E(

SKa

b

a

ARRA

KNLRPMS

1

lia[atpt

h0

Lung Cancer 84 (2014) 156–160

Contents lists available at ScienceDirect

Lung Cancer

jou rna l h om epa ge: www.elsev ier .com/ locate / lungcan

vidence supporting contemporary post-operative radiation therapyPORT) using linear accelerators in N2 lung cancer

uchit H. Patela, Yan Mab, A. Gabriella Wernickea, Dattatreyudu Noria,.S.C. Chaoa, Bhupesh Parashara,∗

Stich Radiation Center, New York Presbyterian Hospital/Weill Cornell Medical Center, 525 East 68th Street, New York, NY 10065, United StatesDivision of Biostatistics and Epidemiology, Weill Cornell Medical College, 1300 York Avenue, New York, NY 10021, United States

r t i c l e i n f o

rticle history:eceived 29 November 2013eceived in revised form 10 February 2014ccepted 25 February 2014

eywords:2ungadiationORTeta-analysis

urvival

a b s t r a c t

Purpose: Post-operative radiotherapy (PORT) treatment for lung cancer declined since a meta-analysisfailed to show benefit in patients with N2 disease. Because several included studies employed out-moded radiation planning and delivery techniques, we sought to determine whether PORT with moderntechnology benefits patients with N2 disease.Methods: We conducted searches of the published literature. For inclusion, studies must have includedpatients with stage III-N2 lung cancer treated with PORT using only linear accelerators, used a controlgroup that did not receive PORT, and reported outcome data for overall survival (OS). Prospective andretrospective analyses were included. Exclusion criteria were the use of cobalt devices or orthovoltageradiation.Results: Data were evaluated with random-effects models. Three prospective and eight retrospectivestudies were included. The PORT and no-PORT groups included 1368 and 1360 patients, respectively.The PORT group had significantly improved OS over the no-PORT group (hazard ratio [HR] = 0.77, 95%confidence interval [CI] 0.62–0.96, P = 0.020). Locoregional recurrence-free survival (LRFS) in 10 studiesfor which data was available was also improved in the PORT group (HR = 0.51, CI 0.41–0.65, P < 0.001).

Conclusions: PORT was associated with significantly lower risk of death and locoregional recurrence inpatients with N2 lung cancer. Our study was limited by lack of access to individual patient data, whichwould have enabled more detailed analyses. Regardless, data thus far suggest PORT may be associatedwith a survival benefit. Given a lack of large-scale prospective data, clinical trials evaluating PORT withmodern technology are warranted.

© 2014 Elsevier Ireland Ltd. All rights reserved.

. Introduction

The standard of care for patients with early stage non-small cellung cancer (NSCLC) is surgical resection [1]. On pathologic stag-ng, a sizeable portion of these patients, as well as those with moredvanced disease, are found to have N2 lymph node involvement2], which results in higher rates of local recurrence and worse over-ll survival (OS) as compared with N0 or N1 patients [3,4]. Givenhis risk, postoperative radiation therapy (PORT) has long been pro-

osed as a way to reduce the risk of local recurrence in N2 diseasehat undergo surgical treatment.

∗ Corresponding author. Tel.: +1 212 746 3612.E-mail address: bup9001@med.cornell.edu (B. Parashar).

ttp://dx.doi.org/10.1016/j.lungcan.2014.02.016169-5002/© 2014 Elsevier Ireland Ltd. All rights reserved.

A meta-analysis published in 1998 that analyzed data from 2343patients with stage I to III NSCLC showed PORT reduced OS in theentire cohort [5]. Subgroup analyses confirmed this adverse effectfor N0 and N1 disease, but favored PORT for N2 disease, though sta-tistical significance was not reached. Following the meta-analysis,use of PORT declined from 65% of patients in 1999 to 37% in 2002.Given that many of the included trials in the analysis used radia-tion planning and delivery techniques that are no longer consideredcurrent standard (i.e. cobalt machines, single field delivery, largedaily fractions), there has been considerable debate whether PORTdelivered with more modern techniques with linear accelerators(LINAC) using two- and three-dimensional computed tomography

(CT) based planning might be beneficial in patients with N2 dis-ease. Furthermore, relative to most of the included trials in thePORT meta-analysis, adjuvant platinum-based chemotherapy hasbeen incorporated into standard treatment in stage IIIA disease

Cance

oc

ad

2

2

aocsprbdbirwwws

2

Oe

2

tswcpnu

3

3

EpwtEdSrmst

wort

S.H. Patel et al. / Lung

nly recently [6–8] and the role of PORT in the setting of adjuvanthemotherapy in this population remains unclear.

