Lobectomy Versus Sublobar Resection for Small (2 cm or Less) Non Small Cell Lung Cancers

Transcription

1 Lobectomy Versus Resection for Small (2 cm or Less) Non Small Cell Lung Cancers Andrea S. Wolf, MD, William G. Richards, PhD, Michael T. Jaklitsch, MD, Ritu Gill, MD, Lucian R. Chirieac, MD, Yolonda L. Colson, MD, PhD, Kamran Mohiuddin, MD, Steven J. Mentzer, MD, Raphael Bueno, MD, David J. Sugarbaker, MD, and Scott J. Swanson, MD Division of Thoracic Surgery, Department of Radiology, and Department of Pathology, Brigham and Women s Hospital, Boston, Massachusetts Background. We evaluated a cohort of patients who underwent resection for small (2 cm or less) non small cell lung cancer (NSCLC) to determine if there is an association between extent of resection (lobar versus sublobar resection) and local recurrence or survival. Methods. We reviewed 468 consecutive patients who underwent resection for small NSCLC at our institution between 2000 and We excluded patients who had neoadjuvant therapy, active noncutaneous malignancies, pure bronchioalveolar carcinoma, lymph node (n 53) or distant metastases at diagnosis, or multicentric cancers. Clinicopathologic data, recurrence, and vital status as of June 15, 2010, were retrieved. Overall and recurrence-free survival from surgery rates were assessed. Results. Two hundred thirty-eight patients underwent resection for primary solitary small NSCLC. Lobectomy (n 84) was associated with longer overall (p ) and recurrence-free (p ) survival. Patients who underwent sublobar resection were older (p < ) and had worse pulmonary function (p < ). While there was a trend toward increased rate of local recurrence for sublobar resection (16% versus 8%, p ), there was no difference in distant recurrence. Moreover, when lymph nodes were sampled with sublobar resection, local recurrence rate and overall and recurrence-free survival distributions were similar to those for lobectomy. Conclusions. resection is reasonable in older patients with limited cardiopulmonary function. For healthy patients, however, lobectomy remains the standard therapy, with sublobar resection with lymph node sampling representing an alternative to consider. These findings support continued effort to conduct a randomized trial of lobar versus sublobar resection, such as CALGB (Ann Thorac Surg 2011;92: ) 2011 by The Society of Thoracic Surgeons Lung cancer is the primary cause of cancer death among men and women in the United States, with 160,000 deaths annually [1]. Stage IA non small cell lung cancer (NSCLC), which in the American Joint Commission on Cancer (AJCC) 6th edition staging system (current at the time the present study cohort was staged) included lesions of 3 cm or less with no nodal or metastatic disease (T1 N0 M0), is associated with the best prognosis, with 5-year survival ranging from 58% to 90% [1-4]. Of the 170,000 new cases of lung cancer annually, approximately 20% of patients are surgical candidates [3]. Although the historic standard for surgical management of stage IA disease has been anatomic lobectomy, a wedge resection with negative margins or a segmentectomy is also performed for peripheral lesions and in patients with poor cardiopulmonary reserve. Whether to treat early-stage NSCLC with lobar or sublobar resection remains controversial. Accepted for publication June 22, Presented at the Forty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31 Feb 2, Address correspondence to Dr Wolf, Division of Thoracic Surgery, Brigham and Women s Hospital, Harvard Medical School, 75 Francis St, Boston, MA 02215; awolf2@partners.org In a multi-institutional retrospective cohort analysis from 1989 to 1994 for stage IA NSCLC, Landreneau and associates [5] found the median time to recurrence was 10 months for wedge resection, compared to 19 months for lobectomy. The rate of local recurrence was slightly higher for patients who underwent wedge resection (compared to those who underwent lobectomy), but this difference was not statistically significant. Although the rate of non cancer-related deaths was higher after wedge resection, the overall 5-year survival was not significantly different between patients who underwent wedge resection and those who underwent lobectomy. In the only prospective randomized controlled trial comparing sublobar and lobar resection (conducted between 1982 and 1988), the Lung Cancer Study Group found that although the mortality rate was slightly higher in the sublobar resection group, this difference was not significant [6]. The study did, however, demonstrate a significantly higher cancer recurrence rate in the sublobar resection group. This finding established lobectomy as the standard surgical approach to the treatment of early-stage lung cancer. With more widespread use of computed tomography (CT) and improvement in CT resolution, NSCLC is being diagnosed earlier and with smaller tumor size than in the 2011 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 1820 WOLF ET AL Ann