Induction Chemotherapy for Resectable Non–Small-Cell Lung Cancer

ABSTRACT: Lung cancer remains the leading cause of cancer death in Americanmen and women. Non–small-cell lung cancer (NSCLC) accountsfor 85% of these cases. Although surgery is the best curative approachfor resectable NSCLC, long-term survival for patients with operabledisease remains poor. More than half of patients who initially presentwith stage I to IIIA disease experience relapse of metastatic disease.Postoperative adjuvant therapy has been evaluated in several randomizedtrials, and provides a survival benefit. It appears reasonable tolook to induction chemotherapy, or preoperative chemotherapy, to providea similar improvement in survival with early treatment ofmicrometastatic disease. Multiple trials of induction therapy have beencarried out with encouraging results. The use of various induction regimenswith chemotherapy alone or chemotherapy combined with radiotherapyfor stage IIIA NSCLC is under investigation. Randomized trialsare under way to better define the role of induction therapy in themultimodality treatment of NSCLC.An estimated 173,000 men andwomen in the United Stateswill develop lung cancer in2004, of whom 160,000 will die.[1]Non-small-cell lung cancer (NSCLC)accounts for approximately 85% ofthese cases. Despite advances in therapyand attempts at prevention throughsmoking cessation programs, limitedadvertising, and restricted publicsmoking, lung cancer continues to bethe leading cause of cancer death inthe United States.Unfortunately, the majority of patientswith NSCLC will present withadvanced, incurable disease. At presentation,approximately 30% of patientswill have stage I or II diseasethat is potentially curable by surgery,and an additional 25% to 30% of patientswill have locally advanced disease(stage III) that is generally treatedwith combined-modality therapy.[2]Treatment BackgroundStandard Surgical Treatment
For patients with early-stageNSCLC (clinical stage I/II), lobectomyor pneumonectomy with mediastinalsampling or lymphadenectomyis the optimal treatment modality forthe potential cure of patients and mayresult in a 5-year survival ranging from67% in patients with pathologic T1,N0 disease to 38% in patients withpathologic T3, N0 disease.[3] Clinicalstaging often understages NSCLC,particularly if newer modalities suchas positron-emission tomography(PET) scanning are not utilized. As aresult, patients with disease at a givenclinical stage fare much worse thanthose at the corresponding pathologicstages (Table 1).[2]The majority of patients with stage IIINSCLC evaluated for resection havea disease extent that precludes immediatecurative surgery. However, evenin patients who are immediately resectable,those with pathologic T1-3,N2 disease (stage IIIA) have a 5-yearsurvival of only 23%.[3] Moreover,when bulky mediastinal involvementis noted on chest radiography, 3-yearsurvival is dismal-less than 10%.[4]Although in all of these patients diseaseis confined to the chest at onsetand can be removed surgically, thevast majority suffer from recurrentNSCLC and ultimately die of metastaticdisease.

