ORIGINAL ARTICLE. Abstract INTRODUCTION

Transcription

1 European Journal of Cardio-Thoracic Surgery 43 (2013) e71 e81 doi: /ejcts/ezs622 Advance Access publication 4 January 2013 ORIGINAL ARTICLE a b c d Long-term results in patients with pathological complete response after induction radiochemotherapy followed by surgery for locally advanced non-small-cell lung cancer Filippo Lococo a, *, Alfredo Cesario a,b, Stefano Margaritora a,c, Valentina Dall Armi d, Francesca Mattei d, Rosalia Romano a, Venanzio Porziella a and Pierluigi Granone a Department of General Thoracic Surgery, Catholic University, Rome, Italy Scientific Direction, IRCCS San Raffaele Pisana, Rome, Italy Centro Oncologico Fiorentico (C.F.O.), Florence, Italy Unit of Clinical and Molecular Epidemiology, IRCCS San Raffaele Pisana, Rome, Italy * Corresponding author. Department of General Thoracic Surgery, Catholic University, Largo F.Vito 1, Rome, Italy. Tel: ; filippo_lococo@yahoo.it (F. Lococo). Received 26 June 2012; received in revised form 20 September 2012; accepted 24 September 2012 Abstract OBJECTIVES: The outcome of locally advanced non-small-cell lung cancer (NSCLC) patients with pathological complete response ( pcr) pt0n0 after induction chemoradiotherapy (IT) followed by surgery has, to date, only rarely been investigated. The long-term results in this highly selected subset of patients were evaluated and reported here to identify any predictive factors associated with prognosis. METHODS: From January 1992 to December 2009, 195 consecutive locally advanced (T1 T4/N0 2/M0) NSCLC patients underwent IT, and after clinical restaging, 137 were operated upon with radical intent. Among these, 37 (19% of the overall and 27% of the surgical cohort) showed a pcr status and were included in this retrospective analysis. Survival rates and prognostic factors were analysed by the Kaplan Meier, the log-rank and Cox regression analyses. RESULTS: The mean age and male/female ratio were 61.9 ± 9.8 years and 33/4, respectively. Before starting IT, the clinical staging was IIb in 2 (5%) patients, IIIa in 20 (54%) and IIIb in 15 (41%). Morbidity and 30-day mortality rates were 27 and 3%, respectively. The overall 3- and 5-year long-term survivals (LTSs) and disease-free survival (DFS) were 67 and 64% and 68 and 71%, respectively. Overall, 17 patients (46%) experienced a recurrence, occurring more frequently in a distant site (32%) than locally (19%). The analysis of the 5- year LTS suggests that (i) the initial single N2 station involvement (P = 0.010); (ii) the resection to a lesser extent than pneumonectomy (P = 0.005) and (iii) the adjuvant therapy (P = 0.005) are all positive prognostic factors. In particular, a 5-year hazard ratio of 8.21 (95% confidence interval , P = 0.002) was estimated by Cox regression analysis for subjects who did not undergo adjuvant therapy vs those who did. CONCLUSIONS: After induction radiochemotherapy followed by surgery in locally advanced NSCLC, a pcr is achieved in a remarkable proportion of cases (27% in our experience). In such patients, a rewarding LTS (64% at 5 years) could be expected, especially when a single N2 station is involved at diagnosis or when an adjuvant treatment is administered. Nevertheless, recurrences after surgery are quite common (46%) and this evidence deserves further investigations and deeper analysis. Keywords: Non-small-cell lung cancer Induction therapy Radiotherapy Surgery INTRODUCTION Locally advanced Stage III non-small-cell lung cancer (NSCLC) represents a therapeutic challenge with variable options [1]. In the American Association of Chest Physicians (AACP) evidencebased clinical practise guidelines, it is stated that in NSCLC with N2 disease identified preoperatively (IIIA) induction therapy followed by surgery is NOT recommended except as part of a clinical trial [2]. Nevertheless, even considering that a positive Presented at the 20th European Conference on General Thoracic Surgery, Essen, Germany, June prognostic impact of surgery in multimodality treatment is not yet completely validated or still based on a small number of trials and patients [3], in numerous institutions such as our own, surgery is the most accepted treatment after induction concurrent radiochemotherapy (IT) in the subset of patients with a complete or partial response to the IT itself or those with a stable disease and evidence of technical feasibility of a resection with radical intent [4]. In fact, there is consistent evidence that downstaging after IT is directly associated with an improvement of disease-free survival (DFS) and a lower distant recurrence rate [5]. The Author Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.

