| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1998;65:220-226
© 1998 The Society of Thoracic Surgeons
Departments of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada
Department of Medicine, McMaster University, Hamilton, Ontario, Canada,
Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York, USA
Accepted for publication September 23, 1997.
Dr Guyatt, Department of Clinical Epidemiology and Biostatistics, McMaster University, Health Sciences Center, Room 2C12, 1200 Main St W, Hamilton, ON L8N 3Z5, Canada (e-mail: guyatt@fhs.csu.mcmaster.ca).
 |
Abstract |
|---|
 TopFootnotes Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
Methods. Two reviewers independently examined the surgical records and pathologic reports from a randomized trial comparing computed tomography versus mediastinoscopy for staging of lung cancer. They recorded pathologic findings at the surgical resection margin, the highest mediastinal nodal station sampled at thoracotomy, histologic type, tumor size, N status, and evidence of vascular or lymphatic invasion. These variables formed the independent variables in logistic regression models to predict recurrence.
Results. Except for 1 patient, follow-up at 3 years for 399 included patients was complete. Significant predictors of recurrence were tumor size (odds ratio [OR], 1.2 (per centimeter); 99% CI [confidence interval], 1.1 to 1.4), and N status (compared with N0, N1: OR, 1.6; CI, 0.8 to 3.1; N2: OR, 3.2; CI, 1.4 to 7.5). Other variables, including positive surgical resection margin, did not predict early recurrence or death.
Conclusions. In patients with non–small cell lung cancer, surgical resection margin or highest nodal station sampled at thoracotomy that are involved by carcinoma do not predict recurrence. The current definition of incomplete resection has limited prognostic significance.
|
Introduction |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
Much uncertainty remains about the prognosis and management of an incompletely resected non–small cell lung cancer (NSCLC). This is partly related to the controversy in the current surgical literature around the definition of "incomplete resection" [8]. Macroscopic or microscopic residual disease at the resection margin has been treated as representing histologic proof that the surgical resection was incomplete [7][9][10]. Many investigators also consider the presence of tumor in the highest mediastinal node sampled at thoracotomy as evidence of incomplete tumor resection [7][10]. Others include perinodal extension as evidence of an incomplete resection [10].
The purpose of classifying patients with respect to completeness of resection is that it provides prognostic information, and thus can guide management. For instance, trials of radiotherapy or chemotherapy might target patients at high risk of recurrence. The power of features such as tumor at the resection margin or in the highest mediastinal node for predicting recurrence after resection has not been well established [8]. If the current definitions of incomplete resection fail to convey prognostic information, they become tautological, and of little use.
In this study, we examined two hypotheses: first, that clinicians’ decisions to offer patients adjuvant therapy is influenced by tumor pathologic characteristics, and among these characteristics are those that currently define incomplete resection; second, that these same pathologic features exert a powerful prognostic significance on recurrent disease.
|
Material and Methods |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
Data Extraction
Two investigators independently reviewed the surgical record and pathologic reports of each patient who had undergone thoracotomy with surgical resection of a lung carcinoma. The reviewers were blinded to the clinical condition of the patients before the operation and to their final vital status (recurrence or death) at 3-year follow-up. Extracted data included (1) surgical procedure (tumor resection versus exploratory thoracotomy alone); (2) tumor size; (3) histologic type; (4) pathologic characteristics at the resection margin (normal tissue, metaplasia, dysplasia, or carcinoma); (5) pathologic description of lymphatic or vascular invasion, microscopic necrosis, cavitation, or hemorrhage within the tumor; and (6) a complete mapping of the sampled mediastinal nodes at thoracotomy [12], including the pathologic characteristics of the highest sampled mediastinal nodal station. Data extractors classified patients with a positive surgical resection margin at the bronchial stump according to the histologic location of the residual disease [13]. We defined "synchronous tumors" as (1) two or more tumors with different histologic types, or (2) two or more histologically similar tumors originating from separate segments, lobes, or lungs, and having no region with common lymphatic involvement [14]; and "satellite nodules" as accessory carcinomatous foci adjacent to but separated from the main tumor [15] and sharing common lymphatic drainage with the main tumor. Final surgical staging [4][16] was established on the basis of both the surgical and pathologic reports. When reports from the thoracic surgeon and the pathologist conflicted regarding the size of the tumor and the mediastinal nodal stations that were sampled, we considered only the pathologist’s report.
