Variability in Defining T1N0 Non-Small Cell Lung Cancer Impacts Locoregional Failure and Survival

Transcription

1 Variability in Defining T1N0 Non-Small Cell Lung Cancer Impacts Locoregional Failure and Survival Mert Saynak, MD, Jessica Hubbs, MS, Jiho Nam, MD, Lawrence B. Marks, MD, Richard H. Feins, MD, Benjamin E. Haithcock, MD, and Nirmal K. Veeramachaneni, MD Department of Radiation Oncology and Division of Cardiothoracic Surgery, University of North Carolina at Chapel Hill and UNC Lineberger Comprehensive Cancer Center, Chapel Hill, North Carolina Background. Locoregional recurrence can occur despite complete anatomic resection of T1N0 non-small cell lung cancer. That may be the result of incomplete resection or inaccurate staging. We assessed the impact of extent of nodal staging on the rate of locoregional failure and patient survival. Methods. The records of 742 patients undergoing lobectomy, bilobectomy, or pneumonectomy for non-small cell lung cancer from 1996 to 2006 were reviewed. Operative reports and pathology reports were reviewed for the number of lymph nodes and the anatomic nodal stations examined. The Kaplan-Meier method was applied to analyze recurrence-free survival. Results. A total of 119 patients with pathologically staged Ia lung cancer were identified. Histology type included 61% (n 73) adenocarcinoma, 27% (n 32) squamous cell cancer, and 12% (n 14) other. Median age was 65 years (range, 34 to 88). Mean follow-up duration was 40 months (median 47; range, 1 to 121). Locoregional recurrence occurred in 20% (n 18). The N2 nodal stations were examined in 94% (n 112). At least one defined N1 nodal station was examined in 70% (n 83). Station undefined N1 nodes were examined in 27% (n 32), and no N1 nodes were examined in 3% (n 4). Median number of N1 lymph nodes analyzed was 5 (range, 0 to 18). The locoregional recurrence rate was 14% (12 of 83) for patients with a defined N1 station node versus 31% (11 of 36) for patients in whom there were undefined N1 nodes (p 0.03). Similar differences were seen in disease-free survival, 78.2% versus 62.6%, respectively (p 0.06). Conclusions. Despite anatomic resection of stage Ia lung cancer and uniform analysis of N2 nodal stations, a high rate of locoregional recurrence occurs. Imprecise staging of N1 lymph nodes may contribute to the understaging and undertreatment of patients with early stage lung cancer. (Ann Thorac Surg 2010;90: ) 2010 by The Society of Thoracic Surgeons Surgery serves as the primary modality to treat early stage non-small cell lung cancer. Despite improvements in preoperative clinical staging with imaging modalities such as positron emission tomography (PET) and computed tomography (CT), and apparent R0 resections, several investigators have demonstrated a significant rate of local and distant failure in patients with completely resected early stage lung cancer [1 3]. Failure to properly identify advanced disease may be due to inadequate analysis of the resected specimen or to inadequate resection. Nodal involvement of lung cancer remains the primary predictor of poor long-term outcome. To gain additional insights into why apparent completely resected T1N0 lung cancer has a high rate of treatment failure, we assessed the impact of extent of nodal staging on the rate of locoregional failure, as well as the precision of nodal analysis. We evaluated not only Accepted for publication June 7, Presented at the Fifty-sixth Annual Meeting of the Southern Thoracic Surgical Association, Marco Island, FL, Nov 4 7, Address correspondence to Dr Veeramachaneni, Department of Surgery, 3040 Burnett-Womack Bldg, Campus Box 7065, Chapel Hill, NC ; nirmalv@med.unc.edu. the use of mediastinoscopy but also the definitions used to define a patient as node negative. Material and Methods This study was approved by the University of North Carolina Institutional Review Board. The medical records and pertinent radiographs of 742 patients who underwent surgery for primary lung tumors