Thorac Surg RESECTION FOR SMALL LUNG CANCERS 2011;92: Table 1. Demographic, Clinical, and Pathologic Characteristics of Patients Undergoing Resection and Lobectomy Characteristic past [7]. Recent studies have demonstrated that even within stage IA, patients with tumors up to 2 cm experience a better 5-year survival rate than those with tumors between 2 cm and 3 cm [2, 3]. It seems reasonable to hypothesize that among T1 tumors, 2- to 3-cm lesions could represent those more prone to recurrence and that smaller tumors might be successfully managed with less radical resection. We sought to determine whether segmentectomy or wedge resection may be appropriate for the management of small lung tumors, particularly as technology leads to the diagnosis of more such lesions in a continually aging population. We therefore evaluated a cohort of patients who underwent resection for small (2 cm or less) NSCLC to determine if there is an association between extent of resection and rate of local recurrence or survival duration. Patients and Methods (n 154)* Median age, y (range) 71 (41 91) 66 (42 85).0001 Male, n 62 (40%) 35 (42%) Median, % predicted FEV 1 68 (12 141) 82 (32 147) (range) COPD, n 49 (32%) 18 (21%) Central tumor (inner half 29 (23%) 26 (35%) of hemithorax), n (%) Median tumor size, cm 1.5 ( ) 1.5 ( ) (range) Visceral pleural invasion, n 23 (15%) 14 (17%) Squamous (versus 32 (21%) 13 (15%) nonsquamous), n Poorlydifferentiated, n 43 (28%) 19 (23%) Lymph nodes sampled, n 45 (29%) 79 (94%).0001 Median distance to nearest stapled margin (sublobar only), cm (range) 0.6 (0-3.8) * Includes 24 segmentectomies. COPD chronic obstructive pulmonary disease; FEV 1 forced expiratory volume in 1 second. With institutional review board approval, we reviewed the records of all patients who underwent resection for NSCLC tumors of 2 cm or less at our institution between January 1, 2000, and December 31, Of these 468 consecutive patients, we excluded patients who had neoadjuvant therapy, other active noncutaneous malignancies, pure bronchioalveolar carcinomas (BAC), lymph node (n 53) or distant metastases at diagnosis, or multicentric cancers. Electronic and written records were reviewed for clinicopathologic data. Original images (chest radiography and CT) were reviewed by a chest radiologist (R.G.) for designation of inner or outer half location within the hemithorax. All relevant pathologic variables were audited with a comprehensive review and resolution of any ambiguities by a pulmonary pathologist (L.C.). Recurrence was defined as a tumor of the same histology occurring within the same lobe (or draining hilum/mediastinum) or a tumor of the same histology in a different lobe or lung with carcinoma in the lymphatics common to both and/or extraplumonary metastases, as described by Martini and Melamed [8]. Local recurrence was defined as recurrent tumor within the same lobe, hilum, or mediastinum with or without the presence of distant metastases. Vital status as of June 15, 2010, was confirmed through records, obituaries, or the Social Security Death Index. Descriptive statistics, including percentages and medians with ranges, are reported. Distributions were plotted to examine compliance with distributional assumptions and the veracity of any outliers. Nonparametric testing was used as appropriate to compare continuous and dichotomous variables in univariate analysis with 2-sided p values reported. Kaplan-Meier plots were used to illustrate the cumulative proportion surviving as a function of months since surgery with predictors of survival compared using the log rank test. Observations were censored at date of last contact or close of data collection at June 15, Overall (living patients censored at last contact) and recurrence-free (other-cause death censored) survival from surgery to death or radiologic recurrence was assessed. Lobectomy and sublobar resection were compared for the primary endpoint of local recurrence and secondary endpoints of overall and recurrence-free survival. Segmentectomies were grouped with wedge resections as there were too few segmentectomies (n 24) in this study to draw meaningful independent conclusions when analyzed separately. Moreover, segmentectomies have been categorized with wedge resections in both the Lung Cancer Study Group [6] trial and the current prospective trial CALGB Because the sample sizes and number of Table 2. Complications by Type of Surgery Complication (n 154)* Perioperative mortality, n 1 ( 1%) 3 (4%) Atrial fibrillation, n 9 (6%) 8 (10%) Pulmonary embolus, n 0 1 (1%) Myocardial infarction, n 1 ( 1%) 1 (1%).9999 Airleak 7 days, n 7 (5%) 9 (11%) Tracheostomy, n 0 1 (1%) Feeding tube, n 0 1 (1%) Adult respiratory distress 2 (1%) 2 (2%) syndrome, n Pneumonia, n 4 (3%) 4 (5%) Bronchopleural fistula, n 0 1 (1%) Chyle leak, n 0 2 (2%) Reoperation for bleeding, n 2 (1%) Acute renal failure (none 2 (1%) 2 (2%) required hemodialysis), n Any complication, n 22 (14%) 18 (21%) * Includes 24 segmentectomies.

