Treatment of Advanced Non-Small-Cell Lung Cancer in Special Populations

September 1, 2004

Lilenbaum’s paper highlightingrecent controversies in the managementof advanced non–small-cell lung cancer (NSCLC) inthe elderly and in vulnerable performancestatus (PS) populations is bothtimely and relevant. A recent Surveillance,Epidemiology and End Results(SEER) analysis suggests that nearly50% of all patients diagnosed withNSCLC are 70 years of age or older.Non–small-cell lung cancer generallypeaks in incidence in the elderly, andthe population of the United States iscontinually aging, with nearly 20%expected to be over age 65 by theyear 2030.[1]

Lilenbaum's paper highlightingrecent controversies in the managementof advanced non-small-cell lung cancer (NSCLC) inthe elderly and in vulnerable performancestatus (PS) populations is bothtimely and relevant. A recent Surveillance,Epidemiology and End Results(SEER) analysis suggests that nearly50% of all patients diagnosed withNSCLC are 70 years of age or older.Non-small-cell lung cancer generallypeaks in incidence in the elderly, andthe population of the United States iscontinually aging, with nearly 20%expected to be over age 65 by theyear 2030.[1]Older Patients in Clinical Trials
Historically, older patients havebeen underrepresented in clinical trials.There are several putative explanationsfor this observation[2]:
(1) Therapeutic nihilisim that imbuesmany practitioners as well astheir patients, and that limits the expectationsof older individuals;
(2) Lack of financial, social, or logisticsupport for trial participation;
(3) Increased incidence of functionaldependency and comorbidities,which increase with age and may compromiseeligibility for therapy;
(4) Outdated eligibility criteria thatexplicitly exclude older individualsfrom trial participation-even thosewith intact PS;
(5) Societal pressure, particularlycommon in European countries withnationalized health care, where resourcesare greatly constrained;
(6) Lack of niche trials designedspecifically for the elderly.The landmark trials of Gridelli etal (Elderly Lung Cancer VinorelbineItalian Study [ELVIS] and MulticenterItalian Lung Cancer in the ElderlyStudy [MILES]) have essentially definedthe state of the art for this population.In addition to yielding improvedsurvival and quality of life, vinorelbine(Navelbine) in the ELVIS trial[3] resulted in only a 10% incidenceof grade 3/4 toxicity. Of those enrolled,24% also had a compromisedperformance status (Eastern CooperativeOncology Group [ECOG] PS 2);among the 19 patients with a compromisedPS randomized to best supportivecare in the ELVIS trial, the mediansurvival was only 8 weeks, comparedto 26 weeks for PS 2 patients randomizedto vinorelbine, despite the absenceof objective responses in thiscohort.SICOG Trial
Multiple investigators in phase IIstudies have demonstrated the feasibilityof single-agent gemcitabine(Gemzar) in the elderly population.Frasci and colleagues,[4] who conductedthe Southern Italy CooperativeOncology Group (SICOG) trial,demonstrated a median survival timeof 29 weeks for the combination ofgemcitabine and vinorelbine vs only18 weeks for single-agent vinorelbine,and commensurate 1-year survivalrates of 30% and 13%, respectively.In addition, outcome was highlycontingent on baseline PS and Charlsoncomorbidity score.[4] Patientswith PS 2 had a median survival ofonly 4.4 months, compared to 6months for those with PS 0/1. Thosewith two or more comorbidities had amedian survival of 3.7 months and a much greater likelihood of discontinuingtherapy early compared to theCharlson 1 group, for whom the mediansurvival was 4.8 months, and theCharlson 0 group, for whom the mediansurvival was 6.5 months.MILES Trial
The track record of the single-agentcontrol arm in the SICOG trial issomewhat disturbing.[4] Median survivalin this group proved no betterthan that in the best supportive carearm in the prior ELVIS trial.[3] Inthis regard, the MILES trial is potentiallymore credible.