival and that it should be abandoned.
3 ReferencesChristophe R. Berney, M.D., Ph.D.
Neil D. Merrett, M.D.
Bankstown Hospital, Bankstown, NSW 2200, Australia
berneyc@hotmail.com 1
Bonenkamp JJ ,Hermans J ,Sasako M ,van de Velde CJH . Extended lymph-node dissection for gastric cancer. N Engl J Med 1999;340:908-914
Full Text | Web of Science | Medline2
Cuschieri A ,Weeden S ,Fielding J , et al. Patient survival after D1 and D2 resections for gastric cancer: long-term results of the MRC randomized surgical trial. Br J Cancer 1999;79:1522-1530
CrossRef | Web of Science | Medline3
Karpeh MS ,Leon L ,Klimstra D ,Brennan MF . Lymph node staging in gastric cancer: is location more important than number? An analysis of 1,038 patients. Ann Surg 2000;232:362-371
CrossRef | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Dr. Abraham and colleagues suggest that we should have randomly assigned the patients to a study group before surgery. This point is not relevant to our study, since the eligibility criteria (resection with negative margins and stage IB through IVM0 disease [according to the 1988 staging criteria of the American Joint Commission on Cancer]) could be determined only postoperatively. They also question the value of a “nine-month median survival benefit.” The survival benefit in our study is superior to any reported previously and represents an advance in the management of stage IB through IVM0 gastric cancer. Moreover, almost all the toxic effects were manageable; only 1 percent of this group of patients, who had good performance status, normal major-organ function, and adequate nutrition, died as a result of a toxic effect related to treatment.
Dr. Atkins raises issues of statistical design and interpretation. He notes the importance of type II errors in interpreting results. It is for this reason that we did not present the results of subgroup analyses. Small numbers within subgroups have poor power to detect clinically meaningful differences. We performed an overall test for the interaction between treatment and base-line variables. An interaction would indicate that the hazard ratio for a treatment effect differs according to stratum. There were no detectable interactions in these data.
Atkins also questions the use of the Cox proportional-hazards model based on unadjusted survival curves. Our formal analyses included an adjustment for stratification factors. The features of Kaplan–Meier curves alone are not sufficient to indicate whether the assumptions of the proportional-hazards model have been violated. Atkins attempts to provide confidence intervals for the three-year survival rate, assuming complete three-year follow-up information for all patients. However, some patients were lost to follow-up, and for some information was not complete, so follow-up information on these patients was censored before the end of the three-year period. Therefore, the formulas Dr. Atkins used are not correct.
Drs. Berney and Merrett comment on the extent of surgery performed. The surgical treatment of our patients1 reflects the current pattern of surgical care in North America.2,3 We agree that D0 lymph-node dissection provides inadequate staging information. However, irrespective of the adequacy of staging, our data demonstrate that postoperative chemoradiotherapy improves disease-free and overall survival after gastrectomy. Berney and Merrett also note differences in survival between our study and a Dutch study.4 Survival in the surgery-only (control) group in our clinical trial was identical to that in large, registry-based, American data bases.2,5 Therefore, the benefit of adjuvant treatment cannot be ascribed to inferior survival in the surgery-only group. In addition, our patients had more advanced cancer than the patients in the Dutch study: 62 percent of the patients in our study had T3 tumors, and 85 percent had positive nodes. Of the patients in the Dutch cohort, 26 percent had T1 tumors, 47 percent T2 tumors, and 25 percent T3 tumors, and only 55 percent had positive nodes.4
5 ReferencesJohn S. Macdonald, M.D.
St. Vincent's Comprehensive Cancer Center, New York, NY 10010
jmacdona@salick.com Jacqueline Benedetti, Ph.D.
Southwest Oncology Group Statistical Center, Seattle, WA 98104Scott A. Hundahl, M.D.
University of Hawaii, Honolulu, HI 968131
Wanebo HJ ,Kennedy BJ ,Chmiel J ,Steele G Jr ,Winchester D ,Osteen R . Cancer of the stomach: a patient care study by the American College of Surgeons. Ann Surg 1993;218:583-592
CrossRef | Web of Science | Medline2
Wanebo HJ ,Kennedy BJ ,Winchester DP ,Fremgen A ,Stewart AK . Gastric carcinoma: does lymph node dissection alter survival? J Am Coll Surg 1996;183:616-624
Web of Science | Medline3
Hundahl SA, Macdonald JS, Benedetti J, et al. Surgical treatment variation in a prospective, randomized trial of chemoradiotherapy in gastric cancer: the impact of under-treatment. Ann Surg Oncol (in press).
4
Bonenkamp JJ ,Hermans J ,Sasako M ,van de Velde CJH . Extended lymph-node dissection for gastric cancer. N Engl J Med 1999;340:908-914
Full Text | Web of Science | Medline5
Hundahl SA ,Phillips JL ,Menck HR . The National Cancer Data Base Report on poor survival of U.S. gastric carcinoma patients treated with gastrectomy: Fifth Edition American Joint Committee on Cancer staging, proximal disease, and the “different disease“ hypothesis. Cancer 2000;88:921-932
CrossRef | Web of Science | Medline
- Citing Articles (4)
Citing Articles
1
Kevin R. Kozak, John S. Moody. (2008) The Survival Impact of the Intergroup 0116 Trial on Patients With Gastric Cancer. International Journal of Radiation Oncology*Biology*Physics 72:2, 517-521
CrossRef2
S. Cascinu, R. Labianca, C. Barone, A. Santoro, C. Carnaghi, A. Cassano, G. D. Beretta, V. Catalano, O. Bertetto, S. Barni, L. Frontini, E. Aitini, S. Rota, V. Torri, I. Floriani. (2007) Adjuvant Treatment of High-Risk, Radically Resected Gastric Cancer Patients With 5-Fluorouracil, Leucovorin, Cisplatin, and Epidoxorubicin in a Randomized Controlled Trial. JNCI Journal of the National Cancer Institute 99:8, 601-607
CrossRef3
Henk H. Hartgrink, Cornelis J.H. van de Velde. (2005) Status of extended lymph node dissection: Locoregional control is the only way to survive gastric cancer. Journal of Surgical Oncology 90:3, 153-165
CrossRef4
Jacques Wils. (2003) Perspectives in adjuvant gastric cancer therapy. Archive of oncology 11:3, 153-154
CrossRef
