Neoadjuvant Chemoradiation for Clinically Advanced Non-Small Cell Lung Cancer: An Analysis of 233 Patients

Transcription

1 Neoadjuvant Chemoradiation for Clinically Advanced Non-Small Cell Lung Cancer: An Analysis of 233 Patients Anthony W. Kim, MD, Michael J. Liptay, MD, Philip Bonomi, MD, William H. Warren, MD, Sanjib Basu, PhD, Erin C. Farlow, MD, and L. Penfield Faber, MD Section of Thoracic Surgery, Yale University School of Medicine, New Haven, Connecticut; and Division of Thoracic Surgery and Section of Medical Oncology, Rush University Medical Center, Chicago, Illinois Background. Surgical intervention after chemoradiation for locoregionally advanced non-small cell lung cancer (NSCLC) is controversial. This study evaluated patient survival after neoadjuvant chemoradiation and anatomic pulmonary resections for locoregionally advanced NSCLC. Methods. Clinicopathologic data were retrospectively collected for 233 patients (110 women, 123 men) with NSCLC who underwent chemoradiation therapy, followed by pneumonectomy, sleeve lobectomy, bilobectomy, and standard lobectomy, from 1989 to Univariate log-rank analysis of Kaplan-Meier survival curves and multivariate Cox regression analysis was performed. Results. Final pathologic stages were complete responders, 52 (22%); I, 56 (24%); II, 39 (17%); and III, 86 (37%). Final pathologic lymph node status was N0, 130 (56%); N1, 28 (12%); and N2, 75 (32%). Overall 5-year survival for the cohort was 43%. The 90-day mortality was 8% (18 of 233). The 5-year survival was 33% for pneumectomy vs 51% for lobectomy (p 0.002). Survival rates at 5 years by stage were complete responders, 58%; I, 50%; II, 41%; and III, 32%; by primary tumor status, T0, 50%; T2, 38%; T3, 29%; and T4, 28%; and by final pathologic nodal status, N0, 51%; N1, 40%; N2, 32% (N0 vs N1, p 0.236; N1 vs N2, p 0.704; N0 vs N2, p 0.019; N0 vs N1 N2, p 0.020). Multivariate analysis demonstrated pneumonectomy was associated with decreased 5-year survival (hazard risk, ; 95% confidence interval, to 2.189, p ). Conclusions. Respectable survival can be achieved after neoadjuvant chemoradiation, followed by anatomic resection, in selected patients with clinically advanced NSCLC. A T0 primary tumor or N0 lymph node status individually, or together as a complete response (T0 N0) status, is associated with the best long-term survival. Survival is most favorable for lobectomies vs pneumonectomies after neoadjuvant chemoradiation therapy. (Ann Thorac Surg 2011;92:233 43) 2011 by The Society of Thoracic Surgeons Anatomic pulmonary resection of non-small cell lung cancer (NSCLC) after neoadjuvant chemoradiation therapy, although still controversial, is a widely accepted practice. The use of trimodality therapy has grown since early reports espoused its benefits [1 3]. Recently, we [4] and others [5] have demonstrated that pneumonectomy after neoadjuvant chemoradiation therapy or chemotherapy alone, respectively, could provide satisfactory 5-year survivals. However, we recommended caution when considering a right-sided pneumonectomy after chemoradiation therapy due to increased perioperative death [4]. This study evaluated the effect of anatomic pulmonary resections on survival after neoadjuvant chemoradiation in selected patients. Accepted for publication March 7, Presented at Fifty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Orlando, FL, Nov 3 6, Address correspondence to Dr Faber, University Thoracic Surgeons, 1725 W. Harrison, Ste 774, Chicago, IL 60612; pfaber@rush.edu. Material and Methods The medical records of all patients undergoing anatomic pulmonary resection for NSCLC after neoadjuvant chemoradiation therapy at a Rush University Medical Center from January 1989 through December 2008 were recorded. Institutional Review Board approval was obtained to perform this retrospective study, and individual patient consent was waived. Patients, generally, with clinically advanced NSCLC or histologically positive mediastinal lymph nodes were carefully selected for therapy by a multidisciplinary group. More specifically, the patients included in this study were those who (1) were diagnosed with locoregionally advanced NSCLC (T1 to T4 N2) or central disease (T3 N0 to N2), (2) underwent completion of chemotherapy and radiotherapy in concurrent splitcourse fashion, and (3) underwent a anatomic resection at the completion of their full course of neoadjuvant therapy. Exclusion criteria were (1) patients who underwent single-modality neoadjuvant therapy (chemotherapy or 2011 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 234 KIM ET AL Ann Thorac Surg NEOADJUVANT CHEMORADIATION FOR NSCLC 2011;92: radiotherapy alone), (2) patients who underwent chemoradiation therapy for another malignancy (head and neck cancer, small cell cancer, etc), (3) patients with incomplete (R1) resections, and (4) patients thought to be unresectable despite neoadjuvant therapy or physiologically unfit to tolerate the regimen. Clinical variables abstracted included age, sex, era of treatment, chemoradiation therapy regimen, type of resection, pretreatment clinical and final pathologic T N M stage, and survival. Patients undergoing resection after neoadjuvant chemoradiation before and including 