[5] The SICOGtrial accrued just over 150 patients,whereas the MILES effort accruednearly 700 individuals.As discussed by Dr. Lilenbaum, incontrast to the SICOG effort, theMILES effort failed to show a survivalbenefit for the combination of gemcitabineand vinorelbine vs theconstituent single agents. Unlike mostNorth American trials in advancedNSCLC, a higher proportion of patients-in this case, 31%-hadstage IIIB disease. In addition, the plurality(49%) had squamous cell histology.Nearly 50% had three or morecomorbidities. The distribution ofcomorbidities was well balanced betweeneach arm. The combination armshowed significantly more neutropenia,diarrhea, fatigue, vesicantreactions, cardiac toxicity, and constipationcompared to the gemcitabinesingle-agent arm and significantlymore myelosuppression compared tosingle-agent vinorelbine.PS 2 patients had no overt increasein toxicity, but sustained a clear reductionin survival: 1-year survivalrates among PS 2 patients were 18%to 23%, compared to 30% to 43% forPS 0/1 patients. Those with higherbaseline quality of life had far bettersurvival compared to those with intermediateor inferior scores. Thosewith a baseline quality-of-life assessment of > 67% had a median survivalof 53 weeks, compared to only 20weeks for those with a baseline assessmentof < 42%. By the same token,those with an intermediate orbetter baseline Instrumental Activitiesof Daily Living (IADL) score hadimproved survival compared to thosewith poorer scores.Retrospective Analyses
Based on these results, Gridelli andcolleagues consider a nonplatinumsingle agent the standard of comparisonin the treatment of elderly individualswith advanced NSCLC. Manyinvestigators, however, dispute thisnotion, citing-as Dr. Lilenbaum hashighlighted-several important retrospectiveanalyses in which platinumbasedtherapy proved feasible in theelderly.In addition to the studies the authorenumerates, Belani and colleagueshave reviewed outcome inolder patients enrolled in TAX 326, arandomized phase III effort evaluatingdocetaxel in combination witheither cisplatin or carboplatin (Paraplatin)vs a reference regimen of vinorelbine/cisplatin.[6] Of 1,218patients enrolled, 410 were 65 yearsof age or older. Elderly individualsenrolled in the docetaxel/cisplatin armhad a median survival of 12.6 monthscompared to only 9.9 months for thosein the vinorelbine/cisplatin arm. Therespective 1- and 2-year survival rateswere 52% and 24% for docetaxel/cisplatinvs 41% and 17% for vinorelbine/cisplatin. This difference wasstatistically significant, with a hazardratio of 1.34 (confidence interval[CI] = 1.011-1.776). There was nosignificant difference in outcome betweenelderly individuals enrolled inthe docetaxel/carboplatin vs vinorelbine/cisplatin arm. The elderly, however,experienced more toxicity-inparticular, a higher incidence of grade3/4 asthenia, infection, and pulmonarytoxicity-and in the cisplatinarms, more diarrhea and sensoryneurotoxicity.In a landmark phase III trial evaluatinga truncated course of paclitaxel/carboplatin (four cycles) vs indefiniteadministration until disease progression, Hensing et al analyzed outcomesegregated by age.[7] The elderly experiencedmore fatigue, but no otherexacerbation of toxicity compared tothe younger cohorts. Median survivaland 1- and 2-year survival rates wereno different (Table 1). Despite theabsence of significant differences intherapeutic outcome, older patientswere much less likely to completetreatment as planned (16% vs 32%,P = .02) and were much more likelyto stop treatment either because oftoxicity or general reservations.Additional Data
In another randomized phase II trialevaluating various applications ofweekly paclitaxel in combination witheither weekly or monthly carboplatin,the least toxic arm (carboplatin at anarea under the concentration-timecurve [AUC] of 6 every 4 weeks andpaclitaxel at 100 mg/m2 on days 1, 8,and 15 every 4 weeks) had the besttherapeutic index, with response ratesas well as median and 1-year survivalrates equaling, if not exceeding, themore toxic approaches.[8] Among the44 elderly individuals enrolled in thisarm, there was a somewhat higherincidence of grade 4 neutropenia, butno other significant differences in toxicitycompared to younger enrollees.Median survival time for those ≥ 70years old was 49 weeks, with 1- and2-year survival rates of 50% and 23%respectively-no different from theresults observed in patients under age70 (median survival time of 48 weeks,1-year survival of 46%, 2-year survivalof 11%).Few data exist for patients overage 80. Although patients betweenthe ages of 70 and 80 who are fitgenerally do as well as, or nearly aswell as, their younger counterparts(albeit with more toxicity, particularlymyelosuppression), those over age80 may fare worse. In a nested analysisof ECOG 1594, with more than1,200 patients enrolled, only 9 wereage 80 or older.[9] In this group, therewere no responses. Only one individualwas able to tolerate more thanthree cycles of therapy. Progressionfreesurvival was only 2.2 months,compared to 3.7 months for patients aged 70 to 79, and median survivalwas 4.2 vs 8.2 months in the cohort of70- to 79-year-olds.