3 Ann Thorac Surg WOLF ET AL 2011;92: RESECTION FOR SMALL LUNG CANCERS 1821 Table 3. Univariate Analysis Table 4. Stratified Analysis Outcome (n 154) Outcome (n 154) Recurrence, n Local 25 (16%) 7 (8%) Distant 23 (15%) 11 (13%) Any 37 (24%) 13 (15%) Survival 5-y survival, % Median overall 6.9 Not reached survival, y 5-y recurrence-free survival, % Median recurrence-free Not reached Not reached survival, y events were too few to support a stable multivariable regression model, stratified analyses were performed to evaluate the relationships between age, tumor size, and lymph node sampling and the effect of type of resection on local recurrence. The sublobar resection cohort was explored further with survival analysis stratified by lymph node sampling. Results There were 238 patients who underwent resection for primary solitary NSCLC of 2 cm or less at our institution between 2000 and 2005 and who met study criteria for inclusion. There were 154 sublobar resections, including 24 segmentectomies. There were 66 (43%) sublobar resections performed via a video-assisted thoracoscopic (VATS) approach versus 8 (10%) lobectomies (p ). Compared to patients who underwent lobectomy, patients who underwent sublobar resection were older and had worse preoperative pulmonary function (Table 1). Although the proportion of central tumors trended to be higher among patients who underwent lobectomy, the difference did not reach statistical significance. Patients who had sublobar resection were significantly less likely to have lymph nodes sampled at the time of surgery. Other clinicopathologic variables were distributed similarly for the 2 cohorts. Moreover, mortality and morbidity Recurrence by node sampling patients with lymph nodes sampled, n patients without lymph nodes sampled, n Recurrence by age patients under age 70, n patients age 70 and older, n Recurrence by tumor size patients with tumors 1.5 cm patients with tumors 1.5 cm 5 (11%) 7 (8%) (18%) (13%) 4 (7%) (19%) 3 (12%) (18%) 5 (11%) (14%) 2 (5%) were not statistically different between the 2 groups (Table 2). The rates of distant recurrence and any recurrence (defined as local, distant, or both) were similar for the 2 cohorts, although lobectomy was associated with a trend toward fewer local recurrences compared to sublobar resection (Table 3). Lobectomy was associated with significantly longer overall and recurrence-free survival (Fig 1; Table 3). There was no association between the development of a second primary lung cancer and type of resection (n 6, 7% for lobectomy; n 14, 9% for sublobar resection; p ). When stratified by tumor size, the local recurrence rates for patients with sublobar resection and lobectomy were similar for tumors smaller versus those larger than the median. Similarly, when stratified by age, patients younger versus those older than the median had similar rates of local recurrence whether they underwent sublobar or lobar resection (Table 4). When stratified by whether lymph nodes were sampled (with mediastinoscopy or intraoperatively), patients who underwent sublobar resection with Fig 1. Cumulative survival distribution for overall survival (A) and recurrence-free survival (B) grouped by type of resection.