at the University of North Carolina between January 1996 through December Of these, 119 patients with pathologic stage Ia (T1a/T1bN0M0) were identified who had undergone an anatomic complete resection (lobectomy, sleeve lobectomy, bilobectomy, or pneumonectomy). Only patients with anatomical resections (ie, no less than a lobectomy) were included. Excluded from consideration were patients treated with neoadjuvant therapy (chemotherapy, radiation therapy, or both) or postoperative radiotherapy with or without chemotherapy. All patients were staged according to the American Joint Committee on Cancer seventh edition guidelines adopted in Medical records, operative notes, pathology reports, and pertinent radiologic imaging were reviewed to char by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 1646 SAYNAK ET AL Ann Thorac Surg IMPACT OF T1N0 NSCLC STAGING 2010;90: acterize each patient s demographic information, obtain surgical and pathologic details, and identify patterns of failure after surgery. Patients were reevaluated every 3 to 6 months after resection with either a chest roentgenogram or, if indicated, a CT scan of the chest. The presence of a failure and the site of failure were determined by review of the available medical records. Treatment failure was classified as either local-regional and/or distant. Locoregional sites were defined as ipsilateral lung, stump, chest wall, mediastinum, or supraclavicular lymph nodes. On imaging, a local/regional nodal failure was defined as a new or enlarging lymph node 1 cm or greater on the short axis on CT or hypermetabolic on PET that on the patient s subsequent clinical follow-up was consistent with disease progression. If the pathology of the recurrent lesion was distinctly different from the initial tumor, this was classified as a new tumor rather than as a recurrence. Patients with equivocal evidence of a local-regional failure, whose subsequent course suggested that there was not a localregional failure, were not considered as a local-regional failure. Statistical Analysis Normally distributed continuous data are expressed as mean SD. Categorical data are expressed as counts and proportions. The 2 or Fisher s exact tests were used to analyze categorical data. Comparisons for means of normally distributed continuous variables were done with independent sample t tests. Logistic regression analysis and multivariate analysis were performed using SPSS (SPSS 16.0 for Windows; SPSS, Chicago, IL). For univariate and multivariate analysis, clinically relevant variables of patient sex, patient age, tumor size, tumor histology, and quality of nodal analysis were analyzed. Survival estimates were calculated using the Kaplan- Meier method and were compared using the log rank test. The electronic medical records at University of North Carolina were the primary source of follow-up information. No attempts were made to contact patients by telephone or mail. Results Median follow-up time for the 119 patients was 40 months (range, 1 to 150). In all, 109 (92%) had 6 months or more of follow-up. The patients had a median age of 65 years (range, 34 to 88); 59 (50%) were male and 60 (50%) were female. All patients had preoperative chest CT scan but only 34 of 119 patients (29%) had PET scan for preoperative radiologic evaluation. Mediastinoscopy was performed in 34 of 119 patients (29%). The mediastinal lymph node stations were assessed by either mediastinoscopy or PET scan preoperatively in 59 of 119 patients (50%) before lung resection. The mean pathologic tumor size was 2 cm (range, 0 to 3 cm). Additional patient characteristics and surgery and pathology details are presented in Table 1. The 5-year observed failure rate (local-regional or distant or both) was 21% (25 of 119). The median time to Table 1. Patient, Tumor, and Treatment Characteristics of Patients With Complete Resection of Stage Ia Non-Small Cell Lung Cancer Characteristic Number a Percent Age, years Mean 64 Median 65 Range Sex Male