Elderly NSCLC Patients:Conclusions
Although the elderly are more likelyto be frail and vulnerable comparedto younger patients, those whoare fit seem to do as well as theiryounger counterparts. While level 1,evidence-based data show a clear benefitfor nonplatinum monotherapy (ie,vinorelbine) vs best supportive carein elderly PS 0-2 NSCLC patients, noconsistent superiority has been shownfor nonplatinum doublets. Comorbiditiesappear to be predictive of outcome.Retrospective analyses ofplatinum-based combinations havedemonstrated similar outcomes for fitelderly and younger cohorts with respectto response rate, time to progression,median survival, and 1- and2-year survival rates, although thesebenefits are vitiated by increasedtoxicity.Decisions regarding therapy in theelderly hinge on multiple factors, includingthe toxicity profile of theagent(s) under consideration, performancestatus, pharmacokinetics, comorbidities,organ function, and socialsupport. Patients between the ages of70 and 80 and those over age 80 mayconstitute two functionally distinct cohorts.In this regard, special considerationmust be given to comorbidity,functional status (ie, IADL) as well asthe mental state (psychological and cognitive) of elderly individuals withadvanced NSCLC. Logistics also playa key role, as does baseline nutritionof potential study participants.Several outstanding issues remain.To date, there has been no elderlyspecific,phase III trial comparing asingle agent to the identical singleagent in combination with platinum.In North America, at least, a morecomprehensive analysis of comorbiditiesand their influence on toxicity,treatment tolerance, quality of life,and survival is needed. We also needto expand our data set on NSCLCpatients 80 years of age or older-arapidly expanding cohort of patients,for whom very little information exists.Finally, in the context of clinicaltrials, we should at least consider nestedpharmacokinetic analyses.PS 2 NSCLC
Using the database of multipleECOG studies through the early1990s, Mike Jiroutec, a statisticianformerly associated with ECOG,conducted a recursive partitioninganalysis of prognostic factors inNSCLC.[10] These trials evaluatedsingle agents as well as platinumbasedcombinations. The three mostimportant prognostic factors includedperformance status, gender, and appetite(and by inference, weight loss).A woman with an intact appetiteand excellent performance status (PS 0)could expect a median survival of over12 months, whereas the same womanwith a compromised status (PS 2) anddiminished appetite had a median survivalof only 2.3 months. This representsa greater than fivefold differencein therapeutic expectations, basedpurely on clinical factors that are readilyidentifiable within the first 20 secondsof a new patient evaluation. Todate, there are no therapeutic combinationsthat can bridge this gap. Thesedifferences speak to the intrinsic biologyof the tumor as well as the intrinsicbiology and physiologic functionof the patient.Dr. Lilenbaum has ably reviewedoutcome for PS 2 individuals enrolledin North-American trials. Outside theUnited States, trends are identical. Ina phase III trial, Alberola et al compared cisplatin/gemcitabine to the triplet,cisplatin/gemcitabine/vinorelbine,as well as to sequential nonplatinumdoublets (gemcitabine/vinorelbine->ifosfamide (Ifex)/vinorelbine).[11]PS 2 individuals had a median survivalof only 4.8 months, comparedto 9.11 months in PS 0/1 individuals.Van Meerbeeck et al initiated aphase III trial comparing cisplatincombinations with either paclitaxel orgemcitabine vs the nonplatinum doublet(paclitaxel/gemcitabine).[12]Median survival in the PS 2 cohortwas only 3.3 months, compared to8.6 months in the PS 0/1 cohort. Anearly identical trend was observedby Kosimidis in a phase III trial evaluatingcarboplatin (AUC 6) in combinationwith paclitaxel at either alow dose (175 mg/m2) or high dose(225 mg/m2).[13] Median survival inthe PS 2 cohort was 3.8 months vs11.25 months in the PS 0/1 cohort.Improved Median Survival
Compared to these historic series,median survival in ECOG 1599[14]-a randomized phase II trial of cisplatinat 60 mg/m2 every 3 weeks andgemcitabine at 1 g/m2 every 3 weeksor carboplatin (AUC 6) and paclitaxelat 200 mg/m2 every 3 weeks-wassomewhat better: 6.8 months for thecisplatin/gemcitabine arm and 6.1months for the carboplatin/paclitaxelarm. That said, these improvementsin median survival did not necessarilytranslate to improved 1-year survivalrates (25% and 19%, respectively).This trial has the distinction of beingone of the first specifically restrictedto PS 2 individuals. In addition, ithas reinforced the notion that combinationtherapy is, at the very least,feasible in vulnerable patients-an observation previously cited byLilenbaum et al in a subanalysis ofPS 2 individuals enrolled inCALGB 9730.