2 e72 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery In detail, the patients operated upon after a favourable clinical restaging who show a pathological complete response ( pcr) are not so numerous. However, these patients are indeed the optimal candidates to undergo surgery after IT since they achieve good to very good long-term survivals (LTSs) [4, 6]. To the best of our knowledge, there is no report about any in-depth analysis performed in this sub-class of patients: information is therefore incomplete and many important issues are still debated. The aim of this study was to assess and report about a comprehensive analysis of the clinical and pathological characteristics of patients with pcr after surgery who had been administered concurrent induction radiochemotherapy. Data on long-term result in terms of LTS and information on prognostic factors are described in detail and discussed. MATERIALS AND METHODS We conducted a retrospective review of a prospective lung cancer database to select the clinical records of 37 locally advanced (non-n3) NSCLC patients who, upon a favourable clinical restaging status, were operated on following IT in the period between January 1992 and December Institutional Review Board approval had been preliminarily obtained for the use of data stemming from standard clinical practise for research purposes. No additional interventions were planned (retrospective observational study). Based on the information available from the clinical records, demographic and clinical features were collected and taken into consideration in the statistical analysis. Follow-up data were obtained from our database or, in some cases, by direct telephonic interview ( privacy-related issues were covered by comprehensive informed consent) with patients or, in the case of a death, with representatives from the family (next of kin). All patients had at least 1 year of follow-up. Inclusion criteria to the multimodal treatment were previously reported [4, 7]. In particular, the patients with multilevel mediastinal nodal disease were treated with induction (neoadjuvant) intent only if a complete surgical resection was deemed feasible before the induction therapy was administered (inoperable but resectable cases). In the present study, we have also excluded patients with contralateral nodal metastases (N3), extranodal extension at mediastinoscopy, unresectable T4 disease ( pleural effusion and oesophageal invasion) or severe cardiopulmonary impairment precluding surgical resection. Staging was performed according to the process detailed in [7]. In brief, all patients had consistent and thorough clinical and pathological staging (before IT) and (clinical) restaging before operation. In particular, in all cases, mediastinal involvement was always pathologically proven (or excluded in those cases with pn0 N1) before IT by endoscopy (trans-bronchial/transoesophageal) or via collar mediastinoscopy or anterior procedures (Chamberlain s mediastinotomy). During the study period, the positron emission tomography (PET) scan was not always available in our institution; however, from 2005, this has been used for staging purposes in selected patients. Similarly, all patients underwent homogeneous restaging procedures (substantially based on imaging technique computed tomography (CT) or PET/CT features more than on invasive or mini-invasive procedures) and surgical operations. One physician (P.G.) coordinated the multidisciplinar activities around the staging (clinical Table 1: Descriptive statistics for the sample (n = 37) Patients characteristics n (%) Sex, n (%) Females 4 (11%) Males 33 (89%) Age (mean ± SD) 61.9 ± 9.8 ctnm, n (%) IIB 2 (5%) IIIA 20 (54%) IIIB 15 (41%) ct, n (%) T1 3 (8%) T2 5 (14%) T3 14 (38%) T4 15 (41%) cn, n (%) N0 4 (11%) N1 2 (5%) N2 31 (84%) Lymph-node involvement, n (%) (N = 31) Single station 10 (32%) Multiple stations 21 (68%) Induction chemoradiotherapy protocols (number of patients enrolled (%)) CHT CBDCA 7 (19%) CDDC + 5-FU 9 (24%) CDDC 21 (57%) RT (50.4 Gy) Standard: 180 cgy daily 13 (35%) Accelerated: 120 cgy twice daily 24 (65%) Toxicity, n (%) No 19 (51%) Yes 18 (49%) Days from IT to surgery (mean ± SD) 45.0 ± 23.3 ytnm, n (%) 0 4 (11%) IA 7 (19%) IB 6 (16%) IIA 3 (8%) IIB 4 (11%) IIA 7 (19%) IIB 6 (16%) Resection type, n (%) Lobectomy 26 (70%) Pneumonectomy 7 (19%) Bilobectomy 4 (11%) Histology, n (%) Adenocarcinoma 11 (30%) Adenosquamous carcinoma 1 (3%) Squamous cells carcinoma 25 (68%) Lymph-node removed (mean ± SD) 9.1 ± 8.1 Number of lymph-node removed, n (%) (51%) >6 18 (49%) Morbidity, n (%) No 27 (73%) Yes 10 (27%) Adjuvant therapy, n (%) None 15 (41%) CHT alone 14 (38%) RT alone 2 (5%) CHT-RT 6 (16%) Status at follow-up, n (%) Alive with no evidence of disease 15 (41%) Alive with disease 4 (11%) Deceased for cancer 14 (38%) Deceased for other causes 4 (11%) CHT: chemotherapy; CBDCA: carboplatinum; CDDC: cisplatinum; 5-FU: 5-Fluoracil; RT: radiotherapy.

3 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery e73 and pathological) and restaging processes and performed or monitored all the surgical operations. This fact, by reducing most of the confounding factors often affecting retrospective analyses, warrants the quality and homogeneity of the information collected. Induction therapy Procedures and schedules regarding the IT protocols are extensively described in [7]; in addition to minor changes in the selection of drugs used uniformly platinum-based compounds (CBDCA and CDDP) plus 5 fluorouracil (5-FU) up to the introduction of gemcitabine (GEM) in the substitution of 5-FU (Table 1) there has been a change from 2- to 3-dimensional conformal radiotherapy methodology in The total administered RT dose was 50.4 Gy in all cases with the classic or hyper-fractionated schedules. In every case, the treatment was supported by CT planning. A complete clinical and radiological restaging was performed 4 6 weeks after the last pulse of the treatment. Clinical restaging status (ytnm) was assigned using the modified World Health Organization criteria [8]. Surgery When indicated, surgery was performed on average 2 weeks (from 10 to 17 days) after the clinical restaging procedures (thus, approximately 6/7 weeks after the last pulse of the IT). A parenchymal resection to an extent lesser than a lobectomy was considered as oncologically inappropriate and never performed. Systematic mediastinal lymph-node dissection of levels 2, 4, 7 and 10 was performed in all cases. Regardless of the gross nodal involvement, Stations 5 and 6 were removed in the case of left upper lobe lesions, whereas a modal dissection of Station 9 was performed after lower lobectomy. All the mediastinal