`
`Twenty-Two Years of Phase III Trials for Patients With
`Advanced Non-Small-Cell Lung Cancer: Sobering Results
`
`By Oscar S. Breathnach, Boris Freidlin, Barbara Conley, Mark R. Green, David H. Johnson, David R. Gandara,
`Michael O’Connell, Frances A. Shepherd, and Bruce E. Johnson
`
`Purpose: To determine the changesinclinical trials
`and outcomes of patients with advanced-stage non-
`small-cell lung cancer (NSCLC) treated on phaseIll ran-
`domized trials initiated in North America from 1973 to
`1994,
`Patientsand Methods: PhaseIll trials for patients
`with advanced-stage NSCLC were identified through a
`search of the National CancerInstitute’s Cancer Ther-
`apy Evaluation Program database from 1973 to 1994,
`contact with Cooperative Groups, and byliterature
`search of MEDLINE. Patients with advanced NSCLC
`treated during a similar time interval were also exam-
`ined in the SEER database. Trends were tested in the
`numberof trials, in the number and sex of patients
`entered on thetrials, and in survival over time.
`Results: Thirty-three phaseIll trials were initiated
`between 1973 and 1994. Twenty-four trials (73%)
`were initiated within thefirst half of this period (1973 to
`
`YSTEMIC CHEMOTHERAPY for patients with ad-
`vanced-stage non-small-cell
`lung cancer
`(NSCLC)
`prolongs survival and palliatcs symptoms compared with
`best supportive care alone.’? Meta-analysis of patients with
`advanced NSCLC who were treated with cisplatin-based
`therapy shows a modest improvementin survival.’ Despite
`systemic chemotherapy,
`lung cancer remains the greatest
`cause of cancer-related mortality within the United States.*
`The 5-year survival of patients with advanced-stage NSCI.C
`is less than 5%.
`
`
`
`Fromthe Lowe Center for Thoracic Oncology, Department ofAdult
`Oncology, Dana-Farber Cancer Institute, and Departments of Medi-
`eme, Brigham and Women’s Hospital and Harvard Medical School.
`Boston, MA; Biometric Researeh Branch, Clinical Investigations
`Branch, Cancer Therapy Evaluation Program, Division of Cancer
`Treaumem and Diagnosis, National Cancer Institute, Bethesda, MD;
`Cancer and Leukemia Group B, Chicago, IL; Eastern Cooperative
`Oncology Group, Pittsburgh, PA; Southwest Oncology Group, San
`Antome, TX; North Central Cancer Treatment Group Philadelphia,
`PA; and National Cancer Institute of Canada Clinical Trials Group.
`Kingston, Ontario, Canada.
`Submined June 30, 2000; accepted November 20, 2000.
`Address reprint requests to Oscar S. Breathnach, MD, Thoracic
`Oncology Program, Dana-Farber Cancer Insutute, Sie 1234, 44
`Binney St, Boston, MA 02115.
`© 200] by American Society of Clinical Oncology.
`0732-183X/01/1906-1734
`
`1983) and accounted for 5,359 (64%) of the 8,434
`eligible patients. The median number of patients
`treated per arm of the trials rose from 77 (1973 to
`1983) to 121 (1984 to 1994) (P < .001).Five trials (15%)
`showed a statistically significant difference in survival
`between treatment arms, with a median prolongation
`of the median survival of 2 months (range, 0.7 to 2.7
`months).
`Conclusion: Analysis of pasttrials in North America
`shows that the prolongation in median survival be-
`tween two armsof a randomized study wasrarely in
`excess of 2 months. Techniques for improved use of
`patient resources and appropriate trial design for
`phase Ill randomized therapeutic trials with patients
`with advanced NSCLC need to be developed.
`J Clin Oncol 19:1734-1742. © 2001 by American
`Society of Clinical Oncology.
`
`trials for patients with
`Prior analysis of phase III
`extensive-stage small-cell lung cancer (SCLC) revealed
`that the median of median survival ofpaticnts treated on
`phaseII] trials and of patients recorded within the SEER
`database increased by 2 months between 1972 and 1990.4
`We therefore chose to review all the published North
`American Cooperative Groups andinstitutional phase III
`randomized therapeutic trials for patients with advanced-
`stage NSCI.C initiated during the time period of 1973 to
`1994 in order to assess the impact of systemic chemo-
`therapy on survival over time. As platinum-based regi-
`mens were uscd throughout the period under review, we
`divided the analyzed trials for comparison purposes
`according lo two equivalent ime periods: 1973 through
`1983 and 1984 through 1994. Trials assessing combined-
`modality therapy of chemotherapy plus chest irradiation
`were not included in the analysis.