4 1822 WOLF ET AL Ann Thorac Surg RESECTION FOR SMALL LUNG CANCERS 2011;92: Fig 2. Cumulative survival distribution for overall survival for lobectomy and sublobar resection, the latter stratified by lymph node sampling. lymph nodes sampled (n 45, 29%) had a local recurrence rate similar to that of patients who underwent lobectomy (Table 4). In contrast, patients who underwent sublobar resection without lymph nodes sampled (n 109, 71%) had a local recurrence rate more than double that observed for lobectomy (Table 4), although this difference did not reach statistical significance (p ). Stratified survival analysis revealed that patients who underwent sublobar resection with lymph nodes sampled had an overall survival function (median not reached; 5-year overall survival, 66%) similar to that of patients treated with lobectomy (p , Fig 2). In contrast, patients who underwent sublobar resection without lymph node sampling experienced shorter survival (median, 6.1 years; 5-year overall survival 56%; p ). Similarly, for patients who underwent sublobar resection with lymph nodes sampling, recurrence-free survival (median not reached; 5-year recurrence-free survival, 81%) was comparable to those treated with lobectomy (p , Fig 3), whereas those without nodal sampling demonstrated shorter recurrence-free survival (median not reached; 5-year recurrence-free survival, 71%; p ). In an effort to explore further the population of patients who underwent wedge resection without nodal staging, these patients were compared to their counterparts who underwent wedge resection with lymph node sampling for the various demographic and clinicopathologic characteristics (Table 5). These data demonstrate that for pulmonary resection of solitary NSCLC tumors of 2 cm or less, recurrence rates were not significantly different for patients who underwent sublobar resection (wedge resection and segmentectomy combined) compared to those who underwent lobectomy, but lobectomy was associated with a trend toward lower rate of local recurrence. Moreover, overall and recurrencefree survival rates were significantly longer for patients treated with lobectomy. When lymph nodes were sampled with sublobar resection, however, the local recurrence rate and overall and recurrence-free survival functions were similar to those seen with lobectomy. It was striking to find that a large number of patients underwent sublobar resection without lymph node sampling. The ability to understand the individual surgeon s decision-making and/or patient-related factors that may have influenced the choice of procedure performed is limited by the retrospective nature of this analysis. Of patients who underwent sublobar resection, 137 (89%) were not staged preoperatively with mediastinoscopy, compared to 40 (48%) lobectomy patients (p ). It is reasonable to hypothesize that many of the wedge resections performed without lymph node sampling (without either preoperative mediastinoscopy or intraoperative lymph node staging) were intended as diagnostic procedures undertaken in older, sicker patients with worse pulmonary function. The ability to test this hypothesis in a retrospective analysis is limited statistically by the small number of events observed in this study. Of particular interest was the observation that sublobar resection with versus without lymph node sampling was associated with significant differences in outcome. In an effort to understand whether a difference between patients who had lymph nodes sampled and those who did not may have confounded the association between extent of resection and recurrence/survival, we compared the characteristics of the patients who were treated with wedge resection without lymph node sampling to those of patients treated with wedge resection alone (Table 5). The 2 groups were found to be similar in all respects except for resection margin, which was statistically smaller in patients who did not have lymph nodes sampled (0.5 cm versus 0.8 cm, p ). While a difference of 3 mm may not be clinically significant when the median margin for both groups was Comment Fig 3. Cumulative survival distribution for recurrence-free survival for lobectomy and sublobar resection, the latter stratified by lymph node sampling.