1994 were considered to be from an early era. Those patients treated after 1994 were considered to be from a later era. This cutoff was based on the end of 1994 being the mean value for time in this study. Long-term survival was established through a multifaceted approach. If survival or death was not confirmed from a review of the medical records available electronically or within the immediately available office records during the 6 months before the initiation of the data analysis, then the Social Security Death Index, referring or consulting physician records, and patient or family contacts were the other means through which survival was determined. Mediastinal Staging Before patients received neoadjuvant therapy, mediastinal lymphadenopathy was confirmed by cervical mediastinoscopy or radiographic evidence of disease as (1) the presence of lymphadenopathy in the mediastinum greater than 2.0 cm on long axis view or 1.0 cm on short axis view, or (2) the presence of lymphadenopathy with positive positron-emission tomography (PET) findings (maximum standard uptake value 2.5). Use of invasive mediastinal staging depended on clinical judgment and the decision of a multidisciplinary team. Radiographic staging alone was relied on more frequently in the earlier period, whereas PET scan and mediastinoscopy were used more systematically later in the study period. Patients who underwent mediastinoscopy to determine eligibility for neoadjuvant chemoradiation did not undergo repeat invasive mediastinal staging and were restaged by computed tomography (CT) or PET scan. Data were not recorded for patients who were initially given neoadjuvant chemoradiation therapy, but in whom progressive disease developed that precluded surgical resection, and they were excluded from this analysis. All patients were evaluated and managed by a multidisciplinary team that included thoracic surgeons, medical oncologists, and radiation oncologists. Treatment Chemotherapy was administered concurrently with split-course radiotherapy in all of the regimens used. All patients received platinum-based chemotherapy. Before 1994, the regimen consisted of cisplatin (16 g/m 2 /day on day 1) and fluorouracil (800 mg/m 2 /day as a 24-hour infusion from days 1 through 5). Etoposide (80 mg/m 2 on days 1 to 3) was added to the regimen later, and with this, the dosage of fluorouracil was decreased (800 mg/m 2 /day as a 24-hour infusion from days 1 through 4). After 1994, the regimen changed to carboplatin (area under the curve [AUC] dose of 4 on day 1), paclitaxel (100 mg/m 2 on days 1 and 8), and etoposide (80 mg/m 2 on days 1 through 3). Etoposide was eliminated from the regimen by The standard course of radiotherapy consisted of three treatment cycles, with each cycle lasting 21 days. Radiotherapy was delivered once daily, 180 to 200 cgy per fraction, on days 1 to 5 and 8 to 9 (or 10), to a goal dose of 4500 cgy. The interval between cycles was 7 to 8 days, with 1-day variability in the break depending on whether the patient received the 10th session. A radiation dose of 4000 was still accepted as a full course of therapy in a few patients. The median dose of radiation was 4300 cgy. Anatomic pulmonary resections consisted of pneumonectomy, sleeve lobectomy, bilobectomy, and standard lobectomy. Sleeve lobectomy, bilobectomy, and standard lobectomy were grouped as lobectomies because they were considered to be variations of this resection for the purposes of analysis. Lymphadenectomy, as distinguished from lymph node sampling, was performed for all of the resections. The right-sided resections underwent dissection of the level 2, 4, 7, 8, 9, 10, and 11 lymph node stations. The left-sided resections underwent dissection of the levels 5, 6, 7, 8, 9, 10, and 11 lymph node stations. Routine buttressing of the bronchial stumps or sleeve anastomosis with muscle flaps was not performed, although pleural or pericardial fat flaps were commonly used. Statistical Analysis For the univariate analysis, the 5-year survivals were calculated and compared using Kaplan-Meier survival curves and log-rank comparisons, respectively. All of the Kaplan-Meier 5-year survival curves were generated based on the time from resection and included all patients, including those who died perioperatively. We also used 2 to compare proportions. Subset analyses were performed in the patients with N2 lymph node disease. Each analysis for this subset compared survival stratified according to type of resection, single station, or multiple station final pathologic lymph node involvement, as well as the change in nodal status from the clinical N2 status before treatment. An additional subset analysis also compared survival of patients with clinical N2 or N3 disease with the survival of patients with clinical N0 or N1 disease. A multivariate analysis using a Cox proportional hazard regression model was performed to evaluate the effect of several covariates on 5-year survival. Variables included in the multivariate analysis were selected from the clinical variables compared on univariate analysis with a p-value of 0.1. Statistical analysis was performed using SAS 9.1 software (SAS Institute Inc, Cary, NC). Statistical significance was defined as p 0.05.