Therefore, we performed a meta-analysis of the current liter-ture to evaluate PORT in the context of modern planning andelivery in the treatment of N2 disease.

. Methods

.1. Analyzed studies

To identify potentially suitable studies, we conducted a system-tic search of Medline, ClinicalTrials.gov, Cochrane Central Registerf Controlled Trials, and Embase for terms relating to PORT and lungancer. Criteria for inclusion were that patients included in eachtudy must have N2 disease, their outcome must have been com-ared to a control group, and that OS data for both groups wereeported. Both prospective and retrospective studies were eligi-le. Furthermore, studies must have used linear accelerators forelivery of external beam radiation and no patients could haveeen treated with equipment using Co-60, since such treatment

s known to result in poorer outcomes [9] and is no longer usedoutinely. Radiation must have been delivered post-operativelyith or without chemotherapy. Each study’s corresponding authoras contacted for individual patient data; however, individual dataere not available for nearly all studies, and thus, only summary

tatistics were analyzed.

.2. Outcome measures

The primary outcome measures sought were hazard ratios forS and local recurrence-free survival (LRFS). Where possible, dis-ase free survival (DFS) was also determined.

.3. Statistical analysis

Data were analyzed using Stata version 12.1 (StataCorp) usinghe metan package. Statistical heterogeneity across the varioustudies was tested with the I2 statistic and was taken into accounthen pooling HRs using random effects models. Forest plots were

reated to show the estimated HRs from component studies and theooled estimate of HR. When HRs were not reported in a compo-ent study, we estimated the HRs with the Cox regression P-valuessing existing statistical methods [10,11].

. Results

.1. Included studies

Characteristics of the included studies are reported in Table 1.leven studies were included in the final analysis. Three wererospective [12–14] and eight were retrospective [15–22]. Sevenere excluded from analysis, usually because studies did not fit

he radiation delivery inclusion criteria. Because the Surveillance,pidemiology, and End Results (SEER) database does not report theetails of radiation treatment and planning, two analyses usingEER did not meet inclusion criteria [23,24]. However, one ret-ospective analysis of the SEER database was included in oureta-analysis because patients were matched with planning and

imulation codes from Medicare physician claims that were usedo determine the complexity of radiation therapy planning [21].

Our analysis includes a total of 2728 patients with N2 disease in

hich 1360 (49.9%) received PORT between 1982 and 2005. A total

f 1026 patients (37.6% of total patients included in this analysis)eceived chemotherapy, usually platinum-based, across seven ofhe included analyses. Where reported, all of the included studies

r 84 (2014) 156–160 157

employed 2 Gy/fraction to deliver a total dose of 45–60 Gy. Targetvolumes typically covered the bronchial stump, ipsilateral hilum,and the mediastinum.

3.2. Effect of PORT on OS, LRFS and DFS

There was significant heterogeneity in the outcomes from eachstudy (I2 = 71.6%) that led us to use a random-effects model for theanalysis [25]. As shown in Fig. 1, we found that PORT was asso-ciated was a significantly improved overall survival (HR = 0.77, CI0.62–0.96, P = 0.020). A significant association with improved OSwas also found on fixed-effects analysis (HR = 0.90, CI 0.82–0.99,P = 0.027). LRFS data were available for ten of the included studies,and showed that PORT was associated with a significant reductionin local recurrence (HR = 0.51, CI 0.41–0.65, P < 0.001; Fig. 2). DFS,defined as time to recurrence or death, was available for only fivestudies, and indicated a significant benefit from PORT (HR = 0.66, CI0.54–0.79, P < 0.001; Fig. 2).

3.3. Radiation toxicity

Numerical data regarding radiation toxicity were not wellreported. Toxicity was most commonly encountered in the formof mild esophagitis, dysphagia, and odynophagia. Cough, andpneumonitis requiring steroid therapy were the most commonpulmonary toxicities. Radiation myelitis was also reported in onepatient. No severe late complications were noted in any of theincluded studies, though this was likely limited by the follow-upduration.