Adjuvant Approaches
In order to improve on the disappointingresults of surgery alone,investigators have focused on chemotherapyand radiotherapy in additionto surgery. The majority of historicalclinical trials examining adjuvant (ie,postoperative) therapy have not foundunequivocal, reproducible evidence ofefficacy for adjuvant chemotherapy.[5-7] However, a meta-analysis of14 trials including over 4,000 patientsand comparing postoperative chemotherapywith no further therapy demonstrateda survival advantage forchemotherapy regimens containingcisplatin.[8] More recently, a largerandomized study showed that cisplatin-based adjuvant chemotherapyimproves survival by 4% at 5 years.[9]More recently, two trials in early-stagehomogeneous populations demonstratedsignificant survival benefits inpatients treated with platin-based adjuvantchemotherapy. [10, 11]Induction Therapy
Given this uncertainty, chemotherapyprior to surgery, or induction ther-apy,has been the subject of ongoinginvestigations. Induction therapy hasbeen attractive to clinicians becauseregression of the primary tumor to chemotherapyserves as a surrogate markerfor the control of clinically undetectablemetastases. In fact, consistentlyhigher response rates and better tolerabilityhave been noted in locally advanceddisease than in metastaticNSCLC.[12-14] Notably, a completepathologic response rate of 13% hasbeen reported in early clinical studiesof induction chemotherapy.[15]Thus, chemotherapy given beforesurgery has the potential to allow forassessment of tumor response to chemotherapy,to treat micrometastaticdisease early, to facilitate the completeeradication of lesions, and toprovide better dose delivery with fewertreatment interruptions. Additionally,decreasing tumor size mayimprove resectability and decrease themorbidity of resection.Resectable NSCLCPatients with stage I/II NSCLC cangenerally undergo immediate surgerywith curative intent. Although patientswith stage III disease have no evidenceof distant metastases and areviewed as potentially curable, theirprognosis is generally poor, and theoverwhelming majority die of theirdisease. Stage IIIA encompasses patientsfor whom extended surgery canremove all known sites of disease.Patients with stage IIIB disease arefrequently considered unresectable,because of contralateral mediastinallymph node involvement (N3) or extensivetumor invasion (T4) that requiresspecialized surgical techniquesto accomplish complete resections. Acarefully selected subset of patientswith T4 disease may be suitable forsurgical resection.The basic division between stage IIIAand IIIB NSCLC was established inthe 1980s.[3] However, in the 1997revision of the International StagingSystem, patients with T3, N0 diseasewere reclassified as having stage IIBbecause their survival was almostidentical to patients with T2, N1 tumors.This grouping complicates interpretationof stage III trials before1997, because treatment groups withT3, N0 patients have improved outcomes.Futhermore, within stage IIIAdisease there is significant variabilitythat can have an impact on prognosis.Larger tumors and multiple involvedlymph node stations result in worsesurvival.[16-19]Staging NSCLCComputed Tomography
Patients with potentially resectableNSCLC, especially those with stage IIIdisease, benefit from rigorous stagingevaluation prior to therapy. Computedtomography (CT) scanning is criticalin evaluating the location andextent of the primary tumor. CT scans,however, may under- or overestimatemediastinal lymph node involvement.Normal-sized lymph nodes (generallyconsidered to be < 1 cm in the shortaxis) can contain tumor cells, and enlargedlymph nodes do not alwayscontain metastatic tumor. Lymph nodemetastases become more likely withincreasing nodal size.[20]In one meta-analysis of 29 studiesinvolving 2,226 patients utilizing CTscans for the detection of mediastinalnodal metastases, the mean sensitivityand specificity were 60% and 77%,with a positive predictive value of50% and a negative predictive valueof 85%. An enlarged mediastinallymph node detected by CT scanning,therefore, only contains cancer 50%of the time, and 15% of normal-sized(< 1 cm) lymph nodes contain tumorcells.[21]Positron-Emission Tomography
PET with 18F-fluorodeoxyglucoseis a newer staging technique that usesthe relatively increased metabolic activityof cancer cells to differentiatethem from normal cells. In the aforementionedmeta-analysis, data from14 studies involving 514 patients wereexamined, and the authors found thatthe mean sensitivity and specificityfor PET were 79% and 91%, respectively,with a positive predictive valueof 90% and a negative predictivevalue of 93%.[21]Pieterman et al reported their experiencein 102 patients with resectableNSCLC with preoperative PETscanning and found sensitivity andspecificity of 91% and 86%, respectively.[22] However, in this group ofpatients, the positive predictive valueof PET was 74%. Therefore, 26% ofpatients with increased uptake on PEThave no evidence of nodal disease onhistopathology. Furthermore, evenlymph nodes that are positive by CTand PET require histopathologic proofof involvement. The negative predictivevalue of both a negative PET andCT scan of 97% was excellent in thisseries, and many surgeons feel thatPET scans can reliably rule out N2disease but that surgical staging is necessaryto confirm its presence.Mediastinoscopy
In many centers, mediastinoscopyis routinely performed to stage patientswith lung cancer who arecandidates for surgical resection. Mediastinoscopyis used to detect N3disease for which surgical therapywould not be advised and to detectN2 disease for which induction therapyprior to definitive surgery wouldbe warranted. Because not all lymphnodes can be sampled, cervical mediastinoscopyprovides sensitivity between72% and 89%.[23-25] Cervicalmediastinoscopy allows access to theright and left paratracheal lymphnodes and subcarinal lymph nodes.The aorticopulmonary window andsubaortic lymph nodes are not accessiblevia cervical mediastinoscopy andrequire anterior mediastinotomy.Combined Modalities
Appropriate staging procedures forpatients with resectable NSCLC,therefore, include a CT scan not onlyto assess the primary tumor and mediastinumbut also to evaluate for evidenceof metastatic disease withinthe lung, liver, and adrenal glands.All patients with mediastinal lymphnodes that are greater than 1 cm byCT imaging or are positive by PETscan should have mediastinal diseaseconfirmed histopathologically. In patientswith normal mediastinal lymphnodes by CT scan, a negative PETscan may obviate the need for mediastinoscopy.Mediastinoscopy, however,may identify patients withmicrometastatic lymph node involvementnot apparent on imaging modalitieswho may benefit from inductiontherapy.Detecting Brain Metastases
The incidence of brain metastasesin the initial staging of patients withprimary lung cancer has been reportedto between 12% and 18%.[26,27]Asymptomatic brain metastases occurmore frequently in patients withmore advanced stages of disease, withrates as high as 30% at 2 years instage II/III patients. Higher stage andnonsquamous histology have beenidentified as risk factors for brain metastases.[28,29] In asymptomatic patients,gadolinium-enhanced magneticresonance imaging (MRI) of the brainis superior for the detection of occultbrain metastases.A recent study randomized 332patients with potentially operableNSCLC, but without neurologicsymptoms, to brain CT or MRI inorder to detect occult brain metastasisbefore lung surgery.[30] MRI showeda trend toward a higher preoperativedetection rate than CT (P = .069),with an overall detection rate of approximately7% from pretreatment to12 months after surgery. In patientswith stage I/II disease, the detectionrate was 4% (8 of 200), whereas forindividuals with stage III disease, therate was 11.4% (15 of 132).Whether the improved detectionrate of brain metastases by MRI translatesinto improved outcome remainsunknown. Brain imaging is currentlyrecommended as a part of the initialevaluation for all patients with potentiallycurable locally advancedNSCLC. Finally, reevaluation of diseaseafter induction chemotherapy andprior to thoracotomy is recommendedto exclude disease progression. Theability of noninvasive tests to predictpathologic response has thus far beensuboptimal.Induction Chemotherapy TrialsInitial Observations
Several clinical trials have testedcombination chemotherapy prior tosurgery to improve the outcome of patientswith resectable NSCLC. Initialstudies using older chemotherapy regimenssuggested that preoperative chemotherapymay be beneficial.[31,32]Martini demonstrated that otherwiseresectable patients with ipsilateral mediastinallymphadenopathy as theirsole site of distant spread can have3-year survivals of 43%, and 5-yearsurvivals of 24%, if both the primarytumor and ipsilateral mediastinalnodes were completely resected andfollowed by mediastinal irradiation.[33,34] These same studies revealedthat individuals with bulkyipsilateral mediastinal lymphadenopathyhad only an 8% 3-year survival.Based on these observations, a preoperativecombination chemotherapyprogram with MVP (mitomycin [Mutamycin],vinca alkaloids, and highdosecisplatin [Platinol] at 120 mg/m2)was developed at Memorial Sloan-Kettering Cancer Center for use instage IIIA patients with clinical N2disease.[15] In a group of 136 patients,the objective major responserate to MVP was 77%, with a 10%complete response rate. Overall, 65%of patients underwent complete resections;14% had achieved a pathologiccomplete response at surgery.The median survival was 19 monthsfor all patients. The 3-year survivalfor completely resected patients was41%-a significant improvement overthe prior surgery-only experience,where the 3-year survival for thisgroup was 8% (P = .001).[15] Particularlynoteworthy was that a pathologiccomplete response was observedin approximately 12% of advancedNSCLC patients who received preoperativeMVP chemotherapy, with survivalestimates in this population of54% at 5 years.[35]The Cancer and Leukemia GroupB (CALBG) conducted a second similarlydesigned trial. A total of 74 patientswith surgically staged IIIA (N2)NSCLC were treated with two cyclesof preoperative cisplatin and vinblastine.[36] Patients who underwent resectionwere then treated with twocycles of adjuvant cisplatin and vinblastinefollowed by thoracic radiation.Approximately 64% of patientsachieved a radiographic response orstable disease, and 62% underwent acomplete resection. Operative mortalitywas 3% (2 deaths). Median survivalwas 15 months.Elias and investigators at the Dana-Farber Cancer Institute treated 34 patientswith pathologically confirmedN2 disease, using infusional cisplatin,fluorouracil (5-FU), and leucovorinplus postoperative radiotherapy to 54-60 Gy.[37] The radiographic responserate to this regimen was 65%. Thoracotomywas performed in 82% of patients.No operative mortality wasnoted, and 18% of patients achieved apathologic complete response. Mediansurvival was 18 months. The authorsnoted that few local recurrencesoccurred and that 15% of first relapseswere solely in the brain.Randomized Studies
Following these initial observationsthat suggested promising results withinduction therapy, a number of randomizedstudies aimed at establishingthe role of induction therapy wereconducted.

• Spanish Study-The first randomizedtrial that compared cisplatinbasedchemotherapy plus surgery withsurgery alone was conducted by Rosellet al in patients with stage IIIA disease.[38] Pathologic confirmation ofN2 was not mandatory, and approximately27% of patients had clinicalT3, N0/1 disease. Chemotherapy consistedof the MIP regimen (mitomycin,ifosfamide [Ifex], cisplatin) givenat 3-week intervals before surgery forthree cycles. All patients receivedmediastinal irradiation (50 Gy) aftersurgery. A total of 60 patients wererandomized, 30 in each arm.After 2 years, the trial was prematurelyterminated because of thesignificant benefits seen in the chemotherapyarm. No toxic deaths occurredduring preoperative therapy.The partial response rate was 53%,with a 7% clinical complete responserate. Postoperative mortality was similarin both groups (two deaths each).Overall median survival was 26 monthsin the chemotherapy arm vs 8 monthsin the surgery arm (P < .001).Survival rates at 2 years were 27%and 0%. Survival in the control armof surgery alone was much lower thanexpected from historic controls. Ananalysis performed 7 years later confirmedthe survival advantage in favorof preoperative chemotherapy,with median survival of 22 monthsin the chemotherapy arm vs 10months in the surgery-alone arm(P = .005).[39] The survival rate at 5years was 17% and 0%, respectively.
• M. D. Anderson Study-Roth andinvestigators at M. D. AndersonCancer Center conducted another randomizedstudy of preoperative chemotherapyin stage IIIA NSCLC.[40]Chemotherapy consisted of three cyclesof cyclophosphamide (Cytoxan,Neosar), etoposide, and cisplatin. Patientswho responded to preoperativechemotherapy were given an additionalthree cycles of adjuvant therapyafter complete resection. Patients whowere found to have unresectable diseasewith incomplete resection couldundergo postoperative radiationtherapy.A total of 28 patients were randomizedto receive preoperative chemotherapy,and 32 were randomizedto receive primary surgery. Almostone-quarter of patients had T3, N0/1disease. The overall response rate was35%, including one pathologiccomplete response. Disease progressiondeveloped in four patients duringchemotherapy; however, theresectability rate and the rate of completeresponses were the same in botharms.An interim analysis showed a statisticallysignificant survival advantagefor the preoperative chemotherapy arm(P < .008). The trial was therefore haltedafter 60 patients were enrolled. Theestimated 2-year survival rate in thepreoperative arm was 60% and in thesurgery arm was 25%. An analysis4 years later revealed that mediansurvival in the chemotherapy arm was21 months and in the surgery arm was14 months (P = .056).[41]
• NCI Study-A third randomizedtrial in histologically confirmed N2NSCLC was conducted by Pass et alat the National Cancer Institute(NCI).[42] Chemotherapy consistedof cisplatin and etoposide. At interimanalysis, 27 patients had been randomized.The response rate was 61%.Preliminary results suggested a trendtoward increased survival for the preoperativechemotherapy arm (mediansurvival = 29 vs 16 months, P = .095).
• Japanese Study-A fourth randomizedstudy, conducted in Japan,compared induction chemotherapywith cisplatin and vindesine (Eldisine)vs surgery alone in patients withstage IIIA (N2) NSCLC. After 62 patientswere enrolled, this study wasterminated due to slow accrual. Therewere no statistically significant differencesin survival in this trial (medianoverall survival was 17 monthsfor patients receiving induction and16 months for patients undergoing surgeryalone).[43]