5 Ann Thorac Surg WOLF ET AL 2011;92: RESECTION FOR SMALL LUNG CANCERS 1823 Table 5. Baseline Demographic and Clinical Characteristics of Patients Undergoing Wedge Resection Characteristic Wedge Resection With Lymph Node Sampling (n 25) Wedge Resection Without Lymph Node Sampling (n 105) p Value Median age, y (range) 71 (51 91) 71 (46 85) Male, n 14 (56%) 39 (37%) Median % predicted FEV 1 (range) 57% (26 123) 67% (23 141) COPD, n 11 (44%) 31 (30%) Central tumor (inner half of hemithorax), n 2 (9%) 20 (23%) Median tumor size, cm (range) 1.5 ( ) 1.4 ( ) Visceral pleural invasion, n 4 (16%) 19 (17%) Squamous (versus nonsquamous), n 4 (16%) 24 (23%) Poorly differentiated, n 9 (36%) 28 (27%) Median distance to nearest stapled margin, cm (range) 0.8 ( ) 0.5 (0 3) COPD chronic obstructive pulmonary disease; less than 1 cm, closer margins may have contributed to the difference seen between the groups. Moreover, the small margins in general support our hypothesis that many of the wedge resections were performed primarily for diagnosis in higher-risk patients. It is also possible that the difference in recurrence and survival seen between patients in the sublobar cohort who had lymph nodes sampled and those who did not was due to understaging of many of the 109 patients who underwent sublobar resection without lymph node sampling. To explore this hypothesis, we must consider the patients who were excluded from analysis due to the various reasons listed earlier (neoadjuvant therapy, BAC, nodal or distant metastases, etc). Of the 344 patients in the original cohort with nodes available for staging, 53 (15%) had nodal metastases, a rate similar to that reported in the literature [9]. It is possible that approximately 15% of the patients who underwent sublobar resection without lymph node sampling would have demonstrated nodal metastases had their nodes been pathologically staged, and thus these patients would have been excluded from the analysis. It is impossible to determine whether this would have altered the results but certainly this study is limited by this bias. In light of the results of this analysis, our current practice emphasizes the importance of adequate lymph node sampling on all NSCLC resections. Moreover, the results of the current study suggest that ablative techniques, such as stereotactic body radiation, which lack a means of pathologic lymph nodal staging, may prove inferior to lobectomy or sublobar resection with lymph node sampling with regard to oncologic outcomes such as local recurrence and survival. While lobectomy continues to represent the gold standard therapy for NSCLC, there is ever increasing interest in parenchyma-sparing resection, including anatomic segmentectomy and nonanatomic wedge resection, particularly as more small peripheral tumors are diagnosed at earlier stages as result of screening efforts and improved imaging technology. As suggested by the updated AJCC seventh edition staging system ([10]), tumors smaller than 2 cm in diameter have more favorable biology. It may be FEV 1 forced expiratory volume in 1 second. reasonable to consider sublobar resection, even in healthy patients, if a tumor is 2cm or smaller. This is particularly relevant as patients live longer and present with metachronous lung tumors (8% in the current study). Thus, there will continue to be 2 types of patients who are considered for sublobar resection: (1) those who are frail, are elderly, and have poor cardiopulmonary function who would not likely tolerate lobectomy and (2) healthy patients with small tumors who are candidates for lobectomy. For the second category of patient, the question of whether sublobar resection with lymph node sampling can achieve outcomes similar to those of lobectomy must be answered with the completion of a prospective randomized trial, such as CALGB References 1. Jemal A, Siegel R, Xu J, et al. Cancer statistics, CA Cancer J Clin 2010;60: Chang MY, Mentzer SJ, Colson YL, et al. Factors predicting poor survival after resection of stage IA non-small cell lung cancer. J Thorac Cardiovac Surg 2007;134: Port JL, Kent MS, Korst RJ, et al. Tumor size predicts survival within stage IA non-small cell lung cancer. Chest 2003;124: Mountain CF. Revisions in the international system for staging lung cancer. Chest 1997;111: Landreneau RJ, Sugarbaker DJ, Mack MJ, et al. Wedge resection versus lobectomy for stage I (T1 N0 M0) non-small cell lung cancer. J Thorac Cardiovasc Surg 1997;113: Ginsberg RJ, Rubnistein LV. Randomized trial of lobectomy versus limited resection for T1 N0 non-small cell lung cancer. Ann Thorac Surg 1995;60: Gajra A, Newman N, Gamble GP, et al. Impact of tumor size on survival in stage IA non-small cell lung cancer: a case for subdividing stage IA disease. Lung Cancer 2003;42: Martini N, Melamed MR. Multiple primary lung cancers. J Thorac Cardiovasc Surg 1975;70: Asamura H, Goya T, Koshiishi Y, et al; Japanese Joint Committee of Lung Cancer Registry. A Japanese Cancer Registry study: prognosis of resected lung cancers. J Thorac Oncol 2008;3: Goldstraw P, executive editor. International Association for the Study of Lung Cancer Staging Manual in Thoracic Oncology. Orange Park, FL: Editorial Rx Press; 2009.