4. Discussion

Our pooled analysis of retrospective and prospective studiessuggests there may be a benefit in OS with the use of PORT withlinear accelerators, conformal radiation fields and standard frac-tionation in patients with resected N2 non-small cell lung cancer.

However, our study has several limitations. First, the lack ofindividual patient data precluded detailed analysis of several inter-esting outcomes, such as disease free survival or association ofradiation toxicity with death or other adverse events. Though wesought individual data from each study, they were unavailable inmost cases. Second, there is a paucity of high-quality randomizedtrials evaluating PORT using current standards of radiation deliveryin conjunction with chemotherapy. Indeed we foresaw this prob-lem and elected to include retrospective analyses, which representthe bulk of the patient population presented here, and which mightprovide rationale for future prospective trials to be conducted.Nonetheless, we cannot rule out bias in the included studies thatmay have skewed our analysis, such as that inherent to patientselection, baseline characteristics and physician preferences in theretrospective studies included in this analysis. Lastly, publicationbias may have precluded the inclusion of potentially relevant trialsthat may have failed to find a strong overall effect of PORT. The latteris at least partially mitigated by the fact that the three prospectivestudies included in this analysis each had failed to show a signifi-cant OS benefit with PORT, as did three of the included retrospectivestudies.

Regardless, our analysis yields several key points. Foremost, itsuggests that PORT has a beneficial role in the treatment of N2disease when treatment is carried out using LINACs with mod-ern planning and standard fractionation. This is not surprising, andaccords with the results of the previous meta-analysis. We interpret

our findings to imply that well-designed prospective trials mightdefinitively prove this benefit.

This leads to the second point: the included studies stronglyunderscore the need for high-quality prospective clinical trials

158 S.H. Patel et al. / Lung Cancer 84 (2014) 156–160

Table 1Characteristics of included studies.

Study Year Type Adjuvant treatment No PORTgroup (n)

PORTgroup (n)

Total dose Fractionation Comments

Mayer [13] 1997 RCT, P RT or O 26 23 56 Gy 2 Gy/f 2-field AP or 3-fielddelivery; customizedblocks;

Feng [12] 2000 RCT, P RT or O 44 61 60 Gy 2 Gy/f AP and oblique fields;Perry [14] 2007 RCT, P Paclitaxel and carboplatin, followed by

RT or O18 19 50 Gy 2 Gy/f

Douillard [16] 2008 Ret Cisplatin/vinorelbine or none, followedby RT or O

108 116 45–60 Gy 2 Gy/f

Matsuguma [18] 2008 Ret RT or O 46 45 25.2–63.9 Gy,50.4 Gy median

2 Gy/f 2D and 3D basedplanning

Moretti [19] 2009 Ret Cisplatin/paclitaxel orcarboplatin/paclitaxel, followed by RTor O

44 39 50–60 Gy,54 Gy median

Not reported 3-field (posterior andlateral); customizedblocks.

Du [17] 2009 Ret Variable regimen or none, followed byRT or O

255 104 50 Gy 2 Gy/f

Scotti [20] 2010 Ret RT or O 56 119 53 Gy median 2 Gy/fZou [22] 2010 Ret Cisplatin with either etoposide,

gemcitabine or paclitaxel, followed byRT or O

79 104 48–54 Gyrange, 50 Gymedian

2 Gy/f

Dai [15] 2011 Ret Cisplatin or paclitaxel, followed by RTor O

125 96 60 Gy 2 Gy/f

Wsinivesky [21] 2012 Ret Platinum-based, other chemotherapy 597 710 N/A N/A SEER and Medicare

A : radia

eawsbto(nb

isr

or none, followed by RT or O

bbreviations: RCT: randomized control trial; Ret: retrospective; P: prospective; RT

valuating PORT using either conformal 3D external beam radi-tion or intensity-modulated radiotherapy (IMRT) in conjunctionith chemotherapy. The last prospective study that fit our inclu-

ion criteria for this analysis ended in 2000, before chemotherapyecame the standard of care. Certainly the rationale for a newrial exists, as just shown by our analysis. Yet to our knowledge,nly two – the LUNG-ART trial in Europe and another in ChinaNCT00880971) – are underway, and results from LUNG-ART willot be available before 2021 [26]. Another trial desperately needse undertaken in the United States.