tissue containing the lymph nodes was dissected and removed systematically within typical anatomical landmarks. The resection was considered complete (R0) if the proximal resection margins and the highest mediastinal resected node were free of tumour. Pathological review was performed using standard techniques, and immunohistochemical staining was used when appropriate. Surgical pathological stages were assigned according to the sixth TNM classification system and the International Staging System for Lung Cancer [9], and data from patients observed and treated before 1997 have been updated to coherently match the sixth TNM and to obtain homogeneous staging information through the entire cohort. In particular, a pathological response was defined as complete when no viable tumour cells were detected in the surgical specimen: a final pathological Stage 0 (T0/N0/M0) was then assigned. Patients with only microscopic foci of disease (<0.5 cm in diameter) were considered to have a residual T1 disease and were excluded from the study group. With respect to the surgical and pathological features, we collected details on the extent of resection, the pathological staging, the completeness of resection, the mediastinal stations involved, the number of the lymph nodes removed and the perioperative morbidity as well as mortality rates. Statistical analysis Patients characteristics were described by mean, standard deviation, median, minimum and maximum for continuous variables, and frequencies and percentages for categorical variables. Furthermore, the summary statistics were stratified by the outcome measures: mortality; total, local and distant recurrence. Time-to-event analysis, at 3 and 5 years from surgery, was performed on these outcomes by means of Kaplan Meier survival curves and log-rank tests. From surgery, it was performed on these outcomes by means of Kaplan Meier survival curves and log-rank tests. Median survival (or DFS) ± standard error, 95% confidence intervals (CIs) and P-values have been provided, considering the entire time of follow-up. When the upper bound of the 95% CI lay outside the observed timeframe, the last time of follow-up was reported. Multivariable analysis was performed with the Cox proportional hazards regression model to assess the prognostic role of several variables on patients survival and development of recurrence. The following variables were considered as potential confounders in the multivariable analysis: age, sex, ctnm, tumour size, lymph-node involvement (single or multiple station), toxicity, ytnm, resection type, histology, adjuvant therapy. Despite the limited sample size, the parameters to be considered for the multivariable analysis were chosen in accordance with the clinical team, based on their impact on the mortality and recurrence of the tumour. Final models included only actual confounders identified through a forward selection procedure. For all tests, P-value <0.05 was regarded as statistically significant. Statistical analyses were performed using SPSS for Windows (version 13.0; SPSS, Inc., USA) and STATA software package (STATA/SE version 10.0, Stata Corp., College Station, TX, USA). RESULTS From January 1992 to December 2009, a total of 195 consecutive locally advanced (T1 T4/N0 2/M0) NSCLC patients entered multimodal treatment with induction concurrent chemotherapy and radiotherapy (IT group). As showed in the consort-type diagram (Fig. 1), after restaging, 137 patients (in good clinical and functional conditions) with partial/complete response or stable disease after IT were explored surgically. Among these, at pathological staging, 37 patients (19% of the overall and 27% of those subject to surgery) showed a pcr and their data were considered eligible for the purposes of this retrospective analysis. The mean age and male/female ratio within this group were 61.9 ± 9.8 years and 33/4, respectively. Table 1 shows data regarding demographic, clinical, surgical characteristics and staging of patients before IT (TNM), after IT (ytnm) and after surgery (ptnm). Before IT, the majority of patients (35 patients, 95%) were in cstage III, apart from 2 cases in cstage IIb (both staged as ct3n0m0). Moreover, patients were mostly enrolled to multimodal treatment due to N2-status (31 patients, 84%) and the remaining (6 patients, 16%), advanced T-status. Toxicity was generally mild and entirely manageable without treatment interruption, although nausea and vomiting were experienced by most, this datum clarifying the high overall toxicity rate (49%). Apart from gastrointestinal symptoms, the most common significant treatment-related toxicity encountered was

4 e74 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery therapy. Regarding the 3 major complications (8%), 1 patient experienced myocardial infarction requiring specific care, surveillance in intensive care unit and long-term rehabilitation, 1 had lobar pneumonia, which required 72 h of mechanical ventilation and lastly 1 had postoperative bleeding treated with surgical re-thoracotomy performed in emergency. After surgery, 22 (59%) patients underwent adjuvant therapies (Table 1). The administration of additional anticancer treatments after surgery was left to the discretion of the treating medical oncologist. The patients received either chemotherapy (14 patients; 38%), radiation (2 patients; 5%) or both (6 patients; 16%). Adjuvant chemotherapy consisted of platinum-based protocols (two or three courses); in detail, 9 patients received docetaxel and cisplatin, 8 received GEM and cisplatin and 3 received GEM and carboplatin combinations. Exclusive radiation therapy was inclusive of a boost of 20 Gy (total dose) at 2.5 3Gyperday. Figure 1: Consort-type diagram of the study group. RT: radiotherapy; CHT: chemotherapy. Table 2: The pattern of failure after surgery following IT in patients with complete pathological response Local recurrence Distant recurrence 1st relapse site, n (%) Brain 5 Liver 3 Bone 2 Contralateral lung 1 Adrenal gland 1 Ipsilateral lung 3 Mediastinal 2 Bronchial stump 2 Surgical approach Pneumonectomy (n = 7) 0 (0%) 3 (43%) (Bi)lobectomy (n = 30) 7 (23%) 9 (30%) Adjuvant therapy None (n = 15) 3 (20%) 8 (53%) Chemotherapy alone 3 (21%) 3 (21%) (n = 14) Radiochemotherapy (n = 6) 1 (17%) 1 (17%) Radiotherapy alone (n = 2) 0 (0%) 0 (0%) Total 7 (19%) a 12 (32%) a a In 2 cases, local and distant relapses occurred simultaneously. myelosuppression. No patient experienced a severe-grade toxicity (G4), neither neurotoxicity nor nephrotoxicity. As expected, the restaging process (substantially based on radiological CT and PET-CT findings procedures) was somewhat inaccurate, where only 11% of all patients were