6 1824 WOLF ET AL Ann Thorac Surg RESECTION FOR SMALL LUNG CANCERS 2011;92: DISCUSSION DR M. BLAIR MARSHALL (Washington, DC): For lymph node sampling, did that include the hilar nodes as well? DR WOLF: It did, Dr Marshall. Just to clarify, lymph node sampling included N1 and/or N2 nodes and we did not distinguish between preoperative mediastinoscopy and intraoperative sampling or dissection. Anyone who had interlobar, hilar, or mediastinal nodes sent for pathology counted as having nodes sampled. DR FRANK C. DETTERBECK (New Haven, CT): I suspect that the patients who had lymph nodes sampled, that that is a marker for potentially 2 things. I think, first of all, it s probably a marker for people who had an anatomic segmentectomy as opposed to just a wedge, and I wonder if that doesn t account for some of the overall differences in your results from the last paper we just heard, that segmentectomy seems to have no difference, whereas your collection of segmentectomy and wedge did seem to have a difference, but perhaps not in the patients who had segmentectomy. So that s the first part. The second thing, I wonder if the patients who had nodes sampled were the patients who had a more elective sublobar resection and the ones who had very limited pulmonary function and had a compromise operation were the ones where you just did a wedge and you got out. There are lots of data that those patients don t do as well as the patients who have an elective sublobar resection. So the issue about the lymph nodes, is that just a marker for other issues that may be the real issue? DR WOLF: That s an excellent question. Thank you, Dr Detterbeck. We explored this issue in great detail because we had the same question with regard to the results, whether lymph node sampling was a marker for some other variable that influenced outcome. For example, was it just that patients who had lymph nodes sampled were more likely to have anatomic resection and thus did better? We did only have 24 segmentectomies in this group. The majority of sublobar resections were wedge resections. We did an exploratory analysis of the patients who underwent wedge resection comparing those who had lymph nodes sampled to those who did not. We found that the patients who did not have nodes sampled had a slightly smaller resection margin (.5 versus.8 for patients who had nodes sampled). But otherwise, they were similar for age, gender, lung function, histology, degree of differentiation and the various other parameters. This small difference in resection margin may help explain why patients who had lymph nodes sampled fared better, but we can t be certain that lymph node sampling is not a marker for some other confounder. DR MICHAEL J. LIPTAY (Chicago, IL): Dr Wolf, that was a nice presentation. Thank you very much. I have a quick question about those lymph nodes. In the number of patients who had lymph node sampling, how many were positive, and did you censor those patients? DR WOLF: Thank you for your question, Dr Liptay. The question is whether or not patients with positive lymph nodes were censored, and the answer is yes. The original dataset was 468 patients. Patients were removed for the various reasons listed: induction therapy, multicentric tumors, nodal or distant metastases, BAC, et cetera. There were about 15% of patients who were censored because of N1 nodes or N2 nodal disease. By default, this censorship only applied to patients who had lymph nodes sampled and this bias may account in large measure for the improved survival observed in these patients compared to those who did not have lymph nodes sampled. DR MATTHEW J. SCHUCHERT (Pittsburgh, PA): Did you compare directly your anatomic segmentectomy and wedge experience in terms of recurrence-free survival, as Dr Detterbeck hinted? And also, with regard to overall survival, do you have any information on the cause of death, that is to say death due to cancer or death due to other causes, such as worsening comorbidities or impaired pulmonary function, in your sublobar resection group? DR WOLF: Thank you, Dr Schuchert. Just to make sure I have the 2 questions, one is whether or not we looked at causes.... DR SCHUCHERT: Did you compare segments versus wedge directly? DR WOLF: And whether we looked at the other causes of death. DR SCHUCHERT: Causes of death explaining the difference in overall survival. DR WOLF: For the first question, we did look at the individual procedures, but with only 24 segmentectomies, our survival curves were actually stair-case-looking and did not meet significance, but we were not certain how to interpret those data. For the second question, with regard to other causes of death, with 92% follow-up through physician records and discussion with patients