3 Ann Thorac Surg KIM ET AL 2011;92: NEOADJUVANT CHEMORADIATION FOR NSCLC 235 Results We identified 233 patients who had undergone neoadjuvant chemoradiation therapy, followed by anatomic pulmonary resection (Table 1). These 233 patients represented those patients that who were successfully treated in the intent-to-treat trimodality paradigm and were 86% of the 272 patients in the total intent-to-treat population. Ultimately, 34 patients (14%) of the intent-to-treat population were excluded: 16 refused the surgical intervention, the disease progressed in 10, and 8 patients died. All of the 16 patients who refused the operation finished a definitive course of chemoradiation therapy, and there were 2 long-term survivors. Another 5 patients (2%) proceeded to thoracotomy but were excluded at the time of operation due to unresectability. The mean age of the total population was years (range, 30 to 78 years). There were 110 women and 123 men. Resections consisted of 96 pneumonectomies (41%), 114 standard lobectomies (49%), 13 sleeve lobectomies (5%), and 10 bilobectomies (4%). Among the 233 patients, the preresection clinical stages were I, 4 (2%); II, 39 (17%); and III, 190 (81%; Table 2). Cervical mediastinoscopy was used in 33% of the patients. Final pathologic stages after resection were complete responders, 52 (22%); I, 56 (24%); II, 39 (17%); and III, 86 (37%; Table 3). The overall change in the pretreatment clinical stage to the final pathologic stage for the cohort is listed in Table 4. The final pathologic status of the surgically resected primary tumors was T0, 54 (23%); T1, 66 (28%); T2, 64 (28%); T3, 31 (13%); and T4, 18 (8%). The final pathologic status of the surgically resected lymph nodes was N0, 130 (56%); N1, 28 (12%); and N2, 75 (32%; Table 3). The overall 5-year survival for all patients was 43%, with a median survival of 40 months (Fig 1). Mortality rates were 3% (8 of 233) at 30 days and 8% (18 of 233) at 90 days. When classified by resection type, 30-day and 90-day mortality rates were 6% (6 of 96) and 13% (13 of 96) for pneumonectomy, and 1% (2 of 137) and 4% (5 of 137) for lobectomy, respectively. Among patients undergoing pneumonectomy, pulmonary complications, including Table 1. Perioperative Patient Characteristics and Operative Information Variable No. (%) or Mean (Range) Total patients 223 (100) Mean age, years (30 78) Sex Female 110 (47) Male 123 (53) Type of operation Pneumonectomy 96 (41) Lobectomy 137 (59) Standard 114 (49) Sleeve 13 (5) Bilobectomy 10 (4) Table 2. Clinical Stage Stage Total No. Lobectomy No. Pneumonectomy No. IB 4 ct2 N IIB 39 ct2 N ct3 N III 190 IIIA 163 ct3 N ct1 N ct2 N ct3 N IIIB 27 ct4 N ct4 N ct4 N ct2 N acute respiratory distress syndrome, were the most frequent cause of death at 90 days (Table 5). Patient survival included the patients who had died within 90 days of their operation. Univariate analysis demonstrated that there were no differences in overall survival according to sex or laterality of resection. No significant differences were observed when survival was stratified according to the individual pretreatment clinical T status, clinical N status, and clinical stage. According to the type of resection, 5-year survival rates were 33% (median, 31 months) for pneumonectomy vs 51% (median, not reached; probability of survival 0.5) for lobectomy (p 0.002; Fig 2). According to the stage after treatment, 5-year survival rates were complete responders, 58% (median, not reached; proba- Table 3. Final Pathologic Features Final Pathologic Variable Total No. (%) Lobectomy No. Pneumonectomy No. Stage ypcr 52 (22) ypi 56 (24) ypii 39 (17) ypiii 86 (37) Primary tumor status ypt0 54 (23) ypt1 66 (28) ypt2 64 (28%) ypt3 31 (13%) ypt4 18 (8%) 5 13 Lymph node status ypn0 130 (56%) ypn1 28 (12%) ypn2 75 (32%) CR Complete responders (T0 N0).

4 236 KIM ET AL Ann Thorac Surg NEOADJUVANT CHEMORADIATION FOR NSCLC 2011;92: Table 4. Pretreatment Clinical to Final Pathologic Stage Change Aggregated by Stage Stage Pretreatment Clinical (%) Final Pathologic (%) % 0 (CR) I II a III a This value does not reflect the absence of change for this specific group, but rather, represents the equivalent number of patients who were largely downstaged out of stage II with the number of patients who were downstaged out of stage III and into stage II. CR complete responder. bility of survival 0.5); stage I, 50% (median, not reached; probability of survival 0.5); stage II, 41% (median, 41 months), and stage III, 32% (median, 22 months). The survival among the complete responders was 58% (median, not reached; probability of survival 0.5), which was significantly greater than the 40% for all the other patients with residual disease (median, 31 months), either in the primary tumor or in the lymph node (p 0.012; Figure 3). The effect of the stage after treatment was further examined by reviewing the individual T and N status. When stratified according to the primary tumor status, the 5-year survival rates were T0, 58% (median survival, not reached; probability of survival 0.5); T1, 50% (median, 40 months); T2, 38% (median, 24 months); T3, 29% (median, 25 months); and T4, 28% (median, 31 months). A significant difference was found in survival between T0 and all other T categories combined (T1 T2 T3 T4; p 0.02). When stratified according to final pathologic nodal status, 5-year survival rates were N0, 51% (median, not reached; probability of survival 0.5); N1, 40% (median, 32 months); and N2, 32% (median, 24 months; Fig 4). The difference in survival was not significant between N0 and N1 disease (p 0.236) and between N1 and N2 disease (p 0.704). In contrast, the difference in survival between N0 and N2 disease was significant (p 0.019). A significant difference was found Table 5. Cause of 90-Day Mortality Cause of Death No. Pneumonectomy 13 ARDS/pulmonary 4 Arrhythmia 2 Bronchopleural fistula 2 Aspiration 1 Esophageal Perforation 1 Intraoperative hemorrhage 1 Thromboembolic disease 1 Unknown 1 Lobectomy 5 ARDS 1 Aspiration 1 Massive hemoptysis 1 Respiratory failure 1 Other operation 1 Total 18 ARDS acute respiratory distress syndrome. in survival between N0 and any node-positive (N1 N2) disease (p 0.020; Fig 5). The era of treatment was associated with a significant difference in survival. The survival of patients treated before or during 1994 was 37% (median, 28 months), which was worse than the 52% after 1994 (median, not reached; probability of survival 0.5). A subgroup analysis of the 75 patients with final pathologic N2 disease found no significant differences in survival according to the age, sex, and laterality of resection. This was also the case when survival was compared by the type of resection. In patients with final pathologic N2 disease, the 5-year survival for the lobectomies was 34% and 30% for the pneumonectomies. Despite the absence of a significant difference, the survival curves demonstrated a consistent pattern of patients with lobectomies faring better than pneumonectomies, irrespective of survival time (Fig 6). When patients with pneumonectomies were further divided into right- and left-sided Fig 1. Overall 5-year survival for the entire cohort of 233 patients was 43%, with a median survival of 40 months; (n) represents the at-risk population.