correctly restaged as Stage 0 and about 35% assigned to Stage III. Pneumonectomy was performed in 7 (19%) patients only and squamous cell carcinoma was the most frequent histotype at the pathological evaluation (Table 1). Overall, the 30-day mortality rate was 3% (1 case of respiratory failure associated with bronchopleural fistula after right pneumonectomy). Postoperative complications occurred in 10 patients (overall morbidity rate = 27%). Among the 7 minor complications (19%), 3 were supra-ventricular arrhythmias (medically treated), 2 were a persistent air-leak and the remaining 2 cases consisted of lobar pneumonia that required prolonged antibiotic Pattern of recurrence after surgery The patterns of recurrence after surgery are described here and summarized in Table 2. Overall, during the follow-up period, the disease recurred in 17 patients (overall recurrence rate = 46%). Among those who experienced a recurrence, the mean (±standard deviation) follow-up time was 35 ± 33.4 months. In most cases (12 patients), the disease relapsed more often in a distant site (distant recurrent rate = 32%) than locally (7 patients, local recurrent rate = 19%). The most common site of relapse was the brain (Table 2). Two patients simultaneously experienced a relapse of disease locally and at distant sites, 12 and 48 months, respectively, after surgery. Interestingly, patients who underwent pneumonectomy presented a different pattern of recurrence after surgery when compared with patients who underwent lobectomy or bilobectomy (Table 2). Indeed, while the overall recurrence rate was similar (43 after pneumonectomy vs 53% after (bi)lobectomy), the local recurrence rate was substantially different (local recurrence rate: 23 vs 0%); in this setting, we could speculate about the achievement of a different control of disease according to the extension of the resection (better local control after pneumonectomy), this in turn very probably influenced the rationale of adoption of different therapeutic strategies in the adjuvant setting. The pattern of recurrence was also influenced by the administration of adjuvant therapy. In particular, of 20 patients who underwent adjuvant chemotherapy (administered alone in 14 patients and combined with radiotherapy in 6, Table 2), only 4 experienced a distant relapse (distant recurrence rate = 20%), while a recurrence occurred in 8 of the 17 who did not undergo systemic treatment after surgery (distant recurrence rate = 47%). The recurrent disease was prevalently treated with nonsurgical therapies (mostly by systemic treatment), although in 3 cases with very long-term disease-free intervals (defined as the period between the pulmonary resection and the time of recurrence), a surgical resection (1 adrenalectomy, 1 single brain metastases resection and 1 pulmonary wedge resection ) was carried out. Long-term outcomes and prognostic factors The follow-up period for mortality and tumour relapse started from the first day after surgery for all patients and ended at 167

5 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery e75 months (13 years and 10 months) with 1 patient still alive and without evidence of disease. The overall mean (±standard deviation) follow-up duration was 59.9 ± 38.5 months. The median time for survival was 86 ± 15.8 months (95% CI ), and 80 ± 15.4 months for recurrence (95% CI ). Three- and 5-year estimated overall survival (LTS) rates for the entire cohort of patients were 67 (95% CI 49 80%) and 64% (95% CI 46 77%) as reported in Fig. 2A. Similarly, 3- and 5-year estimated overall DFS rates were 68 (95% CI 50 81%) and 61% (95% CI 43 76%), respectively (Fig. 3A). Survival rates calculated according to clinical, surgical, pathological and postoperative variables (Table 3) demonstrated (P-values from the log-rank test) that a better outcome could be expected in patients with initial N2 single-level involvement when compared with patients with initial N2 multiple-level involvement in terms of both 5-year LTS (100 vs 50%, P = 0,010) and DFS (90 vs 50%, P = 0,031) as reported in Fig. 2B. The median LTS time for the group of patients with singlelevel involvement was at 127 ± 14.8 months (95% CI ), while the last death in the multilevel involvement group was observed at 37 months of follow-up, with 11 patients at risk (50.4% survival; P = 0.136). The median DFS time for the group with single-level involvement was 124 ± 12.5 (95% CI ) vs 71 ± 37.5 months (95% CI ) for the multilevel group (P = 0.043). The type of resection also influenced the outcome of patients with pcr; in fact, patients who underwent pneumonectomy had a poorer 5-year LTS compared with those subjects who underwent lobectomy/bilobectomy (5-year LTS: 72 vs 29%, P = 0.005; Fig. 2C), but no significant differences were found in terms of 5-year DFS (5-year LTS: 72 vs 42%, P = 0.121). The median LTS time for the group of subjects who had pneumonectomy was 17 ± 9.2 months (95% CI 1 66), while the group who had bi/lobectomy had a median LTS time of 127 ± 2.6 months (95% CI ; P = 0.014). The median DFS time for the two groups was 24 ± 14.9 (95% CI 5 167) and 80 ± 9.5 months (95% CI ; P = 0.581), respectively. Moreover, a better long-term outcome could be predicted in patients who underwent adjuvant treatment compared with those who did not (5-year LTS: 86 vs 40%, P = 0.001, Fig. 2D; 5-year DFS: 81 vs 46%, P = 0.031, Fig. 3B). On the contrary, no differences in long-term outcomes were found when comparing other selected factors (Table 3). The median LTS time for the adjuvant therapy group was 127 ± 9.6 months (95% CI ), while that for the non-adjuvant group was 27 ± 6.4 months (95% CI ; P = 0.016). The median DFS time for the two groups was 124 ± 7.3 (95% CI ) and 24 ± 12.3 months (95% CI 12 72; P = 0.002), respectively. From the Cox multivariable regression analysis (Table 4), it emerged that adjuvant treatment was an independent prognostic factor in such patients. In particular, a 5-year hazard ratio (HR) for mortality of 8.21 (95% CI ; P = 0.002) and a 5-year HR for recurrence of 5.43 (95% CI ; P = 0.007) were estimated for the group of patients who did not undergo adjuvant therapy after surgery vs those patients who did. Moreover, local and distant DFSs were proven to be slightly different in patients with pcr according to the administration of adjuvant therapy (Fig. 3C and D). In detail, adjuvant therapy (mostly composed by systemic treatment in our sample, see Table 4) seems to have guaranteed a better control of disease at Figure 2: Overall LTS function of all cohort of patients (A) and according with lymph nodal mediastinal involvement at diagnosis (B), surgical resection (C) and adjuvant treatment (D).