Early-Stage Disease

Taken together, the four small, randomizedstudies described above demonstrateda potential benefit with theaddition of induction chemotherapy instage IIIA NSCLC. Studies have sincebeen designed to evaluate the role ofinduction chemotherapy before surgeryin even earlier-stage NSCLC.

• French Study-The French CooperativeOncology Group conducteda large randomized study in stage I(except T1, N0), II, and IIIA resectableNSCLC.[44,45] The 355 eligiblepatients were randomized to inductiontherapy with two cycles of mitomycin,ifosfamide, and cisplatin or toprimary surgery. Patients who respondedto chemotherapy received anadditional two cycles postoperatively.In both arms, patients with pT3 orpN2 disease or incomplete surgeryreceived radiotherapy after either surgeryor completion of chemotherapy.This regimen yielded a response rateof 64% with a pathologic completeresponse rate of 11%.The complete resection rate wassimilar between the two arms: 92%with induction vs 86% with surgeryalone. A 10-month improvement inmedian survival in the chemotherapyarm was documented (26 vs 36months), although this was not statisticallysignificant (P = .11). Operativemortality was similar in both arms, butthere was a trend toward increased postoperativemortality (7% vs 4.5%) andincreased postoperative complicationsin the chemotherapy arm. Although astatistically significant improvement insurvival could not be demonstrated,subset analysis showed that patientswith stage N0/1 disease derived a survivalbenefit from induction chemotherapy(hazard ratio = 0.68, 95% CI =0.49-0.96, P = .027). This survival benefitwas not observed in patients withN2 disease (hazard ratio = 1.04, 95%CI = 0.68-1.60, P = 0.85).The study by Depierre et al is thelargest trial of induction chemotherapyconducted to date. However, it remainsnonconclusive as it included aheterogeneous group of patients withvarious stages of disease and its implicationsfor the treatment of patientswith stage IIIA NSCLC are only basedupon subset analysis.
• US Studies-The role of inductionchemotherapy for early-stageNSCLC has also been under investigationin the United States. Pisters etal reported promising results from themulticenter phase II study, the BimodalityLung Oncology Team (BLOT)trial.[46] Patients with stage IB-IIIA(T3, N1, M0) NSCLC received inductiontherapy with two cycles ofcarboplatin (Paraplatin) and paclitaxelfollowed by three additional cyclespostoperatively. Evaluable patientsincluded 94 evaluated for survival andtoxicity and 90 for response. The responserate was 56%, the pathologiccomplete response rate was 6%, and3% of patients had disease progressionduring induction. The 1-year survivalrate was 85%.Results from a second group ofpatients who received three rather thantwo cycles of induction chemotherapywith carboplatin and paclitaxel havealso been reported.[47] In a total of 124patients, the response rate was 51%and 3-year survival was 61%, whichwas superior to historical controls. Thesurgical mortality rate was only 1%.The North Central Cancer TreatmentGroup (NCCTG) conducted asimilar phase II study with inductioncarboplatin and paclitaxel in patientswith stage T1-3, N0/1, M0 disease.[48]Of 52 patients, 3 died postoperatively.The 2-year survival rate was 73%.An ongoing phase III randomizedstudy (S9900) is randomizing patientsto three cycles of induction carboplatinand paclitaxel followed by twoadditional cycles of postoperative carboplatinand paclitaxel or surgeryalone. Over 325 patients have thus farbeen entered into this trial, which hasan accrual goal of 600.

Newer Chemotherapy Regimens

Despite the often impressive resultswith induction chemotherapy,one-quarter or more of patients do notrespond to current primary chemotherapyregimens. Also, the majorityof patients who initially respond andhave complete surgical resectionseventually relapse. Two-thirds of therelapses are systemic. It appearsunreasonable to look to further advancesin surgery or radiation to substantiallyimprove survival. The abilityof a locoregional therapy modalitysuch as radiotherapy to control systemicrelapse is inherently limited.Based on the accumulated resultsand relapse patterns, improvementsin primary chemotherapy regimens aremost likely to result in improvementsin overall outcome. However, whileresponse rates to chemotherapy arehigh in stage III patients, clinical andpathologic complete response rates aregenerally less than 20%. Because individualswith complete pathologicresponses have the highest resectionand survival rates, the completepathologic response rate can serve asa surrogate for patient outcome inthe development of novel inductionregimens.

• Novel Agents-Platinum-basedcombinations with newer agents, suchas gemcitabine (Gemzar) and docetaxelhave attracted the interest ofinvestigators. The European Organizationfor Research and Treatment ofCancer (EORTC) group conducted aphase II trial of cisplatin and gemcitabine(Gemzar) in 47 patients withstage IIIA NSCLC and showed a 70%response rate with mediastinal nodeclearing in 53% of cases. The estimated1-year survival was 69%.[49]

• Another phase II study, by Betticheret al, investigated the efficacyand toxicity of induction docetaxel(Taxotere) and cisplatin in stage IIIANSCLC.[50] A total of 90 patientswere treated in this trial. The responserate in this trial was 66%, and 75patients were able to undergo resection.Pathologic complete responsesoccurred in 19% of patients who underwentresection, and mediastinalnode clearing was strongly predictiveof long-term survival (P = .0003).Another multi-institutional studyrandomized 274 patients with stageIIIA/IIIB NSCLC to receive up tothree cycles of single-agent docetaxel(100 mg/m²) prior to definitive surgeryor radiation.[51] Overall survivalin the docetaxel group was 14.8 monthsand in the control group was 12.6months. Although not statistically significant,there was a trend toward longersurvival in the docetaxel arm.
• Triplet Regimens-Triplet combinationshave also been investigated. Ina phase I study conducted at MemorialSloan-Kettering Cancer Center, thecombination of gemcitabine, cisplatin,and docetaxel was evaluated beforesurgery.[52] The recommended dosefor phase II studies was gemictabineat 1,000 mg/m² on days 1,8, and 15;docetaxel at 100 mg/m² on day 1; andcisplatin, 100 mg/m² on day 15. Theresponse rate in the phase I trial was44%, and the phase II trial has not yetbeen reported.The Spanish Lung Cancer Groupreported a similar triplet combination.In this trial, 120 patients with stageIIIA, N2 and T4, N0/1 NSCLC weretreated with gemcitabine, cisplatin,and docetaxel. All patients receivedthree cycles of therapy prior to surgery.Nodal clearing occurred at resectionin 37% of patients with N2disease at diagnosis. When reported,the median survival for 84 evaluablepatients was 14.6 months.[53]A third study from Italy treated 49patients with pathologically provenstage IIIA (N2) disease with gemcitabine,paclitaxel, and cisplatin priorto surgery or definitive radiotherapy.[54] Approximately 74% of patientsachieved an objective response,and 60% of patients underwent thoracotomy.Mediastinal nodes were freeof tumor in 35% of all cases, andeight pathologic complete responses(16%) were reported. Median survivalwas 23 months.Multiple ongoing phase III randomizedstudies are evaluating the roleof induction chemotherapy basedupon newer chemotherapy regimensin early-stage NSCLC (Table 2.)