5 Ann Thorac Surg KIM ET AL 2011;92: NEOADJUVANT CHEMORADIATION FOR NSCLC 237 Fig 2. The 5-year survival is shown stratified by operative resection type. The 96 patients in the pneumonectomy group (solid line) were compared with the 137 patients in the combined lobectomy group (standard lobectomy bilobectomy sleeve lobectomy, dashed line); (n) represents the at-risk population. subgroups and compared with lobectomies, again, no significant survival differences were observed. However, irrespective of the survival time, the survival curves for right-sided pneumonectomies, in general, were consistently worse than for lobectomies or left-sided pneumonectomies, which were similar to each other (Fig 7). The 5-year survival associated with N2 lymph node involvement in a single station was 38% (median survival, 26 months) vs 23% in multiple stations (median survival, 24 months), which was not significantly different (p 0.42). Further analysis by resection type showed that among pneumonectomy patients, the presence of singlestation or multiple-station N2 lymph node involvement did not affect 5-year survival, 39% (median, 20 months) vs 18% (median, 16 months; p 0.24), respectively. The same was true for lobectomy, 36% (median, 26 months) vs 30% (median, 31 months; p 0.098), respectively. We do not know if these patients had clinical single-station or multiple-station disease before treatment because these data were not recorded. True mediastinal clearance in this study cohort was difficult to assess because imaging was used to diagnose most of the patients with stage IIIA disease and they did not undergo mediastinoscopy. Nevertheless, 74 of the 163 patients with clinical N2 or N3 disease before treatment were believed to be downstaged to a final pathologic N0 status, whereas 89 patients were believed to be downstaged to a final pathologic N0 or N1 status. The survival of the patients downstaged to a final pathologic N0 status was 51% (median, not reached; probability of survival 0.5), whereas the survival of patients downstaged to a final pathologic N1 status or not downstaged at all was 34% (median, 25 months). This difference approached significance (p ). The survival of patients not downstaged at all, exclusively, was 32% (median, 24 months). The difference between those downstaged to an N0 status vs those not downstaged was slightly significant (p ). On the basis of the univariate analysis, type of resection, postoperative T status, postoperative N status, postoperative stage, and era of treatment were included in the multivariate analysis. This multivariate analysis demonstrated that pneumonectomy was prognostic for a worse 5-year survival (hazard ratio, , 95% confidence interval, to 2.189; p ; Table 6) A later era of treatment approached significance as being prognostic for an improved survival among patients treated after 1994 (hazard ratio, ; 95% confidence interval, to ; p ). Comment The use of chemotherapy and radiotherapy in a neoadjuvant paradigm before anatomic pulmonary resection at Fig 3. The 5-year survival is shown stratified by response to chemoradiation therapy. Complete (solid line) denotes the 52 complete responders with final pathologic T0 N0 status, whereas Residual (dashed line) denotes the 181 patients with any residual disease (ie, anything other than a complete response); (n) represents the at-risk population.

6 238 KIM ET AL Ann Thorac Surg NEOADJUVANT CHEMORADIATION FOR NSCLC 2011;92: Fig 4. The 5-year survival is shown stratified by the 130 patients with N0 (solid line), the 28 with N1 (dashed line), and the 75 with N2 (dotted line) nodal involvement; (n) represents the at-risk population. Rush University Medical Center was largely based on the initially successful experience [6] and the subsequent respectable outcomes associated with its use [1, 4, 7 9] by the multidisciplinary lung cancer team. The successful experiences of others [10 12] have served to further support its use. Despite these favorable outcomes, the challenge remains in identifying those patients who benefit the most from surgical resection after neoadjuvant therapy. Some have suggested that the simple presence of a clinical response to neoadjuvant therapy and the ability to operatively eradicate disease is sufficient justification for surgical intervention [13, 14] The improved long-term survival associated with an N0 lymph node status is thought to reflect the elimination of undetected, but suspected systemic disease [13, 15]. The reduction of the tumor burden in the lymph nodes by PET scan has been associated with an improved 5-year survival [16]. Although performed in a subset analysis and found to be significant when complete nodal sterilization was achieved, evaluating lymph node downstaging was not the primary focus of this study. Because mediastinoscopy was selectively used, we do not know which patients were truly downstaged with chemoradiation. Despite this, the results of this study suggest that nodal downstaging may be associated with a more favorable survival, a finding that has been observed by others. We and others believe that resection of residual positive N2 disease after induction chemotherapy can be associated with long-term survival that is improved over medical therapy alone [14, 17]. Cerfolio and colleagues [17, 18] demonstrated that the resection of unsuspected N2 disease after neoadjuvant chemoradiation therapy has been associated with a 5-year survival of 42%. This is not a universally shared opinion, however. Deslauriers and colleagues [19] reported unfavorable outcomes with sleeve lobectomy or pneumonectomy in patients with positive N2 disease after chemoradiation therapy. The addition of preoperative radiation therapy has been thought to affect more of a complete response than induction chemotherapy alone, thereby resulting in aggressive locoregional control [13]. Further supporting the added beneficial effect of preoperative radiotherapy is that higher doses of preoperative radiation have resulted in an increased incidence of complete response (T0 N0) [11, 18]. Also, favorable outcomes have been reported in patients originally deemed unresectable undergoing salvage lung resections after definitive chemoradiation therapy, with radiation doses greater than 59 cgy [20]. In that series of 24 patients, 22 had stage III or higher disease: 14 patients had N2 disease, and 6 had N3 disease. Univariate analysis showed no difference in survival between patients with N0 and N1 or lymph node Fig 5. Comparison of the 5-year survival is shown stratified by the 130 patients with no nodal involvement (N0; solid line) vs the 103 with N1 and N2 (dashed line) nodal involvement; (n) represents the at-risk population.