[15]Although Lilenbaum's subanalysiswas performed in the context of amuch larger prospective randomizedphase III trial, the observations areprovocative.[15] The median survivalof PS 2 individuals receiving combinationpaclitaxel/carboplatin wasdouble that observed for those receiving single-agent paclitaxel (4.7 vs 2.4months); and there was a commensurateincrease in 1-year survival rates(18% vs 10%, log rank P = .0123).There were no 2-year survivors amongthe PS 2 individuals receiving singleagentpaclitaxel, whereas the 2-yearsurvival rate was nearly 10% in thePS 2 cohort receiving combinationtherapy.Neurotoxicity remains a major sequelaof treatment in patients receivingevery-3-week paclitaxel. Belaniand others have looked at weekly therapywith paclitaxel in this setting.[8]In a randomized phase II effort, paclitaxelat 100 mg/m2 on days 1, 8, and15 every 4 weeks in combination withcarboplatin at AUC 6 had the besttherapeutic index compared to otheriterations of this combination. Mediansurvival among PS 2 individuals inthat arm was 26 weeks, and the 1-year survival rate was 28%.[8]Recently Completed Efforts:ASCO 2004
A number of recently completedtrials have also proven revealing. Kosimidiset al evaluated gemcitabine,either as a single agent or in combinationwith carboplatin administeredevery other week.[16] Median survivalin the combination arm was 6.7months, compared to 4.8 months forsingle-agent gemcitabine therapy(P = .49). However, the combinationresulted in significantly more myelosuppressionand no incremental improvementin subjective clinicalresponse.On the other hand, single-agentstudies evaluating either docetaxelalone or sequential vinorelbine followedby docetaxel have yieldedsomewhat poorer results. In a phase IItrial reported by Hesketh et al, among44 patients with a compromised status(PS 2), the response rate was only10%, median survival only 4 months,and 1-year survival rate 14%, comparedto 21%, 9 months, and 40% fora contemporaneous good performancestatus, elderly cohort.[17] In a randomizedphase II trial, Lilenbaum etal evaluated docetaxel at 75 mg/m2every 3 weeks, or 30 mg/m2 on days1, 8 and 15 every 4 weeks.[18] Here again, median survival time was ratherpoor, ie, 2.4 months for the PS 2cohort.CT-2103 (Xyotax) is a polyglutamatedversion of paclitaxel, withmuch lower likelihood of plasma esterasedegradation, and far better tumorcell penetration and intracellularretention.[19-21] Phase I and II studieshave demonstrated the feasibilityof this agent, both alone and in combinationwith carboplatin.[21-23] Twoseparate phase III trials in the PS 2cohort have been mounted: SelectiveTargeting for Efficacy in Lung cancerLower Adverse Reactions(STELLAR) 3 compares standard paclitaxeland carboplatin to combinationCT-2103 (210 mg/m2) andcarboplatin (AUC 6). Both regimensare administered at 3-week intervals.This study completed accrual in thefall of 2003, with over 400 patientsenrolled. STELLAR 4 compares single-agent CT-2103 to either gemcitabineor vinorelbine. This study hasalso completed accrual (n = 460+).The results of each effort should beavailable in the first quarter of 2005.PS 2 NSCLC Patients:Conclusions
Patients designated as PS 2 representa sizable percentage of the individualsevaluated in our clinicalpractices. Heretofore, they have beenlargely excluded from clinical research,but over the past 5 years, renewedinterest has focused on thisgroup. Emerging data suggest thatchemotherapy can improve diseaserelatedsymptoms and may prolongsurvival compared to best supportivecare. Combination regimens may havethe greatest impact on PS 2 patients,but toxicity is clearly exacerbated, andoverall, prognosis remains poor. Studiesneed to differentiate patients whoare PS 2 on the basis of comorbidity(and are, therefore, less likely to benefitfrom cytotoxic therapy) from thosewho have rapidly expanding diseaseburden and who may be more likelyto benefit.Patients designated as PS 2 representa sizable percentage of the individualsevaluated in our clinicalpractices. Heretofore, they have beenlargely excluded from clinical research,but over the past 5 years, renewedinterest has focused on thisgroup. Emerging data suggest thatchemotherapy can improve diseaserelatedsymptoms and may prolongsurvival compared to best supportivecare. Combination regimens may havethe greatest impact on PS 2 patients,but toxicity is clearly exacerbated, andoverall, prognosis remains poor. Studiesneed to differentiate patients whoare PS 2 on the basis of comorbidity(and are, therefore, less likely to benefitfrom cytotoxic therapy) from thosewho have rapidly expanding diseaseburden and who may be more likelyto benefit.