6 e76 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery Figure 3: Overall DFS function of all cohort of patients (A) and according with adjuvant treatment (B D). distance. In fact, the 5-year HR of a distant recurrence was estimated as 3.24 (95% CI ; P = 0.073) times higher in patients who did not have adjuvant therapy than those who did have administration of adjuvant therapy. DISCUSSION Approximately 30% of patients with NSCLC present with Stage IIIA: of these, 32% are lost to distant metastasis and one-third due to insufficient local control [1]. The best strategy of care in these patients is still a forum of open debate as demonstrated by the very numerous number of studies recently published on this topic. Starting from the observation that neither primary surgery ( poor survival as reported in historical series [10]) nor a combination of radiotherapy and chemotherapy without surgical resection [11] could achieve satisfactory long-term outcomes, in the early 1990s, several teams, including ourselves [4], have investigated the role of surgery after induction therapy. The belief was that surgery positively impacted on the chances of achieving loco-regional control in locally advanced diseases, and these experiences have brought encouraging results. As reported by the EORTC multicenter randomized trial [12], the clinical rationale for induction therapy in locally advanced NSCLC was four-fold: (i) the integrity of tumoural vascularization provided a better drug supply to the tumour before any surgery; (ii) regression of the primary cancer and mediastinal involvement could be achieved, thereby facilitating and simplifying or reducing subsequent surgery; (iii) undetected micrometastases could be dealt with at the start of treatment and (iv) the inhibition of the surgical stimulus to any residual cancer was facilitated. Therefore, multimodal therapeutic strategies including surgery performed after full chemoradiation schedules (administered with doses and schedules intended for exclusive treatment [6]) have spread thereafter, probably also due to the advances in surgical technique and postoperative care that have made these approaches feasible with acceptable morbidity. Thus, although a positive prognostic impact of surgery in the multimodality treatment is not yet completely validated or still based on a relatively small number of trials and patients [3], there is some evidence that a complete surgical resection in the subset of patients with a complete or partial response to the IT is associated with the best LTS prognosis as reported in the IASLC Consensus Report published in 2003 [13] (although IT was proposed in the context of controlled clinical trials only). Therefore, following the line of extreme simplification, it is possible to state that surgery should be performed in patients with a complete/partial response after IT only (although salvage surgery for persistent N2 disease after IT may be sometimes considered in selected cases [6, 7]). In this setting, the restaging process after IT has gained an increasing interest in the scientific community, its role being very pivotal to identify who returns to the surgical arena [6]. Even considering the whole spectrum of procedures (from non-invasive radiological tools to more invasive surgical approaches), the restaging assessment often failed to predict the actual pathological response in such patients. In particular, standard staging procedures, such as CT and magnetic resonance imaging, are of limited value in the evaluation of objective response after IT because of their intrinsic inability to differentiate between viable tumour and necrosis or scarring on a morphologic basis, thus resulting in low accuracy in staging either the primary tumour or mediastinal lymph nodes. As well, PET-CT scan provides additional metabolic information on

7 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery e77 Table 3: Overall survival and disease-free survival at 3 and 5 years from surgery; P-value of log-rank test for the comparison of survival curves Overall survival Disease-free survival 3 years (Ns = 25) 5 years (Ns = 24) 3 years (Ndfs = 27) 5 years (Ndfs = 26) Survival % (95% CI) P-value Survival % (95% CI) P-value Survival % (95% CI) P-value Survival % (95% CI) P-value Age <65 years (n = 18) 61% (35 79) % (30 75) % (30 75) % (25 70) years (n = 19) 73% (46 88) 73% (46 88) 87% (59 97) 87% (59 97) Sex Females (n = 4) 75% (13 96) % (13 96) % (13 96) % (13 96) Males (n = 33) 66% (47 79) 63% (43 77) 70% (51 83) 66% (46 80) ctnm IIB (n = 2) 100% ( ) % ( ) % ( ) % ( ) IIIA (n = 20) 63% (38 0.8) 63% (38 81) 69% (44 85) 62% (35 80) IIIB (n = 15) 67% (37 85) 60% (32 80) 69% (37 87) 69% (37 87) ct T1 (n = 3) 100% ( ) % ( ) % ( ) % ( ) T2 (n = 5) 40% (5 75) 40% (5 75) 60% (13 88) 60% (13 88) T3 (n = 14) 70% (38 87) 70% (38 87) 71% (39 88) 62% (31 82) T4 (n = 15) 67% (37 85) 60% (32 80) 69% (37 87) 69% (37 87) cn N0 (n = 4) 75% (13 96) % (13 96) % (13 96) % (13 96) N1 (n =2) 0%( ) 0% ( ) 0% ( ) 0% ( ) N2 (n = 31) 70% (51 83) 67% (47 81) 76% (55 88) 71% (50 84) T-size (mm) <60 (n = 15) 73% (44 89) % (44 89) % (50 93) % (39 88) (n = 17) 69% (41 86) 69% (41 86) 80 (51 93) 80% (51 93) Lymph-node involvement Single station (n = 10) 100% ( ) 100% ( ) 100% ( ) 90% (47 98) Multiple stations (n = 21) 55% (32 74) % (27 69) % (32 76) % (26 71) Number of lymph-node removed, n (%) 1 6 (n = 19) 72% (46 87) 67% (40 83) 67% (41 84) 61% (35 79) >6 (n = 18) 61% (35 79) % (35 79) % (41 86) % (34 81) Resection type Pneumonectomy (n = 7) 43% (10 73) % (4 61) % (6 77) % (6 77) Bi/lobectomy (n = 30) 72% (52 85) 72% (52 85) 76% (56 88) 72% (51 85) Histology Adenocarcinoma (n = 11) 54% (23 78) % (23 78) % (23 78) % (23 78) Adenosquamous carcinoma (n = 1) 100% ( ) 100% ( ) 100% ( ) 100% ( ) Squamous cells carcinoma (n = 25) 71% (49 85) 67% (45 82) 78% (54 90) 72% (48 87) Adjuvant therapy No (n = 15) 40% (16 63) % (12 56) % (19 70) % (19 70) Yes (n = 22) 86% (62 95) 86% (62 95) 86% (63 95) 81% (57 92) Adjuvant therapy type None (n = 15) 40% (16 63) % (12 56) % (19 70) % (19 70) RT alone (n = 2) or CHT alone (n = 16) 87% (56 96) 87% (56 96) 88% (59 97) 79% (48 