and their family members, we were able to verify whether people died of recurrent disease or not, but we did not document specific other causes of death. For the two sets of survival curves I showed (overall and recurrence-free survival), we saw that the sublobar and lobar resection curves were a lot closer together for recurrence-free survival than for overall survival. Our inference from these data was that the sublobar patients had a higher rate of non-cancer-related deaths. DR DANIEL P. RAYMOND (Rochester, NY): I m curious about your lack of use of a multivariate analysis or at least matched controls. There may be so many confounding variables here that you re simply concluding that older, sicker patients die more frequently. What about the list of coexistent comorbidities and utilizing a multivariate analysis to kind of tease out some of these variables? DR WOLF: That s an excellent question. Most of these results show no differences, and the purpose of using the multivariable analysis would be to tease out whether a difference we saw was due to some confounder, and thus multivariable analysis has limited application here. With the major significant difference being in overall survival, the question is whether the relationship between extent of resection (sublobar vs. lobar) and survival was confounded by a variable, such as age or comorbidities. Unfortunately, with such a small sample size, the number of covariates one could include in a multivariable model is very

7 Ann Thorac Surg WOLF ET AL 2011;92: RESECTION FOR SMALL LUNG CANCERS 1825 limited. With only an 8% to 16% local recurrence rate, we could only use up to one or two variables to have a stable logistic regression model. Similarly, for a Cox regression model, we just didn t have a large enough sample size to conduct that kind of analysis in this study. We tried to account for differences between the two patient samples by performing separate stratified analyses for lymph node sampling, age, and tumor size. I appreciate the point. It s an excellent one. DR RAYMOND: In that circumstance, however, you can use matched controls as a rougher estimation that allows you to overcome some of those obstacles. LATE RESPONSE BY DR WOLF: This is also an excellent suggestion. We did use propensity scores based on type of The Society of Thoracic Surgeons: Forty-Eighth Annual Meeting resection (sublobar vs. lobar) to explore these issues, but the plots had very little overlap. In other words, the samples were too small and too disparate to generate enough matches for use as controls. DR SWANSON: I think one last point about this dataset that s interesting, and I think Frank Detterbeck s point is a good one, there probably is some marker that node sampling identifies that we can t analyze; however, if not, if it s simply an upstaging, then if you look at the 3 groups, the wedge without lymph nodes could be similar to what would be a stereotactic radiosurgery group, and you would then expect those people to do significantly worse than any sublobar resection with nodal sampling. So I think we need to keep that in mind as we move forward with some of these stereotactic protocols. Mark your calendars for the Forty-Eighth Annual Meeting of The Society of Thoracic Surgeons (STS) to be held at the Greater Fort Lauderdale/Broward County Convention Center, Fort Lauderdale, Florida, from January 30 February 1, Come to Fort Lauderdale to learn from the experts, network with colleagues from around the world, and prepare for whatever the future may hold. This pre-eminent educational event in cardiothoracic surgery is open to all physicians, residents, fellows, engineers, perfusionists, physician assistants, nurses, or other interested individuals who work with cardiothoracic surgeons. Meeting attendees will be provided with the latest scientific information for practicing cardiothoracic surgeons. Attendees will benefit from traditional Abstract Presentations, as well as Surgical Forums, Breakfast Sessions, Surgical Motion Pictures, and Procedural Hands-On Courses. Parallel sessions on Monday and Tuesday will focus on specific subspecialty interests. An advance program with a registration form, hotel reservation information, and details regarding spouse/ guest activities will be mailed to STS members this fall. Nonmembers may contact the Society s secretary, David A. Fullerton, MD, to receive a copy of the advanced program; however, detailed meeting information will be available on the STS website at David A. Fullerton, MD Secretary The Society of Thoracic Surgeons 633 N Saint Clair St, Suite 2320 Chicago, IL Telephone: (312) Fax: (312) sts@sts.org website: by The Society of Thoracic Surgeons Ann Thorac Surg 2011;92: /$36.00 Published by Elsevier Inc