7 Ann Thorac Surg KIM ET AL 2011;92: NEOADJUVANT CHEMORADIATION FOR NSCLC 239 Fig 6. Comparison of the 5-year survival is shown for patients with N2 involvement stratified by type of resection. Although the 35 patients with lobectomies (dashed line) did not differ significantly from the 40 with pneumonectomies (solid line), there was a consistent trend of patients with lobectomies faring better than those with pneumonectomies, irrespective of the survival period; (n) represents the at-risk population. status or between those with the N1 and N2 lymph node status. However, significant differences in survival were observed when patients with N0 vs N2 lymph node status or N0 vs N1 N2 lymph node status were compared. Other investigators [21, 22] have found significant differences in survival when patients with positive N2 lymph nodes were compared with those patients with N0 and positive N1 lymph nodes after neoadjuvant therapy were grouped. The 5-year survival for patients with an N0 or N1 lymph node status has been approximately 53% in these studies, whereas the survival for patients with N2 involvement has been 10% to 16% for those with N2 lymph node status [21, 22]. The investigators in this study focused on the comparison in survival between patients with N0 vs N1 N2 lymph node involvement due to the general principle that N1 involvement represents the leading edge of the continuum of lymph node involvement, and similar to N2 disease, would merit postoperative adjuvant chemotherapy. The survival reported in this study is higher than the survival reported in other studies. This finding is most likely because the survival reported here represents overall survival, without a distinction between diseasefree or progression-free survival. We did not specifically examine the incidence or pattern of recurrence. Patients with recurrence may have possibly accounted for a substantial number of patients who were alive among the entire cohort of patients, as has others have suggested [23]. Even if the subset of patients with recurrence comprised a fair number of patients, the use of neoadjuvant therapy still speaks to a substantial number of patients who can still have long-term survival despite not experiencing a true disease-free state. Another factor contributing to the favorable results in this study was that the 233 patients undergoing surgical resection underwent a full course of neoadjuvant chemoradiation therapy and a complete surgical resection. Jacklitsch and colleagues [23] showed that long-term overall survival and failure-free survival were very similar when complete surgical resection was achieved after induction chemotherapy. Lastly, a later era of treatment approached significance for being prognostic for an improved 5-year survival. This observation is not very surprising given that with time, surgeon comfort with operating on patients having undergone neoadjuvant chemoradiation therapy increases, refinements in patient care strategies with increasing institutional experience are made, and technologic advances that facilitate patient care are incorporated. The results of this study, therefore, may have been weighted favorably by this added factor. Limitations of this retrospective study include the Fig 7. Comparison of the 5-year survival is shown for patients with N2 involvement stratified by the type of resection. The survival curves demonstrate that the 23 patients with right-sided pneumonectomies (dotted line) experienced a consistently worse survival than the 35 with lobectomies (solid line) or the 14 with left-sided pneumonectomies (dashed line), which had more similar curves; (n) represent the at-risk population.

8 240 KIM ET AL Ann Thorac Surg NEOADJUVANT CHEMORADIATION FOR NSCLC 2011;92: Table 6. Multivariate Analysis Variable HR (95% CI) p Value Type of resection ( ) Pneumonectomies vs lobectomies ypt status ( ) ypt0 vs yp(t1 T2 T3 T4) ypn status ( ) ypn0 vs yp(n1 N2) Postoperative stage ( ) ypcr vs yp(i II III) Era of treatment ( ) vs 1994 CI confidence interval; CR complete responders (T0 N0); HR hazard ratio. inherent selection bias of the study cohort. The scope of this study primarily was to determine final pathologic features that were associated with long-term survival after neoadjuvant chemoradiation therapy and anatomic pulmonary resections. Therefore, identifying factors that could be used to better select patients for treatment with neoadjuvant chemoradiation therapy and anatomic resection, or resection after neoadjuvant chemotherapy, was not possible from the data collected or analyzed. Another major limitation is that cervical mediastinoscopy was not performed in most patients. This limitation has several ramifications, including a reliance on noninvasive clinical staging to determine which patients would receive neoadjuvant therapy. It certainly could be possible that invasive staging would have shown that those patients thought to have N2 disease by imaging criteria might not have truly had positive ipsilateral mediastinal lymph node involvement. We acknowledge the ample evidence demonstrating the relatively limited value of CT alone, PET scan alone, or even CT/PET together for staging the mediastinum [24]. We also recognize that the use of different clinical criteria for declaring lymph nodes that were pathologic might also have affected the outcomes. Clinical staging was accomplished by CT scanning alone early in the series and with PET/CT scanning and selective mediastinoscopy later in the series. The absence of significant differences in 5-year survival when stratifying by clinical primary tumor status, clinical lymph node status, or clinical stage may reflect the inaccuracy of the pretreatment staging modalities used in this study. Approximately one-third of the patients in this study underwent staging with cervical mediastinoscopy. Ultimately, we acknowledge that the inclusion of a relatively heterogeneous population of patients could have affected the survival, given that patients with a truly lower stage received neoadjuvant therapy (eg, T3 N0 thought to be T3 N2). However, all patients were deemed clinically advanced by