`We investigated whether the number oftrials initiated
`had changed, whether the numberof patients enrolled on the
`trials had increased or decreased over time, and whether
`survival ofpatients treated on these trials had improved. ‘lhe
`median and 5-year survival
`information on patients with
`advanced-stage NSCIC fromthe last 25 years is available
`on the SEER database. This database provides the data with
`which to compare the outcomes of patients with advanced-
`stage NSCLC in both cooperative groups and in a popula-
`tion database.
`
`1734
`
`JournalofClinical Oncology, Vol 1%, No 6 (March 15), 2001: pp 1734-1742
`
`Downloaded from jco.ascopubs.org on October 24, 2010. For personal use ei No other uses without permission.
`Ir
`Copyright © 2001 American Society of Clinical Oncology. A\
`ights reserved.
`OSI EXHIBIT 2009
`APOTEXV. OSI
`IPR2016-01284
`
`OSI EXHIBIT 2009
`APOTEX V. OSI
`IPR2016-01284
`
`

`

`PHASEIll TRIALS FOR ADVANCED NON-SMALL-CELL LUNG CANCER
`
`1735
`
`PATIENTS AND METHODS
`
`Phase Ill Trials
`
`Phase If] trials initiated for paticnts with advanced (inoperable stage
`Ill and IV) NSCLC were identified through a search of the National
`Cancer Institute’s Cancer Therapy Evaluation Program database from
`1973 to 1994, by computer-based search of MEDLINE, and by direct
`contact with the Eastern Cooperative Oncology, Southwest Oncology.
`North Central Cancer Treatment, and National Cancer Institute of
`Canada-Clinical Trials cooperative groups, and Cancer and Leukemia
`Group B. We evaluated phase III trials initiated between 1973 and
`1994, because this allowed adequate time for patient entry on study,
`follow-up, and publication of the mature results of the therapeutic
`trials. This period also included the interval examined in the meta-
`analysis by the Non-Small Cell Lung Cancer Collaborative Group,
`which compared systemic chemotherapyand best supportive care with
`best supportive care alone (11 trials; 1970 to 1985).
`The lung cancer committce chairs from cach cooperative group in
`North America. were contacted to notify them of this analysis and to
`inquire whether additional. trials had been performed that were not
`known to the authors. Information was obtained about each trial
`regarding the years of enrollment within the trials,
`the number of
`patients enrolled on study, the sex of the patients, treatment regimens
`used, response rates, median survivals, the number of patient deaths at
`the timeof the analysis, and differencesin overall survival. Tho median
`of the median survival times was computed without adjusting for the
`sample size of the control group, and the median survival time of
`patients before and after 1983 was not adjusted fortrial size. Each study
`median was entered as a data point. Trials included both patients with
`unresectable locally advanced-stage III disease and/or metastatic dis-
`ease,
`in which chemotherapy was the primary therapeutic modality
`used in all paticnts. The inclusion critcria were similar to thosof the
`preceding meta-analysis.’ The control arms were identified in each of
`the phase III trials on the basis of the statements of the authors in the
`published reports. A minority of the phase III trials did not have all of
`this information available in published articles. In these cases,
`the
`author of the study and the cooperative group statistician were
`contacted and additional data were obtained if available and permitted
`by the author.
`The SEER database was also examined to compare the median and
`5-year survival information for patients with advanced-stage NSCLC
`versus our analysis ofpatients treated on phase III trials over the same
`lime period. The lime period from 1973 to 1995 was examined because
`the starting year corresponded most closely with the start of the
`cooperative studies, and 1995 corresponded most closely with the year
`that the cooperative grouptrials finished their accrual. In the SEER
`database, patients with metastatic disease are categorized as “distant
`disease.” The SEER definition of distant disease is a neoplasmthat has
`spread to parts of the body remote from the primary tumoreither by
`direcl extension or by discontinuous metastasis.
`In this analysis,
`patients from the SEER database with distant disease are termed
`“advanced stage” to be consistent with the terminology used in the
`cooperative group studics.