After completing data extraction from the surgical records and pathologic reports, we conducted a separate chart review to determine whether the patients received adjuvant therapy (chemotherapy or radiotherapy or both). Throughout the data extraction process, we initially resolved disagreement between reviewers by consensus or, if this failed, by consulting a third reviewer. An adjudication committee had previously ascertained the patients’ final outcome at 3-year follow-up (recurrence, vital status, and cause of death) [11].
We considered pathologic characteristics of the tumor that were not explicitly mentioned as negative. This decision was based on an informal survey of 11 pathologists who had initially assessed the resection specimens at the time of the thoracotomy. We contacted at least one pathologist from each of the centers participating in the trial. Seven pathologists (63.6%) stated that they do not systematically look for presence of lymphatic or vascular invasion, but would describe such findings if observed. The pathologists from the four remaining centers stated that they systematically look for both lymphatic and vascular invasion. However, in none of these centers is there any established policy as to how to consistently report the findings.
Statistical Analysis
To assess predictors of adjuvant therapy, we used the pathologic characteristics extracted from the patients’ records as independent variables in a unrestricted stepwise multivariable logistic regression analysis, with adjuvant therapy as the dependent variable.
To assess predictors of recurrence, we considered, throughout the analysis, adjuvant therapy as an important potential confounding factor in predicting recurrence. Indeed, the finding of a positive surgical resection margin or a positive highest nodal station sampled at thoracotomy may prompt clinicians to offer the patients adjuvant therapy (chemotherapy, radiotherapy, or both), thus obscuring the natural contribution of these pathologic characteristics to the prognosis. Thus, all our analyses to predict recurrence were adjusted for adjuvant therapy.
We first conducted univariate logistic regression analyses using each of the possible predictors in turn as the independent variable and recurrence as the dependent variable. With respect to positive resection margin we did analyses including all patients, and then an analysis in which we excluded patients with a positive resection margin with carcinoma in situ, those with a positive venous resection margin, and those with mucosal involvement. Patients with recurrence included those with early recurrence and subsequent death as well as those with recurrence who survived at 3-year follow-up. For each independent variable, we also tested its interaction with adjuvant therapy to verify whether adjuvant therapy acted as an effect modifier. Then, in the multivariable analysis, we successively incorporated the predictors into a logistic regression model according to a priori hypotheses regarding the strength of association of each variable and the outcome (recurrence). The order of entry of these variables in the model was as follows: (1) tumor size; (2) N status; (3) tumor histologic type; (4) synchronous or satellite tumors; (5) surgical resection margin; (6) highest sampled nodal mediastinal station; (7) vascular invasion within the main tumor; (8) lymphatic invasion within the tumor; and (9) tumor necrosis or hemorrhage.
We used tumor size rather than T status in the model because we believe that the former better reflects neoplastic burden, especially for T3 and T4 tumors. Again, at each step of the model, we also tested the predictor by adjuvant therapy interaction. At each step, the variables that did not explain a statistically significant proportion of the variance were excluded from the model. An unrestricted multivariable logistic regression analysis, in which none of the variables was forced in the model, was also conducted using the same predictors. Because of the multiple comparisons involved, we defined statistical significance as p less than 0.01. We present the odds ratios (ORs) with their 99% confidence intervals (CIs).
|
Results |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
Of the remaining 417 patients we excluded an additional 18 because of an incomplete mediastinal sampling at thoracotomy that made the pathologic assessment of mediastinal disease uncertain. In the original trial, these 18 patients were all randomized to computed tomographic scan and had limited resection. Thus, we included 399 patients in the final analysis.
Table 1 presents the clinical characteristics of the patients who underwent thoracotomy. Two hundred patients had recurrence at 3-year follow-up, whereas 199 had not.
|
Predicting Recurrence
Table 2
summarizes the distribution of the patients according to the pathologic status of their resection margin and highest node samples at thoracotomy, and whether they had received adjuvant therapy. The surgical resection margin was positive for carcinoma in 25 patients (6.3%). Twenty of these 25 had a positive bronchial resection margin (peribronchial involvement [n = 14, of whom 9 had recurrence]; carcinoma in situ [n = 3, of whom 2 had recurrence]; lymphatic permeation [n = 2, both had recurrence]; mucosal involvement [n = 1, patient remained disease-free]); 4 patients had residual involvement on the chest wall, of whom 3 had recurrence; and one, who remained disease-free, had a positive venous resection margin. Thus, 16 of 25 patients had a recurrence; 10 of the 25 patients received adjuvant therapy.