93) RT-CHT (n = 6) 83% (27 97) 83% (27 97) 83% (27 97) 83% (27 97) Recurrence No (n = 21) 85% (60 95) % (60 95) Yes (n = 17) 47% (23 68) 41% (18 63) Ns: number of survivors; Ndfs: number of disease-free survivors; CHT: chemotherapy; RT: radiotherapy. N-stage of the tumour, but it s role in restaging is still controversial. Consequently, we still do not have a clear understanding of the value of the response to the IT when this is followed by surgery and, thus, the process of selection of best candidates for surgery stays, in the best scenario, sub-optimal. In spite of these limitations, worthy of mention is the fact that patients treated with IT who achieving a pcr at surgery show very rewarding survival patterns [14 21] (Table 5). The aim of this study was to extensively analyse the long-term outcome and to investigate the patterns of failure after surgery and the prognostic factors in this particular group of patients. First of all, our data suggest that pcr may be obtained in a significant proportion of patients treated with an IT protocol (27% of all patients who underwent surgery in our experience), and this result may be achieved with acceptable treatment-related toxicity and surgical morbidity. The addition of radiation therapy (especially when an high-dose treatment is administered) in induction therapy protocols is considered to impact more positively the chances of obtaining a complete response at the moment of surgery than those administered on chemotherapy-only based regimens (Table 5). This evidence has been also reported in the recent ESMO clinical practise guidelines [22] ( the rate of

8 e78 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery Table 4: Hazard ratios (HR [95% CI]) for mortality and disease recurrence, estimated at 3 and 5 years from surgery; P-value from Cox regression Mortality Disease recurrence 3 years (Nd = 3) 5 years (Nd = 5) 3 years (Nr = 10) 5 years (Nr = 11) HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value HR (95% CI) P-value 0.91 ( ) ( ) ( ) < ( ) <0.001 Adjuvant therapy b No 6.87 ( ) ( ) Yes ( ) ( ) Recurrence c No 1 1 Yes d 5.40 ( ) ( ) Disease recurrence locally Disease recurrence at distance 3 years (Nr = 3) 5 years (Nr = 5) 3 years (Nr = 10) 5 years (Nr = 11) Adjuvant therapy HR (95% CI) P-value HR (95% CI) P-value HR a (95% CI) P-value HR a (95% CI) P-value None (n = 15) 3.62 ( ) ( ) ( ) ( ) Yes (n = 22) CHT (n = 14) RT-CHT (n =6) RT (n =2) Nd: number of deceased; Nr: number of subjects with recurrence; CHT: chemotherapy; RT: radiotherapy. a Adjusted for adjuvant therapy. b Adjusted for age (continuous). c Adjusted for adjuvant therapy and age (continuous). d Proportionality of hazards assumption not satisfied (OS 3 years: recurrence: P = 0.021; OS 5 years: recurrence: P = 0.030). Table 5: Incidence and LTS in NSCLC patients with pcr after induction therapy Trial No. of patients Disease stage Induction therapy protocol Surgery (%) pcr (%) a LTS (5 years) Machtay et al. [14] ( ) 45 T1 4/N2 CHT + RT (total dose: Gy) N/A (48% in pcr + ppr) DeCamp et al. [15] ( ) 105 T1 4/N2 3 CHT + RT b (total dose: 60 Gy) N/A Friedel et al. [16] ( ) 74 T1 4/N2 3 CHT + RT (total dose: 45 Gy) % Garrido et al. [17] ( ) 136 T1 4/N2 3 CHT 70 9 N/A (52% in pn0) Decaluwe et al. [5] ( ) 92 T1 4/N2 CHT N/A 5 N/A (53% in pn0) Steger et al. [18] ( ) 55 T1 4/N2 3 CHT + RT (total dose: 45 Gy) % Cerfolio et al. [19] ( ) 216 T1-4/N0 3 CHT + RT (total dose: 60 Gy) N/A Kim et al. [20] ( ) 233 T1 4/N0 2 CHT + RT (total dose: 43 Gy) % Paul et al. [21] ( ) 136 T1 3/N2 CHT alone (87%) CHT + RT (13%) N/A 1 N/A Our series ( ) 137 T1 4/N2 CHT + RT (total dose: >50 Gy) % CHT: chemotherapy; RT: radiotherapy; pcr: pathological complete response; LTS: long-term survival. a % of pcr patients underwent surgery. b Given before and after surgery. complete pathological remission with neoadjuvant chemotherapy is lower than with chemoradiotherapy ), probably due to a more effective loco-regional control (as demonstrated by the low local recurrence rates that are generally reported following the application of schemes including radiation therapy, including our own protocol). In this setting, Cerfolio et al. [6] have recently analysed the flow of patients enrolled in a multimodal protocol for locally advanced Stage III NSCLC. The authors reported that only about 30% of patients who underwent IT come back to the surgical arena ; they further speculated that there are several reasons why patients did not return for definitive surgical ( pathological) restaging. The most common was a lack of response to therapy. Thus, the real challenge in the multimodal strategy of care in locally advanced NSCLC is not exactly the LTS in responder patients who undergo surgery, but actually the intrinsic response to the IT itself. In other words, the search for the best effective induction protocol (optimal clinical response with acceptable risk) is crucial

9 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery e79 to extend to a larger part of subjects the benefits of the surgical resection in terms of better local control of the disease. Our protocol, although with some elements of heterogeneity due to the change of drugs and RT planning in the last two decades, was inspired by the results of the Southwest Oncology Group (SWOG) trial 8805 [23], and based on the consideration that radiation therapy combined with a platinum-based systemic treatment, and administered at full therapeutic doses even in the preoperative setting could be very effective in obtaining a partial/complete response in a significantly large percentage of subjects who underwent surgery and with acceptable toxicities. The SWOG trial, with figures dramatically different from those reported in other similar experiences, reported that 85% of Stage IIIA patients and 80% of IIIB patients could undergo surgery after IT. Our data substantially confirm this finding: in fact, on 195 consecutive locally advanced NSCLC patients who were enrolled in the IT protocol, 137 (70%) were surgically explored and, among these, 37 (27%) showed a pcr. This evidence, along with the good LTS rates in this group (64% at 5 year), is remarkable