Induction Chemoradiotherapy

It is postulated that chemotherapydecreases the distant failure rate ofsurgery by eradication of micrometastaticdisease, whereas radiation canreduce the bulk of the primary tumorand nodal disease, thereby increasingthe rate of complete pathologic response-arguably the most importantpredictor of long-term survival. Therefore,investigators have attempted tointegrate chemotherapy and radiationtherapy into induction therapy for resectabledisease.

Phase II Trials

• CALGB Study-The Cancer andLeukemia Group B reported a phase IIstudy of two cycles of induction cisplatin,vinblastine, and infusional5-FU with concomitant radiation to30 Gy, surgery, then an additionalcycle of chemotherapy and an additional30 Gy radiation postoperativelyin patients with stage IIIA NSCLC.[55]A total of 41 patients were studied(80% with N2 disease). The responserate to induction therapy was 51%.Of the 31 patients who went to surgery,only 25 were resected. Thepathologic complete response ratewas 16%. Toxicity was substantial,with six (15%) treatment-relateddeaths. The 1-year survival was 58%,with a median survival of 15.5 months.
• SWOG 8805-A Southwest OncologyGroup phase II trial (SWOG8805) examining trimodality therapyincluded patients with stage IIIA andIIIB disease.[56] The 75 patients withstage IIIA disease and 51 patientswith stage IIIB disease were treatedwith two cycles of cisplatin and etoposideplus 45 Gy irradiation over5 weeks. Resectable disease was confirmedin 85% of stage IIIA patientsand 80% of stage IIIB patients. Patientswho were not resectable receivedan additional two cycles ofchemotherapy and a 14.4-Gy radiationboost.The pathologic complete responsein this trial was 21%. Treatmentrelatedmortality was 10%. The 6-yearsurvival was 20% for patients withstage IIIA disease and 22% for patientswith stage IIIB disease. Subsetanalysis revealed that patients withT4, N0/1 disease had a particularlyhigh long-term survival rate of 49%,whereas patients with N3 diseasebased on contralateral mediastinallymphadenopathy had a dismal 2-yearsurvival rate of 0%.[57]
• Hyperfractionated RadiotherapyTrials-A number of phase II studiesthat used hyperfractionated radiotherapyschedules along with concurrentchemotherapy as induction have beenreported. Choi et al studied preoperativetwice-daily radiation therapy withtwo cycles of concurrent cisplatin, vinblastine,and 5-FU in stage IIIA (N2)and IIIB patients.[58] They reportedan 84% complete resection rate, and a5-year survival rate of 37%.Eberhardt et al conducted a studyin which 94 stage IIIA/IIIB patientsreceived four cycles of preoperativecisplatin and etoposide, three beforeradiation therapy, and the fourth combinedwith twice-daily radiation (45 Gy,1.5 Gy per fraction, twice daily for3 weeks).[59] A complete resectionwas performed in 53% of patients(60% in stage IIIA and 45% instage IIIB groups). The median survivaland 4-year survival rates were20 months and 31% for stage IIIApatients, and 18 months and 26% forstage IIIB patients.Thomas et al examined preoperativechemotherapy (two cycles of ifosfamide,carboplatin, and etoposide)with subsequent chemoradiotherapy(carboplatin and vindesine plus concurrent45 Gy irradiation) and thensurgery in patients with stage IIIA/IIIBdisease.[60] Median survival (25months in patients with stage IIIAdisease and 17 months in those withstage IIIB disease) compared favorablywith other studies.

Phase III Trial

Based on the promising results ofthe SWOG phase II trial, a phase III,randomized trial of the North AmericanIntergroup (0139) was initiated.The goal of the study was to evaluatethe need for surgical resection afterchemoradiotherapy. Patients withstage IIIA pathologically proven N2NSCLC were randomized to fullcoursechemotherapy with cisplatinand etoposide plus radiotherapy to61 Gy, or to induction chemoradiationwith the same regimen to 45 Gyfollowed by surgical resection. Dueto slow accrual, the study was closedafter 429 patients (411 of whom wereeligible) were enrolled.[61]In a preliminary analysis of 392patients after a median follow-up of69 months, progression-free survivalwas significantly improved in the surgicalvs the nonsurgical arm (29% vs19%,

P

= .02), but overall survivalwas similar between the two arms(38% vs 34%). Although more earlynoncancer deaths occurred in the surgicalarm-primarily postoperatively(12 of 14 deaths occurred in patientsundergoing pneumonectomy)-thesurvival curves crossed at the medianof 22 months and showed an absolutebenefit of approximately 5% for surgeryat 3 years. Importantly, patientswho achieved a pathologic completeresponse in the mediastinal lymphnodes achieved 50% 3-year overallsurvival.The results of this study challengethe role of surgery in traditionally resectablestage IIIA NSCLC. Concurrentchemoradiotherapy alone maybe sufficient therapy for resectablestage IIIA NSCLC. It is likely that theimpact of postoperative mortality observedin the surgical arm obscured apotential survival advantage resultingfrom surgical resection. Final resultsof this Intergroup study are awaited.

Other Small Trials

The addition of radiotherapy to preoperativeprograms deserves furtherevaluation. Two small studies presentedin abstract form only have comparedinduction chemotherapy tochemoradiotherapy in potentially resectablestage III NSCLC. A studyfrom Brazil[62] demonstrated a significantimprovement in resectionrates, freedom from progression, and2-year survival in patients receivinginduction chemoradiotherapy vs chemotherapyalone before surgery. Thelong-term results of this study havenot been published; it therefore remainsto be seen whether these datawill hold up over time.A study from France in 92 patientswith stage IIIA/IIIB NSCLC evaluatedinduction MVP chemotherapyalone vs induction MVP followed by5-FU and cisplatin plus split-courseradiation to 40 Gy.[63] Higher resectionrates were noted in the inductionchemoradiotherapy arm, but this didnot translate into improved survivalover induction chemotherapy alone.Surgical complications were increasedin the chemoradiotherapy arm.The use of both induction chemotherapywith carboplatin and gemcitabinefollowed by concurrentchemoradiotherapy and then surgeryin patients with stage III NSCLC hasalso been investigated.[64] This studyenrolled 39 patients, 19 of whom ultimatelyunderwent surgical resection(16 with stage IIIA disease and 3 withselected T4 tumors). The pathologiccomplete response rate was 16% (42%in the mediastinal lymph nodes), andthe 3-year overall survival rate in surgicallytreated patients was 51%. Furtherinvestigation of this approach iswarranted.

Importance of Patient Selection

A potential advantage of inductionchemoradiotherapy may be that ahigher proportion of pathologic completeresponses is achieved comparedto the use of induction chemotherapyalone. However, postoperative morbidityand mortality appears to besomewhat higher with inductionchemoradiotherapy compared to inductionchemotherapy. Thus, ifsurgery is to be performed afterchemoradiotherapy, careful patient selectionis paramount. Patients withcompromised performance status orborderline cardiopulmonary reserveare unsuitable for this approach. Futuretrials need to address whether theaddition of radiotherapy to inductionchemotherapy translates to a survivaladvantage in patients with resectableNSCLC.

Surgical ConsiderationsAfter Induction Therapy

Although multiple studies havedemonstrated the safety of surgicalresection after induction therapy, severalissues make interpretation of thedata difficult. First, many studies involveheterogeneous populations thatinclude patients with both stage IIIAtumors with nodal involvement andstage IIIB tumors that are locally advanced.Resection of peripheral tumorswith nodal disease is associatedwith significantly lower complicationrates than extended resections for locallyadvanced disease.Additionally, patients who ultimatelyundergo surgical resection area selected group often without comparablecontrols. They are fit enoughfor surgical resection and have tumorswith demonstrated sensitivity toinduction chemotherapy. These patientshave better nutritional status,immunocompetency, and overallfunction, and therefore, an overall betterprognosis than those unable to toleratechemotherapy or who may not becandidates for multimodality therapy.Finally, many studies are retrospective,involve few subjects, and includevariable induction regimens such aschemotherapy alone and inductionchemoradiotherapy. Despite the limitationsof the available data, some conclusionsabout the risks and benefitsof surgical resection after inductiontherapy can be made.