As far as we know, the only known attempt at a trial evaluat-ng PORT with chemotherapy was terminated early as a result oflow accrual despite opening of the trial at 68 institutions, thoughesults from the 37 patients enrolled were promising: median

Fig. 1. Meta-analysis of OS outc

analysis

tion therapy; O: observation; 2D: 2-dimensional; 3D: 3-dimensional.

failure-free survival was 16.8 months on the observation armversus 33.7 months on the radiotherapy arm [14]. The results didnot reach significance because the trial remained underpowered,primarily due to restrictive entry criteria that limited the trial tooccult N2 disease only that was demonstrated on surgical explo-ration or postoperative pathology analysis, and not with positivemediastinoscopy. However, a trial initiated today does not needto face such stringent inclusion criteria. Furthermore, compared topatients included in the previous studies included in the originalPORT meta-analysis in 1998, patients today are would be better

selected for surgery with PET-CT and brain imaging, and nearly allwould receive adjuvant chemotherapy.

Lastly, individual studies suggest that PORT delivered withLINACs and modern planning does not add excess intolerable

omes of included studies.

S.H. Patel et al. / Lung Cancer 84 (2014) 156–160 159

) from

tsaEPdcse[bontdo34t

Fig. 2. Meta-analysis of LRFS (top) and DFS (bottom

oxicity, one of the presumed reasons for its association withignificant morbidity in studies included in the original PORT meta-nalysis. The Eastern Cooperative Oncology Group (ECOG) trial3590 followed mortality in 488 patients treated with PORT versusORT and chemotherapy, and reported no increase in inter-currenteaths (DID) in the trial population when compared to matchedontrols [27]. A similar conclusion was reached in another retro-pective series comparing 202 radiotherapy-treated patients withxpected rates from vital statistics reported by Machtay et al.28]. We should note that this differs from the results reportedy Dautzenberg et al., a large randomized trial included in theriginal PORT meta-analysis, in which intercurrent deaths were sig-ificantly increased in the PORT group [29]. However, patients inhe Dautzenberg trial were treated with Co-60 devices, and largeraily fractions and a greater total dose than that used in E3590, all

f which may explain the greater rate of DID (5-year DID rate of1% in patients that received PORT in Dautzenberg, as opposed to-year DID rates of 12.9% and 13.5% in E3590 and Machtay, respec-ively).

studies for which published data were available.

In a prospective study comparing 171 N2 patients treated with3D-planned PORT against 120 N1 patients treated with no PORT,non-cancer related deaths, cardiopulmonary morbidity and qual-ity of life surveys again showed no significant differences [30].Additionally, two retrospective analyses have actually reportedimproved outcomes with 3D planning. One noted significantlyimproved LRFS and a trend toward improved OS in 35 patients thatreceived 3D-planned radiotherapy as compared to 65 that received2D-planned delivery [31]. In the second, a subgroup analysis ofMatsuguma et al. comparing 3D-planned patients with 2D-plannedpatients revealed significantly improved OS [18].

Thus, based on our analysis and other existing reports, we findan imminent need for a modern trial evaluating PORT. Lung cancerremains the leading cause of cancer deaths worldwide [32] andaccounts for 28% of deaths in the United States [33]. Over 226,000

new cases occur annually in the United States, 26% of which arestage III [34]. If PORT improves 5-year survival in stage III disease –presently at a dismal 24% – by only 10%, that represents nearly 4500deaths averted [35].

1 Cance

F

C

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[[

60 S.H. Patel et al. / Lung

unding

None.

onflicts of interest

No conflicts of interest from any author.

eferences

[1] Manser R, Wright G, Hart D, Byrnes G, Campbell DA. Surgery for early stagenon-small cell lung cancer. Cochrane Database Syst Rev 2005:CD004699.

[2] Koike T, Tsuchiya R, Goya T, Sohara Y, Miyaoka E. Prognostic factors in 3315completely resected cases of clinical stage I non-small cell lung cancer in Japan.J Thorac Oncol 2007;2:408–13.

[3] Andre F, Grunenwald D, Pignon JP, Dujon A, Pujol JL, Brichon PY, et al. Sur-vival of patients with resected N2 non-small-cell lung cancer: evidence for asubclassification and implications. J Clin Oncol 2000;18:2981–9.