|
Table 3 presents the risk of recurrence within 3 years of follow-up associated with each independent variable in both univariate and multivariable analyses. In both the univariate and multivariable analyses, none of the interaction terms was significant, suggesting that adjuvant therapy did not influence the value of any of the predictors of recurrence (that is, the predictors were equally powerful whether or not patients had adjuvant therapy). When we excluded patients with carcinoma in situ, mucosal involvement, or a positive venous resection margin the odds ratio associated with a positive resection margin rose, but still failed to reach statistical significance (2.21; 99% CI, 0.56 to 8.66; p = 0.14).
|
|
Comment |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
The strengths of our study in determining predictors of prognosis include the prospective collection of the data, the careful blinded review of the clinical and pathologic data, and the virtually complete follow-up. The duration of follow-up could be seen as either a strength (3 years of follow-up is relatively long) or a weakness (5 years would have been better). The sample size, although large, yielded only small numbers of patients with positive resection margins or positive highest mediastinal node. As a result confidence intervals around the estimates of risk are relatively wide, and we do not have adequate numbers to examine the prognostic impact of different bronchial resection margin positive subtypes (such as mucosal versus submucosal disease). Another limitation of our study is the necessary assumption that if pathologic characteristics were not mentioned, they were not present.
With respect to this last limitation, if pathologic reports did not actually mention any positive or negative findings related to lymphatic invasion, vascular invasion, or necrosis or hemorrhage within the tumor, we assumed these findings were absent. Investigators looking specifically for arterial invasion reported the finding in 58% of 113 consecutive resected primary lung carcinomas, a higher proportion than in our sample [17]. It is possible that lymphatic or arterial invasion, or necrosis or hemorrhage within the tumor may have been stronger predictors if sought more rigorously by the pathologists.
Table 4 summarizes other investigations of the prognosis of patients with NSCLC with a positive proximal resection margin at thoracotomy [18][19][20][21][22][23][24][25][26]. The investigators have consistently documented a substantial proportion (all 0.20 or more) long-term survivors among affected patients. Most studies did not include any control group, however, and only one [26] directly addressed the issue of contribution of adjuvant therapy to long-term outcome as we did. Those studies with control groups showed only a weak relation between positive resection margin and survival.
|
A positive resection margin meets the first criterion well. It makes intuitive sense that if the resection margin is positive, there will be malignant cells on the other side of the margin and that these cells will grow and spread after thoracotomy. We found, however, that 9 of 25 patients with a positive resection margin were disease-free at 3 years, and that the proportions of patients with a positive resection margin differed little in those who did and did not recur. These results are consistent with the prior literature. Positive resection margin did not reach statistical significance as a predictor of recurrence, even in our univariate analysis. Positivity of the resection margin fails the second criterion for a definition of an incomplete resection.
The other pathologic feature, positivity of the highest mediastinal node, is much weaker from the biological viewpoint. This criterion did not reach conventional levels of statistical significance in either the univariate or the multivariate analysis. We conclude that positivity of the highest mediastinal node also fails the criteria for a definition of incomplete resection.
The second implication of our results has to do with settings in which investigators wish to select patients at high risk for recurrence. For instance, reasons that it would be appropriate to test new adjuvant lung cancer treatments in those at highest risk of recurrence include the ethical mandate to subject only those with substantial risk to toxic experimental regimens, and the importance of enrolling high-risk patients to minimize the required sample size.
These considerations are important in designing studies of new therapy for high-risk patients [7]. Studies that enroll patients on the basis of a traditional definition of high risk focused on positive resection margin will not achieve optimal efficiency in recruiting those at highest risk of recurrence. Investigators could do far better in achieving the goal of enrolling highest risk patients by targeting those with larger tumor size and higher N status.