Tumor and Host Responses

Induction therapy causes tumor andhost responses that have both advantagesand disadvantages in terms ofsurgical resection. Large, bulky tumorsmay be downsized, permittingeasier intraoperative visualization andmanipulation. Induction therapy mayalso result in tumor shrinkage awayfrom vital structures making significantlymore radical resection unnecessary.However, the inflammatoryresponse and fibrosis resulting frominduction therapy can make resectionmore difficult.This is particularly true when tryingto determine whether "close" marginsare clear of tumor. Becauseintraoperative frozen sections showfibrosis around treated tumor, the trueextent of the original tumor cannotbe established. Resection of all fibrosismay require a substantially moreradical operation, but incomplete resectionrisks leaving microscopic depositsof tumor. This issue is nottrivial, as patients with incompletelyresected tumor invariably have poorprognoses. Tumor fibrosis also obscuresthe normal anatomic planes,making dissection around vessels, airways,and the esophagus more difficult.Repeat mediastinoscopy afterprevious mediastinoscopy and inductiontherapy is generally considered arisky procedure and is frequently inadequate.[65]Physical conditioning and immunefunction are weakened by inductiontherapy such that patient tolerance ofperioperative complications may bereduced. Induction therapy likely impairslymphatic drainage of the lungand pleural space and causes parenchymalchanges that alter lung elasticrecoil and healing. These factors maycontribute to postoperative air leaks andpleural space problems. Most studieshave focused only on major complicationrates and mortality. Important althoughless significant difficulties withpostoperative recovery after inductiontherapy may be underreported.

Surgical Mortalityand Complications

Surgical mortality for most resectionsappears comparable to that forresections without induction therapy.However, pneumonectomy, particularlyright-sided, is associated withincreased risk of mortality. A reportfrom Memorial Sloan-Kettering CancerCenter by Martin et al examined470 patients with stage I-IV NSCLCreceiving a variety of regimens, butprimarily MVP. These investigatorsshowed a high rate of surgical mortalityafter right pneumonectomy (24%)compared to an overall rate of surgicalmortality of 4% after inductionchemotherapy.[66] Preoperative chemotherapywas otherwise not foundto be a risk factor for morbidity ormortality in multivariate analysis.Most studies report complicationrates of 15% to 40% in patients undergoingpostinduction therapy resections.With the previously mentionedcaveats, authors report rates equalto[66-70] or above[71-73] those forstandard resections. Most complicationsare pulmonary, such as bronchopleuralfistula, pneumonia,postresectional pulmonary edema,adult respiratory distress syndrome(ARDS), and atelectasis. A retrospectiveanalysis from Vanderbilt Universityreported an increase in morbidityand mortality after induction chemotherapy.[74] This single-institutionexperience from 1997 to 1999 included34 patients with early-stage disease(IB-IIIA) receiving carboplatinand paclitaxel for three cycles comparedwith 67 patients undergoing resectionwithout prior systemic therapy.They observed significant increasesin life-threatening complications (27%vs 3%), major complications (47% vs19%), and reintubation (18% vs 3%)in patients receiving chemotherapy.No hospital mortality was observed.A Spanish case-control study examinedpostoperative morbidity andmortality in 42 patients receiving inductionchemotherapy (74 % clinicalstage III or IV) compared with 42matched controls (all clinical stage I)undergoing surgery alone.[67] Nopostoperative mortality was recorded.Although the authors concludedthat there was no significant influenceof chemotherapy on postoperativecomplication rates (26% in thechemotherapy arm vs 43% in controls),the complication rate in the controlswas unusually high.Depierre et al reported a trendtoward increased mortality and morbiditywhen surgery followed a cisplatin-based induction regimen,compared to a surgery-alone arm.[45]In a large multicenter study, Albain etal observed a surgical mortality rateof 5% in patients who underwentthoracotomy postchemoradiotherapy.[61] Most deaths were due to pulmonarycomplications in patientsundergoing pneumonectomy. Inmulticenter phase II studies with inductioncarboplatin and paclitaxel inearly-stage NSCLC, postoperativemortality has been acceptable, ie, 1%to 6%.[46,48]Extended resections that includebronchoplastic or tracheoplastic reconstructionsprobably have highercomplication rates in the setting ofinduction therapy. This is most likelyto be the case when radiotherapy isinvolved. Macchiarini and colleaguescompared two groups of patients withT4 disease.[71] Induction chemotherapyalone was administered to 11patients, and 12 patients received combinedchemoradiotherapy with 40 Gyto the lesion, mediastinum, and ipsilateralsupraclavicular fossa. Majorcomplication rates were 42% in thechemoradiation group vs 9% in thechemotherapy-alone group. Similarly,bronchial stump breakdown maybe more common in the setting ofradiotherapy.[75]In conclusion, it appears that inductiontherapy, especially chemoradiotherapy,may slightly increasesurgical complications, with a minimal,perhaps negligible, increase insurgical mortality. This potential detrimentaleffect is likely to be outweighedby the potential survivalbenefit from the use of induction therapy.A number of ongoing randomizedcomparisons with a surgicalcontrol arm is expected to provideadditional data on the safety of contemporaryinduction chemotherapyregimens (Table 2).

Incorporation of Novel Agents

Recent years in lung cancer researchhave been an exciting time,with the identification of targetedagents such as gefitinib (Iressa) anderlotinib (Tarceva), which inhibit theepithelial growth factor receptor(EFGR) tyrosine kinase, and bevacizumab(Avastin), a monoclonal antibodyagainst the vascular endothelialgrowth factor (VEGF). Some of thesedrugs, eg, gefinitib and erlotinib, havebeen combined with cytotoxic chemotherapyin the advanced diseasesetting without significantly improvingoutcome.Investigators, however, are look-ing at the induction setting in an effortto understand why a fraction ofpatients does benefit from theseagents. For example, the MemorialSloan-Kettering Cancer Center groupis currently conducting one such trialof gemcitabine and cisplatin plus intermittenterlotinib as induction therapyin patients with locally advanceddisease. In these studies, the availabilityof posttreatment tissue mayhelp further elucidate mechanisms ofaction of novel agents.Another trial, reported by Altorkiet al, included patients with stage IBto IIIA NSCLC.[76] A total of 29patients were treated with the combinationof paclitaxel, carboplatin, andcelecoxib (Celebrex). The addition ofcelecoxib to chemotherapy normalizedlevels of PGE

₂

found in NSCLCpatients after treatment with chemotherapy.The authors felt that the additionof celecoxib enhanced theresponse to preoperative therapy.