when compared with data from similar scenarios available in the literature (Table 5 and ranging from 48 to 67%), and when matched against the very poor survival related with the stage at the moment of diagnosis (ciiia and IIIb). In particular, in a large series of 233 Stage III NSCLC, analysed retrospectively by Kim et al. [20], the reported long-term outcome after a trimodality treatment (induction radiochemotherapy followed by anatomical surgical resection) with many similarities to our own shows a pcr of 22% with a good 58% LTS, this being significantly higher than that of patients with residual diseases (P = 0.012). No further stratification has been realized. In such a similar scenario, we pushed the analysis a bit further and could outline how the pattern of mediastinal lymph nodal involvement at diagnosis ( multiple stations vs single station ) stays as a strong prognostic determinant. This is also true when a pcr is achieved independently at the moment of diagnosis (with IT before surgery Table 3 and Fig. 2B). Indeed, patients with initial mediastinal involvement of a single lymph nodal station showed an impressive 5-year LTS when compared with those with multiple-station involvement at diagnosis (100 vs 50%, P = 0.010; DFS: 90 vs 50%, P = 0.031). Indeed, our data stem from a small population, selected with strict criteria and treated with a scheme that can be easily defined as relatively aggressive compared with those reported in the literature. These features must be kept in mind when giving a clinical value to the reported figures. Moreover, what is to be discussed further about the results of our observation is the fact that 15 patients (a good 41% of the entire cohort) were Stage IIIb at the moment of diagnosis. Of course, these cases were well within the characteristics of those described to have a better prognosis in the IIIb group, sub-classes (selected T4s, N0s) on which we have extensively speculated in a very recently reported experience [24]. On a different aspect, an accurate analysis of the survival data matched with the pattern of failure after surgery may highlight some critical points in the strategy of care of the patients whose data have been analysed. The overall local recurrence proportion is substantially acceptable (19%), while the distant control of the disease is quite poor (32%), with a fairly high occurrence of brain metastases (14%). In detail, only 1 patient (of the 5 patients who had brain metastases) had received adjuvant chemotherapy after surgery and none of the 5 was submitted to prophylactic cranial irradiation (PCI). Despite the fact that PCI has been indicated in a large multicenter randomized trial [25] as a valuable tool to produce a reasonable reduction in the brain metastasis rate even in IT protocols, and considering that our data seem to confirm the clinical rationale of this treatment even in pcr patients, not enough data are available, so far, to draw definitive conclusion in this setting. More generally, the role of adjuvant treatment in patients who underwent radiochemotherapy before surgery is unclear and its value and impact not completely understood. In our population, a significantly better outcome has been observed in those treated with adjuvant therapy. However, these figures come from an observational exercise and suffer from the lack of a randomized approach (5-year LTS: 86 vs 33%, P = 0.001, Fig. 2D; 5-year DFS: 81 vs 46%, P = 0.031, Fig. 3B). From a theoretical point of view, and based on the evidence reported so far, we can assume that a strategy of care based on a trimodal sequential protocol (RT + CHT then surgery then CHT) could represent the most effective treatment in patients with locally advanced Stage III NSCLC. A similar strategy was proposed and reported by DeCamp et al. [15]. In detail, the authors have evaluated the feasibility and efficacy of a multimodal protocol consisting of hyper-fractionated radiotherapy with concurrent chemotherapy ( paclitaxel and cisplatin) followed by resection and postoperative chemoradiation for IIIA/B NSCLC patients. With respect to feasibility, 77% of patients completed all therapy; a pcr was observed in 11% of patients and a partial response in 41%. Survival of patients with pn0 was more than 55% at 5-year; finally, the authors reported a very low loco-regional recurrence rate (about 9%), while distant recurrence rate topped up to 45% (brain metastases occurring in half of these cases). Despite these encouraging results in terms of pcr% and LTS (substantially confirmed by our experience), even considering the clinical issue related to so much aggressive planning of care, we do not recommend any therapeutic algorithm, but strongly advocate a prospective randomized methodology in order to clarify the efficacy of adjuvant (chemo and/or radio and/or PCI) therapy even in the situation where an induction therapy with chemoradiation followed by surgical resection allows the subject to obtain a pcr. We are aware of the fact that a strategy with such a degree of aggressiveness could slightly limit the number of patients completing the study, but we believe that a selection (based on strict clinical criteria) is, de facto, not a limitation but a strength of the multimodal strategy that should be indicated only in carefully selected patients. Limitations and strength Our report has the usual limitations of retrospective monocentric studies (the long duration of patients inclusion and the limited absolute number of patients). After IT, an invasive restaging was not systematically performed in our series. Moreover, the lack of defined randomization algorithms in the inclusion criteria to apply adjuvant therapy after surgery represents a further significant limitation. Finally, we wish to mention other two shortcomings, restricted to the clinical impact of the present study: (i) all clinical, biological or histological predictors of pathological response detectable at the time of diagnosis, during induction therapy or before a planned resection remain still elusive; (ii) the inaccuracy of the restaging procedures limited the detection of the actual (real) rate of patients with pcr.

