|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology
Adequate Lymph Node Staging Is Fundamental to Comparative Study on Resectable Non–Small-Cell Lung CancerThe Chinese University of Hong Kong, Hong Kong, China To the Editor:Depierre et al1 have reported a very important study comparing preoperative chemotherapy followed by surgery with surgery alone in the management of early-stage (stages I, II, and IIIa) non–small-cell lung cancer. The impressive difference they observed in median survival between the two groups (37 v 26 months) was not statistically significant. Only after stratification by stage did they find patients with N0 and N1 disease to have a significant reduction in the relative risk of death (0.68, P = .027), but contrary to previously published phase III trials, no benefit was observed in patients with N2 disease. They concluded that the only significant survival differences were confined to patients with stage I and II disease. However, significant discrepancy between clinical and postoperative pathologic staging of mediastinal lymph nodes would make the current conclusion less credible. Comparison of the clinical and pathologic node status was available only on the control (surgery alone) arm. Twenty-nine patients (24%) with clinical stage N0, N1 disease proved to have pathologic N2 disease, while 22 patients (44%) with clinical N2 disease had pathologic N0, N1 disease. Survival benefit from preoperative chemotherapy could have been attributable to the inferior survival outcome of the large proportion of pathologic N2 disease among clinical stage I and II patients in the control arm. Conversely, lack of efficacy of preoperative chemotherapy among stage IIIa patients could be explained by the "better" survival outcome of the 44% of patients with pathologic N0, N1 disease. Subgroup analysis according to stage becomes less meaningful if the "subgroup" is not a true homogenous population that represents the disease status. Furthermore, the authors used a lymph node size of 1 cm or greater in the short-axis diameter on computed tomography (CT) image as the criterion for disease involvement, while mediastinoscopy was permitted but not routinely performed. This practice is suboptimal for definition of N2 disease in comparative study. Recent reports have shown the weighted-average sensitivity and specificity from comparative studies of mediastinal staging of lung cancer to be 63% and 80%, respectively, for CT and 88% and 93% for positron emission tomography.2 In current study, the sensitivity and specificity of CT scanning for detection of N2 disease were 49% and 75.6%, respectively (calculation based on data from Table 4 of Depierre et al1). Therefore, limitations in their staging technique may explain the discrepancy between clinical and pathologic staging. Mediastinoscopy or mediastinotomy with biopsy in all radiologic N2 disease would have minimized this problem. Both randomized studies that showed survival benefit with preoperative chemotherapy in stage IIIa patients used mediastinoscopy or mediastinotomy to confirm histologic N2 disease before enrollment onto the trial.3,4 Depierre et al1 gave two explanations as to why their results conflict with previous trials on preoperative chemotherapy in stage IIIa patients, but they neglected to mention the most important reason—patient selection. If they had adopted the same selection criteria as used in the previous studies that used histologic N2 as an enrollment criterion, they would have had to exclude 44% of patients from the control arm. It would be difficult to estimate the magnitude of discrepancy in the study arm because preoperative chemotherapy would have eradicated microscopic metastasis from a proportion of patients. Comparison of survival outcome between two groups of patients with unconfirmed and possibly unmatched N2 status does not help to establish the role of preoperative chemotherapy. Short of confirmed N2 status at enrollment, it is not justified to conclude that there is a lack of efficacy of preoperative chemotherapy in patients with stage IIIa non–small-cell lung cancer. Finally, the authors identified nonproportional hazards over time where the instantaneous risk of death is higher in the preoperative chemotherapy arm during the first few months and the instantaneous risk of death is higher in the control arm thereafter. Although the authors did not examine the nonproportionality further by stratification on stage, it is very likely that the nonproportionality occurs only in the N2 disease subgroup. The N2 survival curves in Fig 2 of the article suggest a difference in favor of the control arm at the early part of the survival comparison and a difference in favor of preoperative chemotherapy during the later part of the survival comparison. A more detailed listing and analysis of the causes of deaths during the two time periods would be helpful in interpreting the results of this study. REFERENCES
1.