our multidisciplinary group or thought to be probable R1, R2, or unresectable without neoadjuvant therapy. Other study limitations include the use of different chemotherapy regimens. Given the length of our retrospective study, this limitation reflects the different regimens typically used in our study protocol. Lastly, our standard dose of radiation was 4500 cgy, which was administered in split-course fashion to avoid the risk of radiation toxicity. Not all radiotherapy was delivered at our institution; therefore, there were some variations from the 4500 cgy dosage; in fact, a few patients received radiation dosages to 6000 cgy. Despite the limitations of this study, we must point out that the purpose of this study was not to evaluate the effect of downstaging with chemoradiation therapy (per se) but to evaluate survival from a perspective of the postoperative final pathologic stage. Therefore, we contend that the preoperative invasive staging provided less relevant information than the postoperative final pathologic stage. That is, our data suggest that resection in the presence of positive N2 lymph nodes after treatment should not preclude an attempted complete resection. The differences in survival among patients with single-station or multiple-station lymph node involvement were not significant in this study. A number of studies have concluded that multiple-station N2 lymph node involvement negatively affects survival. However, this finding was derived from studies that did not use neoadjuvant therapy [25 28]. The significance of this specific difference after induction therapy has yet to be completely elucidated. Port and colleagues [14] showed that the presence of multilevel N2 nodal disease did not confer a worse prognosis than single-station disease. Our previous work suggested that the performance of a right-sided pneumonectomy in the presence of persistently positive mediastinal lymph nodes after neoadjuvant chemoradiation therapy should be approached with caution or avoided because of the modest survival benefit and increased perioperative morbidity and mortality [4]. Although the difference was not significant in this study, there was a consistent trend of patients with right-sided pneumonectomies having a worse survival than patients with all other resections when positive N2 lymph nodes were present. Patients undergoing lobectomy, even in the presence of positive mediastinal (N2) lymph nodes, had a respectable 5-year survival of 34%. Almost equally impressive was that the overall 5-year survival of all patients with positive N1 or N2 lymph nodes still had a 5-year survival of 35%. Irrespective of resection type, the presence of singlestation vs multiple-station mediastinal lymph node involvement did not affect survival. In conclusion, respectable survival can be achieved after neoadjuvant chemoradiation, followed by anatomic resection, in selected patients with clinically advanced NSCLC. A T0 primary tumor or N0 lymph node status individually, or combined as a complete response (T0 N0) status, is associated with the best long-term survival. As would be expected, survival is most favorable for lobectomies vs pneumonectomies after neoadjuvant chemoradiation therapy.

9 Ann Thorac Surg KIM ET AL 2011;92: NEOADJUVANT CHEMORADIATION FOR NSCLC 241 References 1. Faber LP, Kittle CF, Warren WH, et al. Preoperative chemotherapy and irradiation for stage III non-small cell lung cancer. Ann Thorac Surg 1989;47:669 75; discussion Eagan RT, Ruud C, Lee RE, Pairolero PC, Gail MH. Pilot study of induction therapy with cyclophosphamide, doxorubicin, and cisplatin (CAP) and chest irradiation prior to thoracotomy in initially inoperable stage III M0 non-small cell lung cancer. Cancer Treat Rep 1987;71: Skarin A, Jochelson M, Sheldon T, et al. Neoadjuvant chemotherapy in marginally resectable stage III M0 non-small cell lung cancer: long-term follow-up in 41 patients. J Surg Oncol 1989;40: Kim AW, Faber LP, Warren WH, et al. Pneumonectomy after chemoradiation therapy for non-small cell lung cancer: does side really matter? Ann Thorac Surg 2009;88:937 43; discussion Mansour Z, Kochetkova EA, Santelmo N, et al. Persistent N2 disease after induction therapy does not jeopardize early and medium term outcomes of pneumonectomy. Ann Thorac Surg 2008;86: Bonomi P, Rowland K%, Taylor SG, et al. Phase II trial of etoposide, cisplatin, continuous infusion 5-fluorouracil, and simultaneous split-course radiation therapy in stage III nonsmall-cell bronchogenic carcinoma. Semin Oncol 1986;13: Recine D, Rowland K, Reddy S, et al. Combined modality therapy for locally advanced non-small cell lung carcinoma. Cancer 1990;66: Reddy S, Faber LP, Baumann LA, et al. Preoperative radiation therapy in regionally localized stage III non-small cell lung carcinoma: long-term results and patterns of failure. Int J Radiat Oncol Biol Phys 1990;19: Taylor MA, Reddy S, Lee MS, et al. Combined modality treatment using BID radiation for locally advanced nonsmall cell lung carcinoma. Cancer 1994;73: Edelman MJ, Suntharalingam M, Burrows W, et al. Phase I/II trial of hyperfractionated radiation and chemotherapy followed by surgery in stage III lung cancer. Ann Thorac Surg 2008;86: Krasna MJ, Gamliel Z, Burrows WM, et al. Pneumonectomy for lung cancer after preoperative concurrent chemotherapy and high-dose radiation. Ann Thorac Surg 2010;89:200 6; discussion Suntharalingam M, Paulus R, Edelman MJ, et al. RTOG 0229: A phase II trial of neoadjuvant therapy with concurrent chemotherapy and high-dose radiotherapy (XRT) followed by resection and consolidative therapy for LA-NSCLC. J Clin Oncol 2010;28:15S (Abstract 7024). 