Dr. Langer receivesgrant/research support from, is a scientific advisorfor, and is a member of the speaker’s bureaufor Bristol Myers Squibb, Pharmacia, Lilly,Schering-Plough Research Institute, Aventis,Amgen, Cell Therapeutics Inc, OrthoBiotech,Celgene, Vertex, Genentech, AstraZeneca,Pfizer, Active Biothech Medimmune, ImClone,Intrabiotics, GlaxoSmithKline, Pharmacyclics,Novartis, and Sanofi-Synthelabo Novartis.



Yancik R, Reis LA: Aging and cancer inAmerica. Demographic and epidemiologic perspectives.Hematol Oncol Clin North Am 14:17-23, 2000.


Hutchins LF, Unger JM, Crowley JJ, etal: Underrepresentation of patients 65 years ofage or older in cancer-treatment trials. N EnglJ Med 341:2061-2067, 1999.


The Elderly Lung Cancer Vinorelbine ItalianStudy Group: Effects of Vinorelbine onquality of life and survival of elderly patientswith advanced non-small cell lung cancer. JNatl Cancer Inst 91:66-72, 1999.


Frasci G, Lorusso V, Panza N, et al:Gemcitabine plus vinorelbine vs vinorelbinealone in elderly patients with advanced nonsmall-cell lung cancer. J Clin Oncol 18:2529-2536, 2000.


Gridelli C, Perrone F, Gallo C, et al: Chemotherapyfor elderly patients with advancednon-small cell lung cancer: The MulticenterItalian Lung Cancer in the Elderly Study(MILES) phase III randomized trial. J NatlCancer Inst 95:362-372, 2003.


Fossella F, Pereira JR, von Pawel J, et al:Randomized, multinational, phase III study ofdocetaxel plus platinum combinations versusvinorelbine plus cisplatin for advanced non–small cell lung cancer: The TAX 326 StudyGroup. J Clin Oncol 21:3016-3024, 2003.


Hensing TA, Socinski MA, Schell MJ, etal: Age does not alter toxicity or survival forpatients with stage IIIB/IV NSCLC treated withcarboplatin and paclitaxel (abstract 1382). ProcAm Soc Clin Oncol 20:346a, 2001.


Belani CP, Barstis J, Perry M, et al: PhaseII multicenter randomized trial of weeklypaclitaxel (P) administered in combination withcarboplatin (C) followed by maintenance P vsobservation for patients (Pts) with advanced &metastatic non-small cell lung cancer (NSCLC)(abstract 1287). Proc Am Soc Clin Oncol20:323a, 2001.