The Optimal Therapy

There are no validated criteriaavailable that can be used to determinewhether a patient with potentiallyresectable, locally advancedNSCLC should be treated with inductionchemotherapy followed by surgicalresection, chemoradiotherapyalone, or chemoradiotherapy followedby surgery. In general, patients withnonbulky, mobile lymphadenopathytend to be offered induction chemotherapy,and patients with more extensivemediastinal disease are treatedwith chemoradiotherapy, possibly followedby surgery. When chemotherapyis combined with concurrentradiotherapy, usually only suboptimalsystemic doses of chemotherapy canbe delivered due to increased toxicity.It is possible that giving full-dosesystemic therapy treats micrometastaticdisease immediately before it hasthe opportunity to progress or becomeresistant to the lower doses of chemotherapythat are typically given withradiation.Ichinose et al attempted to identifyrisk factors for recurrence and survivalin 406 patients with resected N2NSCLC that might allow selection ofpatients who should have adjuvanttherapy.[6] In the multivariate analysis,survival was worse in patients withmultiple N2 stations, pathologic T2,T3, or N1 disease, and age older than65. Local recurrence was increasedwith multiple N2 lymph nodes andclinical N1 or N2 disease. Patientswith multiple involved mediastinallymph nodes had a 5-year survivalrate of 17% and a freedom from localrecurrence rate of 48%, compared with43% and 75%, respectively, for patientswith an isolated mediastinallymph node. Ichinose et al surmisedthat the number of mediastinal lymphnodes involved is an important prognosticfactor. It is necessary to betterselect patients who would benefit fromeach treatment approach, and it couldwell be that patients with bulkier diseasefare worse regardless of the additionof radiation to induction therapy.Similiarly, Andre et al evaluatedprognostic factors in 702 patients whounderwent surgical resection forstage IIIA disease.[14] Negative prognosticindicators were clinical N2disease, multiple lymph node involvementof the mediastinum, pathologicT3 or T4 disease, and no preoperativechemotherapy.The role of surgery in the multimodalitytherapy of resectable stage IIINSCLC has been questioned by someclinical studies. In fact, three smallphase III trials with nonsurgical armsfailed to show benefit for surgery overradiation.[77-79] More recently, asdiscussed previously, the results ofIntergroup trial 0139 seem to fuel thedebate. Many have questioned the roleof radiotherapy in the therapy regimen.It may be that the early noncancerdeaths may have been a functionof the radiation, not the chemotherapy.Emerging data on adjuvant therapyfor NSCLC has certainly changedthe practice of oncology. The InternationalAdjuvant Lung Cancer Trial(IALT) Collaborative Group concludedthat cisplatin-based adjuvant therapyimproved survival in patients withcompletely resected NSCLC.[9] Inthis study of 1,867 patients who wererandomly assigned to adjuvant therapyor to observation, overall survivalwas significantly higher in the adjuvantchemotherapy group (

P

< .03),with an absolute increase of 4% in the5-year survival rate. Similarly, both theNational Cancer Institute of Canadatrial JBR.10[10] and the Cancer andLeukemia Group B (CALGB)[11]study 9633 presented at the AmericanSociety of Clinical Oncology meetingin 2004 conducted in early-stage homogeneouspopulations demonstrateabsolute survival benefits of 15% at 5years (

P

= .0022) and 12% at 4 years(

P

= .028) respectively.No randomized study has yetreported results of a comparison betweeninduction and adjuvant chemotherapyfor resectable NSCLC, butmost believe that at least a comparablesurvival benefit will be demonstratedwith the use of inductiontherapy. The ongoing study enrollingpatients with stage IB to IIIA NSCLC,orchestrated by Rosell et al, iscomparing induction chemotherapy,adjuvant chemotherapy, and surgeryalone. Patients in both chemotherapyarms receive carboplatin andpaclitaxel.

Conclusions

Despite what appears to be completeresection of NSCLC, the majorityof patients with non-small-celllung cancer die from recurrent, metastaticdisease. Therapy aimed at eradicatingmicrometastatic disease hasbeen the goal of induction chemotherapy.It is hoped that improvementsin systemic chemotherapy with theadvent of targeted agents will translateinto a survival benefit for patientswith resectable NSCLC.Induction treatment represents animportant paradigm for clinical researchin NSCLC, allowing in vivochemosensitivity assessment and offeringa testing field for new, targetedagents. Furthermore, the delivery ofinduction chemotherapy seems moreeffective than the same therapy in theadjuvant setting.For patients with stage I and IINSCLC, the use of induction has beenshown to be feasible but is not recommendedoutside of a clinical trial. Forpatients with resectable stage III disease,induction chemotherapy-withor without radiotherapy-appears todefinitively answer lingering questionsis by the timely accrual of patientsto well designed clinical trials.be the best approach. The only way todefinitively answer lingering questionsis by the timely accrual of patientsto well designed clinical trials.

Disclosures:

The authors have nosignificant financial interest or other relationshipwith the manufacturers of any productsor providers of any service mentioned in thisarticle.

References:

1.

Jemal A, Tiwari R, Murray T, et al: Cancerstatistics, 2004. CA Cancer J Clin 54:8-29,2004.

2.

Mountain CF: Revisions in the InternationalSystem for Staging Lung Cancer. Chest111:1710-1717, 1997.

3.

Mountain CF: A new international stagingsystem for lung cancer. Chest 89:225-233,1986.

4.

Martini N, Flehinger B, Zaman M, et al:Results of resection in non-oat cell carcinomaof the lung with mediastinal lymph node metastases.Ann Surg 198:386-397, 1983.

5.

Keller SM, Adak S, Wagner H, et al: Arandomized trial of postoperative adjuvanttherapy in patients with completely resectedstage II or IIIA non-small-cell lung cancer.Eastern Cooperative Oncology Group. N EnglJ Med 343:1217-1222, 2000.

6.

Ichinose Y, Tada H, Koike T, et al: Randomizedphase III trial of postoperative adjuvantchemotherapy in patients with completelyresected stage IIIa-N2 NSCLC: Japanese ClinicalOncology Group (JCOG 9304) (abstract1241). Proc Am Soc Clin Oncol 20:311a, 2001.

7.

Scagliotti V, Fossati R, Torri V, et al: Randomizedstudy of adjuvant chemotherapy forcompeltely resected stage I, II, or IIIA nonsmall-cell lung cancer. J Natl Cancer Inst95:1422-1424, 2003.

8.

Non Small Cell Lung Cancer CollaborativeGroup: Chemotherapy in non-small celllung cancer: A meta-analysis using updateddata on individual patients from 52 randomisedclinical trials. Br Med J 311:899-909, 1995.

9.

Arriagada R, Bergman B, Dunant A, etal: Cisplatin-based adjuvant chemotherapy inpatients with completely resected non-smallcell lung cancer. N Engl J Med 350:351-360,2004.

10.

Winton TL. Livingston R, Johnson D, etal: A prospective randomized trial of adjuvantvinorelbine and cisplatin in completely resectedstage 1B and II NSCLC Intergroup JBR.10 (abstract7018). Proc Am Soc Clin Oncol 23:619,2004.

11.

Strauss GM, Herndon J, Maddaus MA,et al: Randomized clinical trial of adjuvantchemotherapy with paclitaxel and carboplatinfollowing resection in stage IB NSCLC: Reportof CALGB 9633 (abstract 7019). Proc AmSoc Clin Oncol 23:619, 2004.

12.

Donnadieu N, Paesmans M, Sculier JP:Chemotherapy of non-small cell lung canceraccording to disease extent: A meta-analysisof the literature. Lung Cancer 7:243-252, 1991.

13.

Dillman R, Seagren S, Propert K, et al:A randomized trial of induction chemotherapyplus high-dose radiation versus radiation alonein stage III non-small cell lung cancer. N EnglJ Med 323:940-945, 1990.

14.

Green MR, Kramar A, Schilsky R, et al:Platinum-based combination chemotherapy inadvanced non-small cell lung cancer: A randomizedphase II trial of the Cancer and LeukemiaGroup B. Med Pediatr Oncol 18:197-202, 1990.

15.

Martini N, Kris MG, Flehinger BJ, et al:Preoperative chemotherapy for stage IIIa (N2)lung cancer: The Memorial Sloan-Ketteringexperience with 136 patients. Ann Thorac Surg55:1365-1374, 1993.

16.

Andre F, Grunenwald D, Pignon JP, etal: Survival of patients with resected N2 nonsmall-cell lung cancer: Evidence for a subclassificationand implications. J Clin Oncol18:2981-2989, 2000.

17.

Goldstraw P, Mannam G, Kaplan D, etal: Surgical management of non-small cell lungcancer with ipsilateral mediastinal node metastasis(N2 disease). J Thorac Cardiovasc Surg107:19-27, 1994.