10 e80 F. Lococo et al. / European Journal of Cardio-Thoracic Surgery Consequently, the prognostic stratification may be done only on the basis of the pathological assessment of the disease and not before IT administration or before surgery. Despite all the limitations mentioned above, this study investigates (in a large long-term monocentric clinical setting) a highly selected group of NSCLC patients who underwent IT followed by surgery and had a pcr. This adds, in our opinion, new information potentially useful to tailor a likely better strategy of care for these patients. CONCLUSIONS In spite of the very encouraging results in multimodality treatment for locally advanced NSCLC, the challenge still remains in identifying the best IT protocol able to obtain high response rates with acceptable associated toxicity. The preoperative selection of patients who really benefit most from surgical resection is still unclear as well. Our data suggest that a combined induction chemoradiotherapy approach may obtain a high rate of complete pathological response (27% of the surgical group), and a very rewarding survival (5-year LTS up to 64%) could be expected in such patients, especially if a single mediastinal station was involved at diagnosis. Moreover, after surgery, adjuvant chemotherapy seems to be associated with a better LTS, particularly in terms of better control of the distant disease. A possible approach based on a trimodal four-step sequential protocol (RT + CHT followed by surgery followed by CHT) needs to be validated by large prospective, controlled clinical trials. Conflict of interest: none declared. REFERENCES [1] Takeda S, Maeda H, Okada T, Yamaguchi T, Nakagawa M, Yokota S et al. Results of pulmonary resection following neoadjuvant therapy for locally advanced (IIIA-IIIB) lung cancer. Eur J Cardiothorac Surg 2006;30: [2] Alberts WM, American College of Chest Physicians. Diagnosis and management of lung cancer executive summary. Chest 2007;132:S1 19. [3] Song WA, Zhou NK, Wang W, Chu XY, Liang CY, Tian XD et al. Survival benefit of neoadjuvant chemotherapy in non-small cell lung cancer: an updated meta-analysis of 13 randomized control trials. J Thorac Oncol 2010;5: [4] Trodella L, Granone P, Valente S, Margaritora S, Macis G, Cesario A et al. Neoadjuvant concurrent radiochemotherapy in locally advanced (IIIA-IIIB) non small cell lung cancer: long term results according to downstaging. Ann Oncol 2004;15: [5] Decaluwe H, De Leyn P, Vansteenkiste J, Dooms C, Van Raemdonck D, Nafteux P et al. Surgical multimodality treatment for baseline resectable stage IIIA-N2 non-small cell lung cancer. Degree of mediastinal lymph node involvement and impact on survival. Eur J Cardiothorac Surg 2009; 36: [6] 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: [7] Meacci E, Cesario A, Cusumano G, Lococo F, D Angelillo R, Dall armi V et al. Surgery for patients with persistent pathological N2 IIIA stage in non-small-cell lung cancer after induction radio-chemotherapy: the microscopic seed of doubt. Eur J Cardiothorac Surg 2011;40: [8] Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer 1981;47: [9] Mountain CF. Revisions in the international system for staging lung cancer. Chest 1997;111: [10] Martini N, Flehinger BJ, Zaman MB, Beattie EJ Jr. Prospective study of 445 lung cancer carcinomas with mediastinal lymph node metastases. J Thorac Cardiovasc Surg 1980;80: [11] Ginsberg RJ, Vokes EE, Raben A. Non-small cell lung cancer. In: DeVita VT Jr, Hellman S, Rosenberg SA (eds). Cancer: Principles and Practice of Oncology, 5th edn. Philadelphia, PA: Lippincott, 1997 pp [12] Gilligan D, Nicolson M, Smith I, Groen H, Dalesio O, Goldstraw P et al. Preoperative chemotherapy in patients with resectable non-small cell lung cancer: results of the MRC LU22/NVALT 2/EORTC multicentre randomised trial and update of systematic review. Lancet 2007;369: [13] Eberhardt WE, Albain KS, Pass H, Putnam JB, Gregor A, Assamura H et al. Induction treatment before surgery for non-small cell lung cancer: IASLC consensus report. Lung Cancer 2003;42:S9 14. [14] Machtay M, Lee JH, Stevenson JP, Shrager JB, Algazy KM, Treat J et al. Two commonly used neoadjuvant chemoradiotherapy regimens for locally advanced stage III non-small cell lung carcinoma: long-term results and associations with pathologic response. J Thorac Cardiovasc Surg 2004;127: [15] DeCamp MM, Rice TW, Adelstein DJ, Chidel MA, Rybicki LA, Murthy SC et al. Value of accelerated multimodality therapy in stage IIIA and IIIB non-small cell lung cancer. J Thorac Cardiovasc Surg 2003;126: [16] Friedel G, Budach W, Dippon J, Spengler W, Eschmann SM, Pfannenberg C et al. Phase II trial of a trimodality regimen for stage III non-small-cell lung cancer using chemotherapy as induction treatment with concurrent hyperfractionated chemoradiation with carboplatin and paclitaxel followed by subsequent resection: a single-center study. J Clin Oncol 2010; 28: [17] Garrido P, González-Larriba JL, Insa A, Provencio M, Torres A, Isla D et al. Long-term survival associated with complete resection after induction chemotherapy in stage IIIA (N2) and IIIB (T4N0 1) non small-cell lung cancer patients: the Spanish Lung Cancer Group Trial J Clin Oncol 2007;25: [18] Steger V, Walles T, Kosan B, Walker T, Kyriss T, Veit S et al. Trimodal therapy for histologically proven N2/3 non-small cell lung cancer: mid-term results and indicators for survival. Ann Thorac Surg 2009;87: [19] 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 723. [20] Kim AW, Liptay MJ, Bonomi P, Warren WH, Basu S, Farlow EC et al. Neoadjuvant chemoradiation for clinically advanced non-small cell lung cancer: an analysis of 233 patients. Ann Thorac Surg 2011;92:233 41; discussion [21] Paul S, Mirza F, Port JL, Lee PC, Stiles BM, Kansler AL et al. Survivalof patients with clinical stage IIIA non-small cell lung cancer after induction therapy: age, mediastinal downstaging, and extent of pulmonary resection as independent predictors. J Thorac Cardiovasc Surg 2011;141: [22] Crino L, Weder W, van Meerbeeck J, Felip E, On behalf of the ESMO Guidelines Working Group. Early stage and locally advanced (nonmetastatic) non-small-cell lung cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2010;21 (Suppl. 5):v [23] Albain KS, Rusch VW, Crowley JJ, Rice TW, Turrisi AT 3rd, Weick JK et al. Concurrent cisplatin/etoposide plus chest radiotherapy followed by surgery for stages IIIA (N2) and IIIB non-small-cell lung cancer: mature results of Southwest Oncology Group phase II study J Clin Oncol 1995;13: [24] Lococo F, Cesario A, Margaritora S, Dall armi V, Nachira D, Cusumano G et al. Induction therapy followed by surgery for T3-T4/N0 non-small cell lung cancer: long-term results. Ann Thorac Surg 2012;93: [25] Eberhardt W, Wilke H, Stamatis G, Stuschke M, Harstrick A, Menker H et al. Preoperative chemotherapy followed by concurrent chemoradiation therapy based on hyperfractionated accelerated radiotherapy and definitive surgery in locally advanced non-small cell lung cancer: mature results of a phase II trial. J Clin Oncol 1998;16: APPENDIX. CONFERENCE DISCUSSION Dr S. Barnett (Melbourne, Australia): I think I want to ask the same question as Dr De Leyn asked the last presenter. In a retrospective study where you are drawing a conclusion that adjuvant chemotherapy is improving survival, when the chemotherapy was administered at the discretion of the oncologist; can you comment on the basis for your conclusion? One of the reasons for that improved survival is because of the patient who died in hospital, or the