13. Bueno R, Richards WG, Swanson SJ, et al. Nodal stage after induction therapy for stage IIIA lung cancer determines patient survival. Ann Thorac Surg 2000;70: Port JL, Korst RJ, Lee PC, et al. Surgical resection for residual N2 disease after induction chemotherapy. Ann Thorac Surg 2005;79: Lorent N, DeLeyn P, Lievens Y, et al. Long-term survival of surgically staged IIIA-N2 non-small-cell lung cancer treated with surgical combined modality approach: analysis of a 7-year prospective experience. Ann Oncol 2004;15: Dooms C, Verbeken E, Stroobants S, et al. Prognostic stratification of stage IIIA-N2 non-small-cell lung cancer after induction chemotherapy: a model based on the combination of morphometric-pathologic response in mediastinal nodes and primary tumor response on serial 18-fluoro-2-deoxyglucose positron emission tomography. J Clin Oncol 2008;26: Cerfolio RJ, Maniscalco L, Bryant AS. The treatment of patients with stage IIIA non-small cell lung cancer from N2 disease: who returns to the surgical arena and who survives. Ann Thorac Surg 2008;86:912 20; discussion Cerfolio RJ, Bryant AS, Jones VL, Cerfolio RM. Pulmonary resection after concurrent chemotherapy and high dose (60Gy) radiation for non-small cell lung cancer is safe and may provide increased survival. Eur J Cardiothorac Surg 2009;35:718 23; discussion Deslauriers J, Gregoire J, Jacques LF, et al. Sleeve lobectomy versus pneumonectomy for lung cancer: a comparative analysis of survival and sites or recurrences. Ann Thorac Surg 2004;77:1152 6; discussion Bauman JE, Mulligan MS, Martins RG, et al. Salvage lung resection after definitive radiation ( 59 Gy) for non-small cell lung cancer: surgical and oncologic outcomes. Ann Thorac Surg 2008;86:1632 8; discussion Takeda S, Maeda H, Okada T, et al. Results of pulmonary resection following neoadjuvant therapy for locally advanced (IIIA-IIIB) lung cancer. Eur J Cardiothorac Surg 2006;30: Carretta A, Ciriaco P, Melloni G, et al. Results of surgical treatment after neoadjuvant chemotherapy for stage III non-small cell lung cancer. World J Surg 2008;32: Jaklitsch MT, Herndon JE 2nd, DeCamp MM Jr, et al. Nodal downstaging predicts survival following induction chemotherapy for stage IIIA (N2) non-small cell lung cancer in CALGB protocol #8935. J Surg Oncol 2006;94: Silvestri GA, Gould MK, Margolis ML, et al. Noninvasive staging of non-small cell lung cancer: ACCP evidencedbased clinical practice guidelines (2nd edition). Chest 2007; 132: S. 25. Cerfolio RJ, Bryant AS. Survival of patients with unsuspected N2 (stage IIIA) nonsmall-cell lung cancer. Ann Thorac Surg 2008;86:362 6; discussion Inoue M, Sawabata N, Takeda S, et al. Results of surgical intervention for p-stage IIIA (N2) non-small cell lung cancer: acceptable prognosis predicted by complete resection in patients with single N2 disease with primary tumor in the upper lobe. J Thorac Cardiovasc Surg 2004;127: Misthos P, Sepsas E, Kokotsakis J, Skottis I, Lioulias A. The significance of one-station N2 disease in the prognosis of patients with nonsmall-cell lung cancer. Ann Thorac Surg 2008;86: Tanaka F, Yanagihara K, Otake Y, et al. Prognostic factors in resected pathologic (p-) stage IIIA-N2, non-small-cell lung cancer. Ann Surg Oncol 2004;11: DISCUSSION DR JOHN A. HOWINGTON (Evanston, IL): Tony, nice talk. A question for you. You commented on the limitation. I was just going to ask you, in your own current practice, if you have a patient with clinical, either by positron emission tomography (PET) or computed tomography (CT), N2 disease, are you doing invasive mediastinal staging before doing chemoradiation therapy? And then just clarify for me, were you suggesting that if you do invasive mediastinal staging after chemoradiation and you have multi-station N2, you are still recommending resection, or are you leaving these patients alone? DR KIM: Thank you, Dr Howington for your comment and excellent questions. With respect to your first question, given the fallibility of imaging studies in consistently and accurately identifying N2 disease, we are using either endobronchial

10 242 KIM ET AL Ann Thorac Surg NEOADJUVANT CHEMORADIATION FOR NSCLC 2011;92: ultrasound (EBUS) or invasive mediastinoscopy to stage all our patients going on to multimodality therapy at this point, and that is what I do in my practice. With respect to the second question, I think that Dr Faber would, arguably, go ahead and move on with therapy with multistation lymph node involvement. For me, I think it is a judgment call. You have to put it together with the clinical picture: What is the degree of multinodal involvement? Currently where I practice, we try to do all our resections following induction therapy on protocol, and we don t have a protocol open for stage III disease with a surgical arm. We are just going to definitive chemoradiation. I think that there may be an argument based on the data that was just presented that if you have what you think are nonbulky lymph nodes with multistation involvement defined by only two nodal stations, it might be reasonable to consider. I again qualify statements by acknowledging that the data we presented is very biased data, and so I think it has to be interpreted cautiously. DR ROBERT J. CERFOLIO (Birmingham, AL): Just to follow up on that, would you do a pneumonectomy with multistation N2? DR KIM: I personally probably would not perform a pneumonectomy with multistation N2 disease. DR CERFOLIO: Okay. The next question is the radiation dose. We published, as has Dr Krasna, that if you give them a dose of 60 or 66 up front, since 20% or 30% aren t going to get resected, you have at least maximized their medical therapy; thus, why have you continued with your dose of 42? Have you thought about increasing the preoperative dose? DR KIM: I will let Dr Faber speak to the current Rush practice, but I think that they still continue to use doses to 4500 Gy. I can say that prior to my arrival at Yale, the thoracic oncology program used doses of just over 60 Gy, and I believe, they did not observe any significant changes in their outcomes. DR CERFOLIO: And the final question is, some are now moving more towards using chemo alone instead of chemoradiation. And I think the reason is because of testing biology. Now that we are getting more chemo-specific and targeted therapy based on some of our molecular markers, maybe we can judge biological response better without adding the radiation to the field that treats locally. I know I don t know the right answer, but I still favor 6600 cgy of radiation, but I understand why some are moving towards chemo alone. I would like your opinion