Langer CJ, Vangel M, Schiller J, et al:Age-specific subanalysis of ECOG 1594: Fitelderly patients (70-80 YRS) with NSCLC doas well as younger pts (10. Jiroutek M, Johnson D, Blum R, et al:Prognostic factors in advanced non-small celllung cancer (NSCLC): Analysis of EasternCooperative Oncology Group (ECOG) trialsfrom 1981-1992 (abstract 1774). Proc Am SocClin Oncol 17:461a, 1998.


Alberola V, Camps C, Provencio M, etal: Cisplatin plus gemcitabine versus acisplatin-based triplet versus nonplatinum sequentialdoublets in advanced non-small-celllung cancer: A Spanish Lung Cancer Groupphase III randomized trial. J Clin Oncol21:3207-3213, 2003.


Van Meerbeeck JP, Smit EF, Lianes P, etal: Three-arm randomized study of twocisplatin-based regimens and paclitaxel plusgemcitabine in advanced non-small-cell lung cancer: A phase III trial of the European Organizationfor Research and Treatment of CancerLung Cancer Group-EORTC 08975. J ClinOncol 21:3909-3917, 2003.


Kosmidis P, Mylonakis N, Nicolaides C,et al: Paclitaxel plus carboplatin versusgemcitabine plus paclitaxel in advanced nonsmall-cell lung cancer: A phase III randomizedtrial. J Clin Oncol 20:3578-3585, 2002.


Langer C, Stephenson P, Schiller J, etal: ECOG 1599: Randomized phase II study ofpaclitaxel/carboplatin vs cisplatin/gemcitabinein performance status (PS) 2 patients with treatment-naïve advanced NSCLC (abstract O-52).Lung Cancer 41(suppl 2):S18, 2003.


Lilenbaum RC, Herndon J, List M, etal: Single-agent (SA) versus combination chemotherapy(CC) in advanced non-small celllung cancer (NSCLC): A CALGB randomizedtrial of efficacy, quality of life (QOL), and costeffectiveness(abstract 2). Proc Am Soc ClinOncol 21:1a, 2002.


Kosimidis PA, Dimopoulos MA, SyrigosC, et al Gemcitabine vs gemcitabinecarboplatinfor patients with advanced nonsmallcell lung cancer and PS 2: A prospectiverandomized phase II study of the Hellenic CooperativeOncology Group (abstract 7058).Proc Am Soc Clin Oncol 23:627, 2004.


Hesketh PJ, Chansky K, Lau KH, et al:Sequential vinorelbine and docetaxel in advancednon-small cell lung cancer patients age> 70 or with performance status 2: A SWOGphase II trial (S0027) (abstract 7056). Proc AmSoc Clin Oncol 23:627, 2004.


Lilenbaum R, Rubin M, Samuel J, et al:A phase II randomized trial of docetaxel weeklyor every 3 weeks in elderly and/or poor performancestatus patients with advanced non-smallcell lung cancer (abstract 7057). Proc Am SocClin Oncol 23:627, 2004.


Singer JW, Baker B, De Vries P, et al:Poly-(L)-glutamic acid-paclitaxel (CT-2103)[XYOTAX], a biodegradable polymeric drugconjugate: Characterization, preclinical pharmacology,and preliminary clinical data. AdvExp Med Biol 519:81-99, 2003.


Li C, Price JE, Milas L, et al: Antitumoractivity of poly(L-glutamic acid)-paclitaxel onsyngeneic and xenografted tumors. Clin CancerRes 5:891-897, 1999.


Skubitz K, Kudelka A, Bolton MG:Phase 1 study of CT-2103/cisplatin in patientswith solid tumors (abstract 554). Eur J Cancer1(5):S168, 2003.


Nemunaitis J, Garzone P, Bolton MG:Phase I study of XYOTAX (CT-2103) andcarboplatin in patients with solid tumors: Preliminarydata (abstract 438). Eur J Cancer38(suppl 7):132, 2002.


Bodkin D, Neubauer M, Bolton MG:Phase 2 study of first line chemotherapy usingCT-2103 (XYOTAX) in patients with nonsmall-cell lung cancer who are >69 years ofage or who have performance status (PS) = 2(abstract 806). Eur J Cancer 1(suppl 5):S242,2003.