[4] Rusch VW, Crowley J, Giroux DJ, Goldstraw P, Im JG, Tsuboi M, et al. The IASLCLung Cancer Staging Project: proposals for the revision of the N descriptorsin the forthcoming seventh edition of the TNM classification for lung cancer. JThorac Oncol 2007;2:603–12.

[5] Postoperative radiotherapy for non-small cell lung cancer. Cochrane DatabaseSyst Rev. John Wiley & Sons, Ltd.; 2005.

[6] Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon JP, Vansteenkiste J.International Adjuvant Lung Cancer Trial Collaborative G Cisplatin-based adju-vant chemotherapy in patients with completely resected non-small-cell lungcancer. N Engl J Med 2004;350:351–60.

[7] Hotta K, Matsuo K, Ueoka H, Kiura K, Tabata M, Tanimoto M. Role of adjuvantchemotherapy in patients with resected non-small-cell lung cancer: reap-praisal with a meta-analysis of randomized controlled trials. J Clin Oncol2004;22:3860–7.

[8] Sedrakyan A, Van Der Meulen J, O’Byrne K, Prendiville J, Hill J, Treasure T. Post-operative chemotherapy for non-small cell lung cancer: a systematic reviewand meta-analysis. J Thorac Cardiovasc Surg 2004;128:414–9.

[9] Phlips P, Rocmans P, Vanderhoeft P, Van Houtte P. Postoperative radiotherapyafter pneumonectomy: impact of modern treatment facilities. Int J Radiat OncolBiol Phys 1993;27:525–9.

10] Parmar MK, Torri V, Stewart L. Extracting summary statistics to performmeta-analyses of the published literature for survival endpoints. Stat Med1998;17:2815–34.

11] Tierney J, Stewart L, Ghersi D, Burdett S, Sydes M. Practical methods for incor-porating summary time-to-event data into meta-analysis. Trials 2007;8:16.

12] Feng QF, Wang M, Wang LJ, Yang ZY, Zhang YG, Zhang DW, et al. A studyof postoperative radiotherapy in patients with non-small-cell lung cancer: arandomized trial. Int J Radiat Oncol Biol Phys 2000;47:925–9.

13] Mayer R, Stuecklschweiger GF, Hackl A, Smolle-Juettner F-M, Friehs G, SzolarD, et al. Postoperative radiotherapy in radically resected non-small cell lungcancer. CHEST J 1997;112:954–9.

14] Perry MC, Kohman LJ, Bonner JA, Gu L, Wang X, Vokes EE, et al. A phase III studyof surgical resection and paclitaxel/carboplatin chemotherapy with or withoutadjuvant radiation therapy for resected stage III non-small-cell lung cancer:Cancer and Leukemia Group B 9734. Clin Lung Cancer 2007;8:268–72.

15] Dai H, Hui Z, Ji W, Liang J, Lu J, Ou G, et al. Postoperative radiotherapyfor resected pathological stage IIIA–N2 non-small cell lung cancer: a retro-spective study of 221 cases from a single institution. Oncologist 2011;16:641–50.

16] Douillard J-Y, Rosell R, De Lena M, Riggi M, Hurteloup P, Mahe M-A. Impactof postoperative radiation therapy on survival in patients with completeresection and stage I, II, or IIIA non-small-cell lung cancer treated with adju-vant chemotherapy: The Adjuvant Navelbine International Trialist Association(ANITA) Randomized Trial. Int J Radiat Oncol Biol Phys 2008;72:695–701.

[

r 84 (2014) 156–160

17] Du F, Yuan Z, Wang J, Zhao L, Pang Q, Gong L, et al. The role of postopera-tive radiotherapy on stage N2 non-small cell lung cancer. Chin J Lung Cancer2009;12:1164–8.

18] Matsuguma H, Nakahara R, Ishikawa Y, Suzuki H, Inoue K, Katano S, et al.Postoperative radiotherapy for patients with completely resected pathologicalstage IIIA-N2 non-small cell lung cancer: focusing on an effect of the numberof mediastinal lymph node stations involved. Interact Cardiovasc Thorac Surg2008;7:573–7.