Depierre A, Milleron B, Moro-Sibilot D, et al: Preoperative chemotherapy followed by surgery compared with primary surgery in resectable stage I (except T1N0), II, and IIIa non–small-cell lung cancer. J Clin Oncol 20: 247-253, 2002
2.
Lowe VJ, Naunheim KS: Positron emission tomography in lung cancer. Ann Thorac Surg 65: 1821-1829, 1998
3.
Roth JA, Fossella F, Komaki R, et al: A randomized trial comparing perioperative chemotherapy and surgery with surgery alone in resectable stage IIIA non-small-cell lung cancer. J Natl Cancer Inst 86: 673-680, 1994
4.
Rosell R, Gomez-Codina J, Camps C, et al: A randomized trial comparing preoperative chemotherapy plus surgery with surgery alone in patients with non-small-cell lung cancer. N Engl J Med 330: 153-158, 1994
Response
J. Minjoz University Hospital, Besançon In Reply:The authors of these two letters made the same comment on the need for a mediastinoscopy before inclusion in a phase III trial of preoperative chemotherapy. We agree with their theoretical arguments. In our study,1 we made the choice of a pragmatic attitude because of the very conflicting opinions of investigators concerning systematic mediastinoscopy (especially in clinical stage I and II). Guidelines from the American Society of Clinical Oncology,2 American Thoracic Society, and European Respiratory Society3 recommend mediastinoscopy only in patients with clinical N2 evident on computed tomography (CT) scans (short-axis diameter > 1 cm). However, many institutions do not perform this procedure for each patient before surgery, as they claim that mediastinoscopy is inefficient for numerous mediastinal lymph node chains (sites 2L, 5L, 6L, 7, 8, 9L, and R). Therefore, the question we have addressed was: "Is preoperative chemotherapy useful in clinically resectable tumors?" This is a useful question to be answered for an everyday practice. Patients were randomized before receiving any treatment; therefore, the rate of minimal N2 (N0 to N1 on CT scan) and of false clinical N2 should be well balanced between the two groups of patients. In the control arm (primary surgery), sensitivity and specificity of CT scan were within the limits reported in the literature. This study was therefore correctly designed to answer the above-cited question. We clearly cannot make definitive conclusions from subgroup analyses on the interest of preoperative chemotherapy in N0, N1 disease on the one hand and in N2 disease on the other hand. However, it can be a sufficient argument for recommending the separate study of early stages and stage IIIa disease in upcoming trials of preoperative treatments. Mok and Zee think it is very likely that the risk of death during the treatment period is only increased in clinical N2 disease. They are correct, especially for postoperative deaths in N2 patients treated with chemotherapy. This point was presented during the 2001 annual meeting of the American Society of Clinical Oncology.4 Estimated risks of death in both randomized arms were not proportional over time. During the treatment period, which was the first 5 months, the relative risk was 2.28 (95% confidence interval, 0.84 to 6.23) for N2 disease and 0.81 (95% confidence interval, 0.31 to 2.14) for N0 to N1 non–small-cell lung cancer, but differences were not statistically significant in any group. Within the first 30 postoperative days and thereafter, there were four lethal complications in the control arm and 14 in the preoperative chemotherapy arm for N2 disease, and five and four, respectively, for N0 to N1 disease. REFERENCES 1. Depierre A, Milleron B, Moro-Sibilot D, et al: Preoperative chemotherapy followed by surgery compared with primary surgery in resectable stage I (except T1N0), II, and IIIa non–small-cell lung cancer. J Clin Oncol 20: 247-253, 2002 2. American Society of Clinical Oncology: Clinical practice guidelines for the treatment of unresectable non–small-cell lung cancer. J Clin Oncol 15: 2996-3018, 1997[Abstract]
3.
American Thoracic Society: Pretreatment evaluation of non-small-cell lung cancer. Am J Respir Crit Care Med 156: 320-332, 1997 4. Breton JL, Westeel V, Milleron B, et al: Postoperative morbidity and mortality after neoadjuvant chemotherapy (NCT) for non small cell lung cancer (NSCLC) in the French phase III trial. Proc Am Soc Clin Oncol 20: 311a, 2001 (abstr 1239)
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2002 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