Female Race White African-American Other 2 2 Surgical procedure Lobectomy Bilobectomy 4 3 Pneumonectomy 5 4 Surgical approach Open VATS Mediastinoscopy Yes No Pathologic size of primary lesion, cm Mean 2 Median 2 Range Location Right upper lobe Right middle lobe 4 3 Right lower lobe Left upper lobe Left lower lobe Histology Adenocarcinoma Squamous cell Large cell 5 4 Bronchioloalveolar 3 3 Other 6 5 Histologic differentiation Well Moderate Poor Not stated 2 2 Lymphovascular invasion Yes 11 9 No Not stated 10 8 a Number for all except age and pathologic size row. VATS video-assisted thoracoscopic surgery. failure was 25 months (range, 3 to 87). In patients with cancer recurrence, sites of failure were local only in 7.5% (9 of 119), local and distant in 12% (14 of 119), and distant

3 Ann Thorac Surg SAYNAK ET AL 2010;90: IMPACT OF T1N0 NSCLC STAGING 1647 only 2% (2 of 119). In total, local recurrence was identified in 19% of patients (23 of 119). The median time to local recurrence was 28 months (range, 3 to 87). Of these local recurrences, 78% (n 18) were identified within 3 years of surgery, and 91% (n 21) within 5 years. We investigated the quality of lymph node analysis in this series of pathologically staged Ia patients. Mediastinal lymph nodes (levels 2 to 9) were sampled or dissected in 95% of patients (n 113 of 119). Although only 29% of patients (n 34 of 119) underwent mediastinoscopy, 92% (n 109 of 119) had an N2 lymph node sampled or resected at the time of lung resection. In 3% of patients (n 4), mediastinoscopy was the only means of N2 nodal assessment, and no additional nodes were assessed at the time of thoracotomy. The median number of mediastinal lymph node stations examined was 2 (range, 0 to 6; Fig 1). The N1 lymph nodes (levels 10 to 14) were sampled in 97% of patients (n 115). The median number N1 lymph nodes sampled was 5 (range, 0 to 18). At least one defined N1 nodal station was examined in 70% (n 83 of 119). Station undefined N1 nodes were examined in 27% (n 32), and no N1 nodes were examined in 3% (n 4). These undefined nodes were reported as peribronchial nodes in the pathology report, and had no further identification. Review of the operative notes did not elucidate the location of these peribronchial nodes. Although the local recurrence rate for patients with a defined N1 station node was 14% (12 of 83), local recurrence occurred in 31% (11 of 36) in whom N1 nodes were analyzed without specification of nodal station (p 0.03; Fig 2). In patients who had a named N1 lymph node analyzed, only (36%) had more than a single N1 nodal station analyzed. No N1 nodes were analyzed in 3% of specimens (n 4). There was no difference in the number of nodes analyzed in patients with a defined N1 node and with unspecified N1 nodes. The actuarial 2-, 3-, and 5-year local recurrence-free survival rates were 90.5%, 80.7%, and 75.5%, respectively (Fig 3). In univariate analysis, squamous cell histology and unspecified N1 lymph nodes were risk factors for local recurrence (Table 2). Multivariate analysis revealed Fig 2. Freedom from locoregional failure in patients with defined N1 nodal stations (green line) analyzed versus unspecified nodal stations, namely, peribronchial lymph nodes (blue line). All patients had pathologic stage Ia lung cancer. Data from patients with pathologic T1N1 disease (orange line) are presented for comparison. The data for pt1 N1 patients were obtained through a similar review of the UNC surgical patients contemporaneous with the 119 patients who are the focus of this report. both squamous cell histology (hazard ratio 2.5, 95% confidence interval: 2.12 to 2.93, p 0.029) and unspecified N1 nodal analysis (hazard ratio 2.6, 95% confidence interval: 2.21 to 3.05, p 0.02) to be risk factors for local recurrence. Kaplan-Meier analysis demonstrated an improved lo- Fig 1. Distribution of the number of N2 stations analyzed. Fig 3. Freedom from local failure for patients with completely resected stage Ia lung cancer.