19] Moretti L, Yu DS, Chen H, Carbone DP, Johnson DH, Keedy VL, et al. Prognosticfactors for resected non-small cell lung cancer with pN2 status: implicationsfor use of postoperative radiotherapy. Oncologist 2009;14:1106–15.

20] Scotti V, Meattini I, Saieva C, Agresti B, de Luca Cardillo C, Bastiani P, et al.Post-operative radiotherapy in N2 non-small cell lung cancer: a retrospectiveanalysis of 175 patients. Radiother Oncol 2010;96:84–8.

21] Wisnivesky JP, Halm EA, Bonomi M, Smith C, Mhango G, Bagiella E. Postop-erative radiotherapy for elderly patients with stage III lung cancer. Cancer2012;118:4478–85.

22] Zou B, Xu Y, Li T, Li W, Tang B, Zhou L, et al. A multicenter retrospective analysisof survival outcome following postoperative chemoradiotherapy in non–small-cell lung cancer patients with N2 nodal disease. Int J Radiat Oncol Biol Phys2010;77:321–8.

23] Lally BE, Zelterman D, Colasanto JM, Haffty BG, Detterbeck FC, Wilson LD.Postoperative radiotherapy for stage II or III non-small-cell lung cancerusing the surveillance, epidemiology, and end results database. J Clin Oncol2006;24:2998–3006.

24] Rescigno J. Use of postoperative radiotherapy for node-positive non-small-celllung cancer. Clin Lung Cancer 2002;4:35–44.

25] Huedo-Medina TB, Sánchez-Meca J, Marín-Martínez F, Botella J. Assessingheterogeneity in meta-analysis: Q statistic or I2 index. Psychol Methods2006;11:193–206.

26] Le Pechoux C. Role of postoperative radiotherapy in resected non-small celllung cancer: a reassessment based on new data. Oncologist 2011;16:672–81.

27] Wakelee HA, Stephenson P, Keller SM, Wagner H, Herskovic A, Komaki R, et al.Post-operative radiotherapy (PORT) or chemoradiotherapy (CPORT) followingresection of stages II and IIIA non-small cell lung cancer (NSCLC) does notincrease the expected risk of death from intercurrent disease (DID) in EasternCooperative Oncology Group (ECOG) trial E3590. Lung Cancer 2005;48:389–97.

28] Machtay M, Lee JH, Shrager JB, Kaiser LR, Glatstein E. Risk of death fromintercurrent disease is not excessively increased by modern postoperativeradiotherapy for high-risk resected non-small-cell lung carcinoma. J Clin Oncol2001;19:3912–7.

29] Dautzenberg B, Arriagada R, Chammard AB, Jarema A, Mezzetti M, Mattson K,et al. A controlled study of postoperative radiotherapy for patients with com-pletely resected nonsmall cell lung carcinoma. Groupe d’Etude et de Traitementdes Cancers Bronchiques. Cancer 1999;86:265–73.

30] Kepka L, Bujko K, Orlowski TM, Jagiello R, Salata A, Matecka-Nowak M, et al.Cardiopulmonary morbidity and quality of life in non-small cell lung cancerpatients treated with or without postoperative radiotherapy. Radiother Oncol2011;98:238–43.

31] Masson-Cote L, Couture C, Fortin A, Dagnault A. Postoperative radiotherapyfor lung cancer: improvement in locoregional control using three-dimensionalcompared with two-dimensional technique. Int J Radiat Oncol Biol Phys2011;80:686–91.

32] Ferlay J, Shin H, Bray F, Forman D, Mathers C, Parkin D. GLOBOCAN 2008 v2.0,Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 10. Lyon,France: International Agency for Research on Cancer; 2008.

33] Cancer Facts & Figures. American Cancer Society; 2012.34] Morgensztern D, Ng SH, Gao F, Govindan R. Trends in stage distribution for

patients with non-small cell lung cancer: a National Cancer Database survey. JThorac Oncol 2010;5:29–33.

35] Goldstraw P, Crowley J, Chansky K, Giroux DJ, Groome PA, Rami-Porta R, et al.International Association for the Study of Lung Cancer International Staging C,Participating I. The IASLC Lung Cancer Staging Project: proposals for the revisionof the TNM stage groupings in the forthcoming (seventh) edition of the TNMClassification of malignant tumours. J Thorac Oncol 2007;2:706–14.