if you think it makes sense to start moving more towards just using chemo alone vs chemoradiation, because I am sure that we don t know the answer. What do you think? DR KIM: That is a great point. I think it is reasonable to consider moving forward with targeted therapies for reasons we can discuss later. Combining this philosophy with an EBUS-driven stage IIIA-N2 disease trial could be an example of where this approach would be worthwhile and really benefit the patient. In general, profiling lymph node positive patients to facilitate their targeted therapy would seem to be a very logical approach. DR CERFOLIO: And then put the radiation on adjuvantly as they are doing in Pittsburgh. It is just another way to do it that we are moving a little bit towards. DR MARK J. KRASNA (Towson, MD): I am sorry, Dr Kim, I wasn t going to do this, but Cerf opened the door. So you have stepped backwards in time. You are going to throw away 20 years of data now where we have learned that there is actually a benefit to using a definitive dose of chemotherapy, definitivedose radiation, and definitive surgery. You might as well just do a robotic wedge resection, Cerf, for these patients and then just throw in targeted therapy. It is going to cure everything just by taking targeted therapy. Last week at the American Society for Radiation Oncology, we presented the results of Radiation Therapy Oncology Group (RTOG) 0229, which was a prospective phase II trial using carboplatin and taxol with radiation of 60 Gy in all 60 patients. There was one death, and that included patients who had a pneumonectomy, not just a lobectomy. So it can be done. Actually, your data support that the patients that had a complete pathologic response did better! So if more is better, get more; give more patients more radiation. If you could achieve a higher pathologic complete response rate by giving more of the patients more than 45 Gy, I know Pen Faber could operate on those. Perhaps Cerf can t do that with the robot, but if he goes back doing it open, he could even operate on those patients after high-dose radiation. That would be my first question. My second question is that actually there is a very important subgroup and that is the patients with T3 N1s. This is now the second paper to talk about these patients, Ben Daly s group also reported on this early this year. There are no prospective data now on T3 N1s, but they always are mixed in with these. Pen, I would urge you and your colleague, Dr Kim, to look at this subset, because this is now going to be the second largest series where you treated T3 N1 patients with neoadjuvant chemotherapy and radiation. I would like to know if you were able to tease that group out, because looking at your curves, you never pulled out the T3 N1s: you had T3 alone and N1 alone. If the T3 N1s in fact did as well as a subgroup and you didn t need to do pneumonectomies, that actually might be a new paradigm. I am curious about your answer to both of those. DR KIM: Simply put, I think your first point is an excellent one and I am in agreement with you. With regard to your second point, we did not tease out the differences the T3 N1 patients, but I think that is a subset of patients that deserve a closer look. For the purposes of the presentation, we eliminated several other slides because there was too much information, including the survival of the T3 and N1 patients individually. We, however, did not evaluate patients with T3 N1 disease specifically, but I think we should in the future. DR BILL PUTNAM (Nashville, TN): Two clinical questions. The first is a patient who has confirmed N2 disease, induction chemoradiotherapy, and now N0 by restaging. The operation is begun, and you find that you will have to do a pneumonectomy. Would you do the pneumonectomy or would you stop and back out? DR KIM: So they had N2 disease? DR PUTNAM: At initial staging, then they had chemoradiotherapy, and now they are restaged as N0, the best that you can tell. DR KIM: Okay. DR PUTNAM: You do the operation for the resection, thinking it will require only a lobectomy; however, you discover you will

11 Ann Thorac Surg KIM ET AL 2011;92: NEOADJUVANT CHEMORADIATION FOR NSCLC 243 have to do a pneumonectomy. Would you do a pneumonectomy in that patient or would you back out? DR KIM: I think this question is a little bit of a loaded question, but the short and qualified answer is that I would do the pneumonectomy, but only if I could achieve a complete resection. DR PUTNAM: Even with a 13% mortality and a limited survival? DR KIM: Correct. I think our data and other data show that perioperative mortality is weighted by the right pneumonectomy. That is why I think your question is a loaded one. DR PUTNAM: I would probably disagree with that. DR KIM: May I ask you what if it was a left pneumonectomy? DR PUTNAM: I think the data would suggest that the pneumonectomy would probably carry an added incremental risk and no benefit over and above that benefit already established with chemotherapy and radiation. DR KIM: I would beg to differ in that I think that with right pneumonectomy, you may be correct regarding the added incremental risk, but may not be entirely correct with left pneumonectomy. I think there is a difference, having recently reviewed the data over the past 20 years. From a survival standpoint, I think your points do warrant additional consideration. DR PUTNAM: And I do need to ask one more question, if I could just beg the moderators for one moment. A healthy patient with mediastinoscopy and confirmed single station N2 disease: In this patient, lobectomy would be needed. Given the single station N2 disease would you recommend definitive chemoradiation therapy? Resection only? Or resection followed by chemotherapy, as the International Adjuvant Lung Cancer Trial (IALT) proposed? I enjoyed the talk. Thank you. DR KIM: You have two options. I believe proceeding with surgery first is not what I would do. The evidence shows that survival is poor with surgically discovered N2 disease. I think that either induction therapy, followed by lobectomy or definitive therapy, are reasonable options.