4 1648 SAYNAK ET AL Ann Thorac Surg IMPACT OF T1N0 NSCLC STAGING 2010;90: Table 2. Impact of Clinical and Pathologic Factors on Risk of Local Recurrence Factors Assessed Univariate Analysis Hazard Ratio 95% CI p Value a Sex Male (versus female) Histology Squamous (versus nonsquamous) Surgical approach VATS (versus open) Grade 3 (versus grades 1 and 2) Lymphovascular space invasion Pathologic T stage T1b (versus T1a) Unspecified N1 nodal station a Kaplan-Meier analysis, log rank test. CI confidence interval; surgery. VATS video-assisted thoracoscopic cal recurrence-free survival in patients with named N1 nodal station analyzed (p 0.03; Fig 2). Patients who were deemed N0 without specification of an N1 node had local recurrence-free survival similar to that of patients who were pathologically T1N1. Disease-free survival demonstrated a trend toward improvement among patients with named N1 nodal station analysis. The 5-year disease-free survival was 62.6% for patients with no specification of the N1 nodes analyzed, and 78.2% for patients with defined N1 nodal stations analyzed (p 0.06; Fig 4). Numerous studies have demonstrated the inaccuracy of clinical staging of early stage lung cancer. Even with the application of high resolution CT scans and PET imaging, as much as 15% of clinical T1N0 lung cancer is, in fact, more advanced stage [4, 5]. In this report, we describe variability in how pathologic T1N0 cancer is defined. Although most patients had at least a single N2 station identified and analyzed, we are able to demonstrate inconsistent analysis of N1 lymph node stations. Not all patients had a named N1 station analyzed. From both the pathology reports and the operative records reviewed, it is unclear what stations are defined by the peribronchial descriptor used in 27% of patients. For example, in many operative notes, the surgeon indicated that all lymphatic tissue was dissected onto the bronchus, the bronchus was then transected, and the specimen submitted to the pathologist. In the final analysis, the pathologist use of peribronchial does not provide information as to the nodes being mediastinal, hilar, or lobar (a potential difference of N2 or N1). Strengthening the importance of accurate N1 nodal identification, the authors noted similar local recurrence-free survival when comparing T1N1 patients with patients deemed T1N0 without specified N1 nodes examined. Meticulous attention to nodal dissection improves pathologic staging [6, 7]. The apparent improvement in outcome of early stage lung cancer patients with better lymph node analysis may be simply a function of stage migration. Patients who were truly stage III by N2 nodal involvement may have been inappropriately classified at either stage I or stage II, because the N2 nodes were never assessed. That would lead to an apparent worse long-term outcome in the stage I or II patients. Conversely, accurate pathologic staging would minimize this error in staging, and the cohort of stage I and II patients would have better survival given the precise exclusion of pathologic stage III patients. In this retrospective study, we demonstrate two distinct populations of stage Ia lung cancer patients. Patients who had even a single N1 defined nodal station demonstrated not only diminished local regional failure but also a trend toward improved disease-free survival. Although all of the patients had an apparent R0 resection, the outcomes of the cancer is distinct. That may be because patients who had a defined N1 station node analyzed were truly N0 and therefore were not undertreated by lack of adjuvant therapy. Alternatively, detailed nodal analysis may be a surrogate marker for a larger and more comprehensive resection and lymphadenectomy. There are limited data to support the conclusion that more radical lymphadenectomy improves survival. Although some authors have demonstrated improved survival with mediastinal nodal dissection (N2 nodes) [7 10], and ongoing studies are evaluating the role of mediastinal nodal resection [11], Comment Fig 4. Disease-free survival of patients with defined N1 nodal stations (green line) analyzed versus unspecified nodal stations, namely, peribronchial lymph nodes (blue line). All patients had pathologic stage Ia lung cancer.

5 Ann Thorac Surg SAYNAK ET AL 2010;90: IMPACT OF T1N0 NSCLC STAGING 1649 there are limited data on the extent of N1 dissection necessary for optimal treatment of lung cancer. Unfortunately, this limited retrospective study is not able to answer these important questions. In the United States, there exists considerably variability in the extent and quality of lymph node analysis performed during the surgical treatment of lung cancer [12, 13]. Even in this study, in which most patients had mediastinal nodal assessment, only 42% of patients had at least three mediastinal nodal stations evaluated. Only recently have thoracic surgery societies offered uniform recommendations on what constitutes adequate N1 and N2 station analysis [14]. Among patients who had a named N1 station, only 36% had more than a single N1 station analyzed. Even