We conclude that traditional definitions of incomplete resection are problematic. This conclusion is based on more than the patients in this trial. As we have described, prior studies are consistent in finding positive resection margin a weak predictor of recurrence (see Table 4). We believe that a positive resection margin has persisted as a definition of incomplete resection because, despite the lack of supporting data regarding its predictive power, the intuitive biologic rationale is compelling. In combination with previous studies, the current data strongly suggest that, despite their intuitive appeal, it is time to reconsider traditional definitions of incomplete resection.
|
Acknowledgments |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
|
Footnotes |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
|
Appendix 1 |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
Study Coordination: Gordon H. Guyatt, MD, Deborah J. Cook, MD, Michael S. Lefcoe, MD, Timothy L. Winton, MD, Jean Deslauriers, MD, Richard I. Inculet, MD, Farid M. Shamji, MD, Michael Coughlin, MD, and Susan M. Troyan, BA
Writing: Gordon H. Guyatt, MD, Deborah J. Cook, MD, Stephen Walter, PhD, Robert J. Ginsberg, MD, Bernie J. O’Brien, PhD, Thomas R. J. Todd, MD, and Richard J. Eddy, MD
Economic Analysis: Ron A. Goeree, MA, and Bernie J. O’Brien, PhD
Adjudication: W. Fred Bennett, MD, Stephen Walter, PhD, Jemi Olak, MD, Deborah J. Cook, MD, Gordon H. Guyatt, MD, Farid M. Shamji, MD, Gail E. Darling, MD, and Susan M. Troyan, BA
Methods Center: Deborah J. Cook, MD, Gordon H. Guyatt, MD, Stephen Walter, PhD, Susan M. Troyan, BA, Lauren E. Griffith, MS, Deborah L. Maddock, Suzanne J. Duchesne, Barbara Jedrzejowski, and Sandi M. Harper
Radiology Coordination: Michael S. Lefcoe
Clinical Centers
Hamilton: Geoffrey Evans, MD, W. Fred Bennett, MD, Carl Zylak, MD, and Susan M. Troyan, BA
Toronto: Timothy J. Winton, MD, Robert J. Ginsberg, MD, Melvyn Goldberg, MD, G. Alec Patterson, MD, F. Griffith Pearson, MD, Donald P. Jones, MD, Robert A. Zeldin, MD, Gary L. Stoik, MD, Hensley A. B. Miller, MD, Glen A. Taylor, MD, Martin F. McKneally, MD, PhD, Stephen J. Herman, MD, Gordon L. Weisbrod, MD, Marvin Steinhardt, MD, Robert Bruce, MD, Susan M. Rosgen, BScN, Jane Flett, RN, Leslie Steward-Pichette, BScN, Ann Fedyk, BScN, and Kathie Roche, BScN
Quebec City: Jean Deslauriers, MD, Michel Piraux, MD, Guy Carrier, MD, Sylvie Ferland, MD, Brigitte Fournier, BScN, Carolle St. Pierre, BScN, and Pierre Bedard, BScN
London: Richard I. Inculet, MD, Richard J. Finley, MD, Michael S. Lefcoe, MD, and Natalie Zankowicz, BScN
Ottawa: Farid M. Shamji, MD, Harold J. Sachs, MD, Thomas R. J. Todd, MD, Rebecca Peterson, MD, Nancy M. Hickey, MD, and Diane C. Lister, RNA
Windsor: Michael Coughlin, MD, Craig R. Pierce, MD, Dermot R. Feore, MD, Peter Tadros, MD, and Judy Imeson, BScN
|
References |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsAppendix 1References |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Riquet, P. Bagan, F. Le Pimpec Barthes, E. Banu, F. Scotte, C. Foucault, A. Dujon, and C. Danel Completely Resected Non-Small Cell Lung Cancer: Reconsidering Prognostic Value and Significance of N2 Metastases Ann. Thorac. Surg., December 1, 2007; 84(6): 1818 - 1824. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Wind, E. J. Smit, S. Senan, and J.-P. Eerenberg Residual disease at the bronchial stump after curative resection for lung cancer Eur. J. Cardiothorac. Surg., July 1, 2007; 32(1): 29 - 34. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Vesselle, J. D. Freeman, L. Wiens, J. Stern, H. Q. Nguyen, S. E. Hawes, P. Bastian, A. Salskov, E. Vallieres, and D. E. Wood Fluorodeoxyglucose Uptake of Primary Non-Small Cell Lung Cancer at Positron Emission Tomography: New Contrary Data on Prognostic Role Clin. Cancer Res., June 1, 2007; 13(11): 3255 - 3263. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Sakao, H. Miyamoto, A. Yamazaki, T. Oh, R. Fukai, K. Shiomi, and Y. Saito Prognostic Significance of Metastasis to the Highest Mediastinal Lymph Node in Nonsmall Cell Lung Cancer Ann. Thorac. Surg., January 1, 2006; 81(1): 292 - 297. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Balasubramanian, J. Au, and J. Dunning Should lobectomy patients with microscopic involvement of the bronchial resection margin undergo re-operation to improve their long-term survival? Interactive CardioVascular and Thoracic Surgery, December 1, 2005; 4(6): 531 - 537. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Pasic, K. Grunberg, W. J. Mooi, M. A. Paul, P. E. Postmus, and T. G. Sutedja The Natural History of Carcinoma In Situ Involving Bronchial Resection Margins Chest, September 1, 2005; 128(3): 1736 - 1741. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Guo, M. G. House, C. Hooker, Y. Han, E. Heath, E. Gabrielson, S. C. Yang, S. B. Baylin, J. G. Herman, and M. V. Brock Promoter Hypermethylation of Resected Bronchial Margins: A Field Defect of Changes? Clin. Cancer Res., August 1, 2004; 10(15): 5131 - 5136. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Refaely, S. Sadetzki, A. Chetrit, D. A. Simansky, M. Paley, B. Modan, and A. Yellin The sequence of vessel interruption during lobectomy for non-small cell lung cancer: Is it indeed important? J. Thorac. Cardiovasc. Surg., June 1, 2003; 125(6): 1313 - 1320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Lequaglie, B. Conti, P.P. Brega Massone, and G. Giudice Unsuspected residual disease at the resection margin after surgery for lung cancer: fate of patients after long-term follow-up Eur. J. Cardiothorac. Surg., February 1, 2003; 23(2): 229 - 232. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Higashiyama, K. Kodama, K. Takami, N. Higaki, T. Nakayama, and H. Yokouchi Intraoperative lavage cytologic analysis of surgical margins in patients undergoing limited surgery for lung cancer J. Thorac. Cardiovasc. Surg., January 1, 2003; 125(1): 101 - 107. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Brundage, D. Davies, and W. J. Mackillop Prognostic Factors in Non-small Cell Lung Cancer* : A Decade of Progress Chest, September 1, 2002; 122(3): 1037 - 1057. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Lopez-Encuentra, J. L. Duque-Medina, R. Rami-Porta, A. G. de la Camara, and P. Ferrando Staging in Lung Cancer: Is 3 cm a Prognostic Threshold in Pathologic Stage I Non-small Cell Lung Cancer? : A Multicenter Study of 1,020 Patients Chest, May 1, 2002; 121(5): 1515 - 1520. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G J M Herder, P Verboom, E F Smit, P C M van Velthoven, J H A M van den Bergh, C D Colder, I van Mansom, J C van Mourik, P E Postmus, G J J Teule, et al. Practice, efficacy and cost of staging suspected non-small cell lung cancer: a retrospective study in two Dutch hospitals Thorax, January 1, 2002; 57(1): 11 - 14. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Passlick, I. Sitar, W. Sienel, O. Thetter, and A. Morresi-Hauf Significance of lymphangiosis carcinomatosa at the bronchial resection margin in patients with non-small cell lung cancer Ann. Thorac. Surg., October 1, 2001; 72(4): 1160 - 1164. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Massard, C. Doddoli, B. Gasser, X. Ducrocq, R. Kessler, C. Schumacher, G.-M. Jung, and J.-M. Wihlm Prognostic implications of a positive bronchial resection margin Eur. J. Cardiothorac. Surg., May 1, 2000; 17(5): 557 - 565. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Asamura Complete resection: Yes or no? J. Thorac. Cardiovasc. Surg., January 1, 2000; 119(1): 193 - 194. [Full Text]
|
|
|
|
|
|
N. Sawabata, T. Mori, K. Iuchi, H. Maeda, M. Ohta, and O. Kuwahara CYTOLOGIC EXAMINATION OF SURGICAL MARGIN OF EXCISED MALIGNANT PULMONARY TUMOR: METHODS AND EARLY RESULTS J. Thorac. Cardiovasc. Surg., March 1, 1999; 117(3): 618 - 619. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