18.

Martini N, Flehinger BJ, Zaman MB, etal: Results of resection in non-oat cell carcinomaof the lung with mediastinal lymph nodemetastases. Ann Surg 198:386-397, 1983.

19.

Vansteenkiste JF, De Leyn PR, DeneffeGJ, et al: Clinical prognostic factors in surgicallytreated stage IIIA-N2 non-small cell lungcancer: Analysis of the literature (see comments).Lung Cancer 19:3-13, 1998.

20.

McLoud TC, Bourgouin PM, GreenbergRW, et al: Bronchogenic carcinoma: Analysisof staging in the mediastinum with CT by correlativelymph node mapping and sampling.Radiology 182:319-323, 1992.

21.

Dwamena B, Sonnad S, Angobaldo J, etal: Metastases from non-small cell lung cancer:Mediastinal staging in the 1990s-metaanalyticcomparison of PET and CT. Radiology213:530-536, 1999.

22.

Pieterman RM, van Putten JW,Meuzelaar JJ, et al: Preoperative staging of nonsmall-cell lung cancer with positron-emissiontomography [comment]. N Engl J Med343:254-261, 2000.

23.

Funatsu T, Matsubara Y, Ikeda S, et al:Preoperative mediastinoscopic assessment ofN factors and the need for mediastinal lymphnode dissection in T1 lung cancer. J ThoracCardiovasc Surg 108:312-318, 1994.

24.

Gdeedo A, Van S, Corthouts B, et al:Prospective evaluation of computed tomographyand mediastinoscopy in mediastinal lymphnode staging. Eur Respir J 10:1547-1551,1997.

25.

Leschber G, Holinka G, Freitag L, et al:Mediastinoscopy in the staging of bronchialcarcinoma-a critical assessment. Pneumologie54:489-493, 2000.

26.

Hooper R, Tenholder M, Underwood G,et al: Computed tomographic scanning of thebrain in initial staging of bronchogenic carcinoma.Chest 85:774-776, 1984.

27.

Mintz B, Tuhrim S, Alexander S, et al:Intracranial metastases in the initial staging ofbronchogenic carcinoma. Chest 86:850-853,1984.

28.

Earnest FT, Ryu JH, Miller GM, et al:Suspected non-small cell lung cancer: Incidenceof occult brain and skeletal metastasesand effectiveness of imaging for detection-pilot study. Radiology 211:137-145, 1999.

29.

Robnett T, Machtay M, Stevenson J, etal: Factors affecting the risk of brain metastasesafter definitive chemoradiation for locally advancednon-small cell lung carcinoma. J ClinOncol 19:1344-1349, 2001.

30.

Yokoi K, Kamiya N, Matsuguma H, etal: Detection of brain metastasis in potentiallyoperable non-small cell lung cancer: A comparisonof CT and MRI. Chest 115:714-719,1999.

31.

Takita H, Regal A, Antkowiak J, et al:Chemotherapy followed by lung resection ininoperable non-small cell lung carcinomas dueto locally far-advanced disease. Cancer 57:630-635, 1986.

32.

Rapp E, Pater JL, Willan A, et al: Chemotherapycan prolong survival in patients withadvanced non-small cell lung cancer: Reportof a Canadian multicenter randomized trial. JClin Oncol 6:633-641, 1988.

33.

Martini N, Flehinger B, Zaman M, et al:Prospective study of 445 lung carcinomas withmedastinal lymph node metastases. J ThoracCardiovasc Surg 80:390-397, 1980.

34.

Martini N, Flehinger BJ, Nagasaki F, etal: Prognostic significance of N1 disease incarcinoma of the lung. J Thorac CardiovascSurg 86:646-653, 1983.

35.

Rusch VW, Albain KS, Crowley JJ, etal: Surgicial resection of stage IIIA and stageIIIB non-small lung cancer after concurrent inductionchemoradiotherapy. J ThoracCardiovasc Surg 105:97-106, 1993.

36.

Sugarbaker DJ, Herndon J, Kohman LJ,et al: Results of cancer and leukemia group Bprotocol 8935. A multiinstitutional phase IItrimodality trial for stage IIIA (N2) non-smallcelllung cancer. Cancer and Leukemia GroupB Thoracic Surgery Group. J ThoracCardiovasc Surg 109:473-83; (discussion 483-485), 1995.

37.

Elias A, Skarin A, Leong T, et al:Neoadjuvant therapy for surgicallly staged IIIAN2 non-small cell lung cancer. Lung Cancer17:147, 1997.

38.

Rosell R, Gomez-Codina J, Camps C, etal: A randomized trial comparing preoperativechemotherapy plus surgery with surgery alonein patients with non-small cell lung cancer. NEngl J Med 330:153-158, 1994.

39.

Rosell R, Gomez-Codina J, Camps C, etal: Preresectional chemotherapy in stage IIIAnon-small cell lung cancer: A 7-year assessmentof a randomized controlled trial. LungCancer 47:7-14, 1999.

40.

Roth JA, Fossella F, Komaki R, et al: Arandomized trial comparing perioperative chemotherapyand surgery with surgery alone inresectable stage IIIA non-small cell cancer. JNatl Cancer Inst 86:9, 1994.

41.

Roth JA, Atkinson EN, Fossella F, et al:Long-term follow-up of patients enrolled in arandomized trial comparing perioperative chemotherapyand surgery with surgery alone inresectable stage IIIA non-small-cell lung cancer(see comments). Lung Cancer 21:1-6, 1998.

42.

Pass HI, Pogrebniak HW, Steinberg SM,et al: Randomized trial of neoadjuvant therapyfor lung cancer: Interim analysis. Ann ThoracSurg 53:992-998, 1992.

43.

Nagai K, Tsuchiya R, Mori T, et al: Arandomized trial comparing induction chemotherapyfollowed by surgery with surgery alonefor patients with stage IIIA N2 non-small celllung cancer (JCOG 9209). J Thorac CardiovascSurg 125:254-260, 2003.

44.

Depierre A, Milleron B, Moro D, et al:Phase III trial of neo-adjuvant chemotherapy(NCT) in resectable stage I (except T1N0), II,IIIa non-small cell lung cancer (NSCLC): TheFrench experience (abstract 1792). Proc Am SocClin Oncol 18:465a, 1999.

45.

Depierre A, Milleron B, Moro-Sibilot D,et al: Preoperative chemotherapy followed bysurgery compared with primary surgery in resectablestage I (except T1N0), II, and IIIa nonsmall-cell lung cancer. J Clin Oncol 20:247-253, 2002.

46.

Pisters KM, Ginsberg RJ, Giroux DJ, etal: Induction chemotherapy before surgery forearly-stage lung cancer: A novel approach. BimodalityLung Oncology Team. J ThoracCardiovasc Surg 119:429-439, 2000.

47.

Pisters KM: Stage I/II non-small-celllung cancer: Is there a role for combined-modalitytherapy, in: Perry MC (ed): AmericanSociety of Clinical Oncology 2002 EducationalBook, pp 465-470. Alexandria, American Societyof Clinical Oncology, 2002.

48.

Marks R, Streitz J, Deschamps C, et al:Response rate and toxicity of pre-operativepaclitaxel and carboplatin in patients with resectablenon-small cell lung cancer: A NorthCentral Cancer Treatment Group (NCCTG)Study (abstract 1355). Proc Am Soc Clin Oncol20:340a, 2001.

49.

Van Zandwijk N, Smit E, Kramer G, etal: Gemcitabine and cisplatin as induction regimenfor patients with biopsy proven stage IIIAN2 non-small-cell lung cancer: A phase II studyof the European Organization for Research andTreatment of Cancer, Lung Cancer CooperativeGroup (EORTC 08955). J Clin Oncol18:2685-2664, 2000.

50.