though patients had an N2 station analyzed routinely, there was little consistency in the extent of N2 analysis, and even less consistency in the extent of N1 analysis. This study supports the need for a defined standard of care for lung cancer resection. We believe that no patient should be labeled as pathologic N0 unless both N1 and N2 nodes have been analyzed at a minimum. The extent of the analysis remains to be defined. Current prospective studies on the role of mediastinal nodal dissection incompletely address this issue, as only analysis of a single level 10 node is required in the American College of Surgeons Oncology Group z0030 study for adequate N1 analysis [11]. At the time of surgical resection, both the surgeon and the pathologist have the opportunity to accurately stage the patient. Accurate staging is both a function of detailed analysis of the resected specimen as well as of adequate resection and sampling of both N1 and N2 lymph node stations. By lack of attention to this important aspect of lung cancer treatment, the role and value of surgical resection in the treatment of early stage lung cancer is being diminished. In this limited study, we have demonstrated that we are better able to predict patient outcome by utilizing basic standards for pathologic staging. Prospective studies of this basic aspect of surgical treatment of lung cancer are necessary to provide guidelines for the surgeon and the pathologist. In conclusion, despite anatomic resection of T1N0 lung cancer and uniform analysis of N2 nodal stations, a high rate of locoregional recurrence occurs. Imprecise staging of N1 lymph nodes may contribute to the understaging and undertreatment of patients with early stage lung cancer. Mert Saynak was supported by a grant from the Turkish Association of Radiation Oncology. References 1. Rena O, Oliaro A, Cavallo A, et al. Stage I non-small cell lung carcinoma: really an early stage? Eur J Cardiothorac Surg 2002;21: Kelsey CR, Marks LB, Hollis D, et al. Local recurrence after surgery for early stage lung cancer: an 11-year experience with 975 patients. Cancer 2009;115: Fujimoto T, Cassivi SD, Yang P, et al. Completely resected N1 non-small cell lung cancer: factors affecting recurrence and long-term survival. J Thorac Cardiovasc Surg 2006;132: Stiles BM, Servais EL, Lee PC, et al. Point: clinical stage IA non-small cell lung cancer determined by computed tomography and positron emission tomography is frequently not pathologic IA non-small cell lung cancer: the problem of understaging. J Thorac Cardiovasc Surg 2009;137: Veeramachaneni NK, Battafarano RJ, Meyers BF, et al. Risk factors for occult nodal metastasis in clinical T1N0 lung cancer: a negative impact on survival. Eur J Cardiothorac Surg 2008;33: Massard G, Ducrocq X, Kochetkova EA, et al. Sampling or node dissection for intraoperative staging of lung cancer: a multicentric cross-sectional study. Eur J Cardiothorac Surg 2006;30: Izbicki JR, Passlick B, Pantel K, et al. Effectiveness of radical systematic mediastinal lymphadenectomy in patients with resectable non-small cell lung cancer: results of a prospective randomized trial. Ann Surg 1998;227: Wu Y, Huang ZF, Wang SY, et al. A randomized trial of systematic nodal dissection in resectable non-small cell lung cancer. Lung Cancer 2002;36: Lardinois D, Suter H, Hakki H, et al. Morbidity, survival, and site of recurrence after mediastinal lymph-node dissection versus systematic sampling after complete resection for nonsmall-cell lung cancer. Ann Thorac Surg 2005;80: Keller SM, Adak S, Wagner H, et al. Mediastinal lymph node dissection improves survival in patients with stages II and IIIa non-small cell lung cancer. Eastern Cooperative Oncology Group. Ann Thorac Surg 2000;70: Allen MS, Darling GE, Pechet TT, et al. Morbidity and mortality of major pulmonary resections in patients with early-stage lung cancer: initial results of the randomized, prospective ACOSOG Z0030 trial. Ann Thorac Surg 2006;81: Farjah F, Flum DR, Varghese TK, et al. Surgeon specialty and long-term survival after pulmonary resection for lung cancer. Ann Thorac Surg 2009;87: Little AG, Rusch VW, Bonner JA, et al. Patterns of surgical care of lung cancer patients. Ann Thorac Surg 2005;80: Lardinois D, De Leyn P, Van Schil P, et al. ESTS guidelines for intraoperative lymph node staging in non-small cell lung Cancer. Eur J Cardiothorac Surg 2006;30: DISCUSSION DR TODD L. DEMMY (Buffalo, NY): The authors propose that the attention to detail regarding N1 node identification in early stage lung cancer may translate into improved survival. If replicated by others, the mechanism behind this phenomenon will need study. Many could surmise that the benefit is not from the labeling per se but rather reflects a disciplined, programmatic, multifaceted, and probably surgeon-led commitment to oncologic principles. Possibly amplifying the usual concerns of selection bias in a retrospective trial, it is important to note that in this particular 10-year time interval there were considerable national practice changes occurring that might have influenced the management of N1 nodes or the population referred for surgery. This includes a time where plain PET was hard to get up to a time where PET-CT imaging was dominant, the emergence of VATS lobectomy, and the introduction of an updated staging system. The latter is important because many lobes were fine-

6 1650 SAYNAK ET AL Ann Thorac Surg IMPACT OF T1N0 NSCLC STAGING 2010;90: sectioned (to locate reclassified T4 satellite nodules), thus making it harder to keep the nodal anatomy distinct. Besides these national trends, it is reasonable to expect that there are even more important institutional changes over this time such as surgeon or pathologist turnover with different staging philosophies. This leads to question one. Do you know whether the adverse finding of nonspecified was associated with surgeon or pathologist, type of approach such as VATS, other staging modality, or a specific time interval that might suggest a different mechanism, and did you consider using a multivariate analysis to determine whether or not there was an interaction among some of these factors that you studied? Question two, one interaction I would like you to comment upon was the other adverse effect of squamous histology on recurrence. Accepting your hypothesis, I wonder whether plucking an interesting N1 node in an adenocarcinoma patient, labeling it, and finding out that it is negative may be more prognostic than a squamous patient where there is generalized nodal enlargement from necrosis clouding the picture. Do you think squamous patients need different nodal processing and do you feel that there should be standards on nodal assessment like the number of nodes processed like exists for colon cancer? Was the median of five nodes in your study adequate? I would like to thank the Association for the privilege of discussing this paper and the authors for providing the manuscript. Thank you. DR VEERAMACHANENI: Thank you, Dr Demmy. In answer to your question regarding squamous cell carcinoma, I did not present that data but just mentioned it. In multivariant analysis, both the lack of specificity of N1 nodal analysis as well as squamous cell histology both predicted increased failure. After analyzing these data, I have changed how I analyze lymph nodes intraoperatively. The use of taking frozen sections has certainly increased, as has been how I present the lymph nodes to our pathologists. I am far more conscientious in patients with squamous cell cancer given the increased local recurrence observed in these patients. I have resorted to dissecting out the specimens a lot more thoroughly, and instead of giving the whole specimen to the pathologist, I provide individual nodal packets in individual labeled containers. As far as the number of nodes, I don t think we have a clear idea of what a true number should be. Especially looking at the N2 data, it is impossible to tell how many nodes were actually taken because they are often represented as fragments. So to be fair to the pathologists, I limited that information to the number of stations analyzed rather than the number of nodes. There are certainly standards with other malignancies, such as colon cancer, regarding number of nodes to be analyzed, and, unfortunately, we seem to lack one for lung cancer. DR BRYAN FITCH MEYERS (St. Louis, MO): The techniques have changed with more and more VATS lobectomy operations being done. The way we get lymph nodes out with a VATS lobectomy is often different than what you do with an open lobectomy. How do you think the numbers of lymph nodes that are being removed are going to change with increased use of VATS lobectomy, and do you see any challenges with analyzing them? DR VEERAMACHANENI: For an operation to be sound even with different techniques, you should pretty much be doing the same oncologically sound operation regardless of VATS versus open. I do the operations the same way you taught me how to do them, which is careful attention to detail, finding the interlobar nodes, and classifying level 10, level 11 for the pathologist. I have taken one step further, and, once the specimen is out of the patient, actually dissecting it a little further, because there is such considerable variability in how the pathologists are handling the specimens. DR ROBERT CERFOLIO (Birmingham, AL): A recent report from the multi-institutional study from ACOSOG suggests that removing all of the N2 nodes offers no real advantage over just sampling the N2 nodes. If true, then how do you see this study affecting the incidence of doing VATS or robotic lobes over open minimally invasive, rib and nerve-sparing lobes since I find that the biggest part or challenge of the VATS lobectomy is removing the nodes. Do you think there is going to be an upswing with VATS lobectomy based on that, or do you think that, in your hands, sampling is as good as a lymph node resection with a scope? Do you think it makes a difference? DR VEERAMACHANENI: I have been trying to do the same operation by either technique. DR CERFOLIO: But do you remove or sample? DR VEERAMACHANENI: My practice is to remove.