Betticher DC, Hsu Schmitz SF, TotschM, et al: Mediastinal lymph node clearanceafter docetaxel-cisplatin neoadjuvant chemotherapyis prognostic of survival in patients withstage IIIA pN2 non-small-cell lung cancer: Amulticenter phase II trial. J Clin Oncol21:1752-1759, 2003.

51.

Mattson KV, Abratt RP, ten Velde G, etal: Docetaxel as neoadjuvant therapy for radicallytreatable stage III non-small-cell lungcancer: A multinational randomised phase IIIstudy. Ann Oncol 14:116-122, 2003.

52.

Krug L, Ng K, Miller V, et al: Inductionchemotherapy with gemictabine, docetaxel, andcisplatin in patients with non-small cell lungcancer: Phase I results (abstract 2053). ProcAm Soc Clin Oncol 20:523a, 2000.

53.

Lopez MG, Lago J, Gonzales-Larriba J,et al: Neoadjuvant chemotherapy withgemcitabine, cisplatin, and docetaxel in stageIIIA N2 and T4N0 non-small cell lung cancer(NSCLC) patients: Spanish Lung Cancer Groupphase II trial 9901 (abstract 2620). Proc AmSoc Clin Oncol 22:651, 2003.

54.

Hanna N, Shepherd F, Rosell R, et al: APhase III study of pemetrexed vs docetaxel inpatients with recurrent non-small cell lung cancer(NSCLC) who were previously treated withchemotherapy (abstract 2503). Proc Am SocClin Oncol 22:622, 2003.

55.

Strauss GM, Herndon JE, Sherman DD,et al: Neoadjuvant chemotherapy and radiotherapyfollowed by surgery in stage IIIA nonsmall-cell carcinoma of the lung: Report of aCancer and Leukemia Group B phase II study.J Clin Oncol 10:1237-1244, 1992.

56.

Albain KS, Rusch VW, Crowley JJ, etal: Concurrent cisplatin/etoposide plus chestradiotherapy followed by surgery for stagesIIIA(N2) and IIIB non-small-cell lung cancer:Mature results of Southwest Oncology Groupphase II study 8805. J Clin Oncol 13:1880-1892, 1995.

57.

Albain KS: Induction chemotherapywith/without radiation followed by surgery instage III non-small-cell lung cancer. Oncology11:51-57, 1997.

58.

Choi NC, Carey RW, Daly W, et al: Potentialimpact on survival of improved tumordownstaging and resection rate by preoperativetwice-daily radiation and concurrent chemotherapyin stage IIIA non-small-cell lungcancer. J Clin Oncol 15:712-722, 1997.

59.

Eberhardt W, Wilke H, Stamatis G, etal: Preoperative chemotherapy followed by concurrentchemoradiation therapy based onhyperfractionated accelerated radiotherapy anddefinitive surgery in locally advanced nonsmall-cell lung cancer: Mature results of a phaseII trial. J Clin Oncol 16:622-634, 1998.

60.

Thomas M, Rube C, Semik M, et al: Impactof preoperative bimodality induction includingtwice-daily radiation on tumor regressionand survival in stage III non-small-celllung cancer. J Clin Oncol 17:1185, 1999.

61.

Albain K, Scott C, Rusch V, et al:Phase III comparision of concurrent chemotherapyplus radiotherapy and CT/RT followedby surgical resection for stage IIIA (pN2) nonsmallcell lung cancer (NSCLC): Initial resultsfrom intergroup trial 0139 (RTOG 93-09) (abstract2497). Proc Am Soc Clin Oncol 22:621,2003.

62.

Fleck J, Camargo J, Godoy D, et al:Chemoradiation therapy versus chemotherapyalone as a neoadjuvant treatment for stage IIINSCLC: preliminary report (abstract 1108).Proc Am Soc Clin Oncol 12:333, 1993.

63.

Sauvaget J, Rebischung J, Vannetzel J:Phase III study of neo-adjuvant MVP versusMVP plus chemoradiotherapy in stage IIINSCLC (abstract 1935). Proc Am Soc ClinOncol 19:495a, 2000.

64.

Argiris A, Liptay M, LaCombe M, et al:A phase I/II trial of induction chemotherapywith carboplatin and gemcitabine followed byconcurrent vinorelbine and paclitaxel withchest radiation in patients with stage III nonsmallcell lung cancer. Lung Cancer 45:243-253, 2004

65.

Pitz CC, Brutel de la Riviere A, ElbersHR, et al: Results of resection of T3 non-smallcell lung cancer invading the mediastinum ormain bronchus. Ann Thorac Surg 62:1016-1020, 1996.

66.

Martin J, Ginsberg R, Abolhoda A, etal: Morbidity and mortality after neoadjuvanttherapy for lung cancer: The risk of right pneumonectomy.Ann Thorac Surg 72:1149-1154,2001.

67.

Novoa N, Varela G, Jimenez M: Morbidityafter surgery for non-small cell lung canceris not related to neoadjuvant chemotherapy.Eur J Cardiothorac Surg 20:700-704, 2001.

68.

Galetta D, Cesario A, Margaritora S, etal: Enduring challenge in the treatment ofnonsmall cell lung cancer with clinical stageIIIB: Results of a trimodality approach. AnnThorac Surg 76:1802-1808, 2003.

69.

Ruffini E, Parola A, Papalia E, et al: Frequencyand mortality of acute lung injury andacute respiratory distress syndrome after pulmonaryresection for bronchogenic carcinoma.Eur J Cardiothorac Surg 20:30-36, 2001.

70.

Grunenwald D, Andre F, Pechoux C, etal: Benefit of surgery after chemoradiotherapyin stage IIIB (T4 and/or N3) non-small cell lungcancer. J Thorac Cardiovasc Surg 122:796-802,2001.

71.

Macchiarini P, Chapelier A, Monnet I,et al: Extended operations after inductiontherapy for stage IIIB (T4) non-small cell lungcancer. Ann Thorac Surg 57:966-973, 1994.

72.

Doddoli C, Thomas P, Thirion X, et al:Postoperative complications in relation withinduction therapy for lung cancer. Eur JCardiothorac Surg 20:385-390, 2001.

73.

Stamatis G, Eberhardt W, Stuben G, etal: Preoperative chemoradiotherapy and surgeryfor selected non-small cell lung cancerIIIB subgroups: Long-term results. Ann ThoracSurg 68:1144-1149, 1999.

74.

Roberts J, Eustis C, Devore R: Inductionchemotherapy increases perioperativecomplications in patients undergoing resectionfor non-small cell lung cancer. Ann Thorac Surg72:885-888, 2001.

75.

Hubaut J, Baron O, Habash OA, et al:Closure of the bronchial stump by manual sutureand incidence of bronchopleural fistula ina series of 209 pneumonectomies for lung cancer.Eur J Cardiothoracic Surg 16:418-423,1999.

76.

Altorki N, Keresztes R, Port J, et al:Celecoxib, a selective cyclo-oxygenase-2 inhibitor,enhances the response to preoperativepaclitaxel and carboplatin in early-stage nonsmallcell lung cancer. J Clin Oncol 21:2645-2650, 2003.

77.

Shepherd FA, Johnston MR, Payne D,et al: Randomized study of chemotherapy andsurgery versus radiotherapy for stage IIIA nonsmall-cell lung cancer: A National Cancer Instituteof Canada Clinical Trials Group Study.Br J Cancer 78:683-685, 1998.

78.

Johnstone D, Byhardt R, Ettinger D, etal: Phase III study comparing chemotherapy andradiotherapy with preoperative chemotherapyand surgical resection in patients with non-smallcell lung cancer with spread to mediastinallumph nodes (N2): Final report of RTOG 89-01. Radiation Oncology Group. Int J RadiatOncol Biol Phys 54:365-369, 2002.

79.

de Boer RH, Smith IE, Pastorino U, etal: Pre-operative chemotherapy in early stageresectable non-small-cell lung cancer: A randomizedfeasibility study justifying amulticentre phase III trial. Br J Cancer79:1514-1518, 1999.