`
`'
`
`Statistical Method
`
`Theinformation from the phaseIII trials was evaluated with multiple
`regression analyses using the ordinary least squares regression method,
`including year of studyinitiation, platinum-based chemotherapy, max-
`imum performance status, treatment group, and brain metastases in
`order to determine whether these factors independently impacted the
`
`survival of patients with advanced-stage NSCLC whoweretreated on
`the. phase IH studies over time. The significance of each explanatory
`variable was tested by a ¢ test. The trend in median survival and 5-year
`survival among patients with advanced NSCLC from the SEER
`database wascalculated using 4 least squares regression analysis. The
`two-sample 1 test and Wilcoxon test were used to compare data for
`1973 to 1983 versus 1984 to 1994 periods. All P valués corresponded
`to two-sided tests.
`
`RESULTS
`
`Phase Ill Trials
`
`Forty-three randomized, controlled phase III trials initi-
`ated between 1973 and 1994 in North America involving
`9,215 patients with advanced non-small-cell lung cancer
`were identified.>*? Ten trials were found to be incligible for
`the purpose of this analysis. Seven trials with 637 patients
`did not
`report
`the dates of
`initiation of
`the
`tn-
`als,13:14:1733.36.45 two trials with 333 patients were reported
`as subgroup analyses, reporting only on their patients with
`squamous cell lung cancer and epidermoid lung cancers,
`respectively;??*! and onetrial with 109 patients included
`patients with small-cell lung cancer.* Several attempts were
`made to contact the authors of the seven trials that had no
`reported date of initiation, but we were unsuccessful
`in
`gaining this information. ‘The exclusion of this data did not
`appreciably altcr the outcomeofthe study. We analyzed the
`treatment arms from the remaining 33 published phase-III
`trials that were initiated from 1973 to 1994. The results of
`these 33 trials are listed in Table 1.
`A total of 8,434 patients with advanced-stage NSCLC
`were treated on these 33 phase III trials. Twenty-four trials
`(73%) were initiated within the first half of the period under
`analysis (1973 to 1983) and accounted for 64% of treated
`patients (n = 5,359). The percentage of womentreated did
`not change significantly between the 1973-to-1983 and
`1984-to-1994 time periods (40% and 39%, respectively).
`Eight of the nine trials (89%) initiated between 1984 and
`1994 included only patients with performance status 0-2,
`compared with 50% of trials initiated between 1973 and
`1983. Patients with brain metastases were included in three
`
`of the trials initiated between 1984 and 1994, which
`compares to 14 ofthetrials initiated between 1973 to 1983.
`In six trials-initiated between 1973 and 1983, the authors do
`not specify whether the presence of brain metastases was
`considered in the eligibility criteria.0?4!9?'%? The vast
`majority of trials (97%) included only patients not previ-
`ously treated with chemotherapy. Seventeen percent of
`patients in a single trial had received prior chemotherapy.’
`A medianof94 patients were treated per regimen in the
`33 phaseIII trials initiated from 1973 to 1994 (range, 19 to
`209). The median numberofpatients treated per arm ofthe
`
`Downloaded from jco.ascopubs.org on October 24, 2010. For personal use only. No other uses without permission.
`Copyright © 2001 American Society of Clinical Oncology. All rights reserved.
`.
`
`

`

`1736
`
`BREATHNACHET AL
`
`Table 1. North American Phase Ill Trials of Advanced-Stage Non-Small-Cell tung Cancer
`No.of
`
`
`No.of 1-Year—No.ofMedian
`Patients
`Sex Male/
`Disease
`Female
`per
`StudyYears
`
`Brain Survival~—sSurvival. DeathsResponders
`
`Am
`PS
`{no.}
`First Author (ref)
`Stage
`Metastases
`CR/PR
`Regimens
`{months)
`{%)
`{n=}
`1973-75
`n/N
`
`0-4
`
`NS
`
`VV
`
`V/r
`
`IV
`
`I/W/r
`
`VV
`
`W/V
`
`Vv
`
`In/IV
`
`0-4.
`
`Yes
`
`Yes
`
`Yes
`
`Yes
`
`No
`
`0-3
`
`0-3
`
`0-4
`
`0-4
`
`0-3
`
`0-2
`
`0-2
`
`0-2
`
`0-2
`
`N/A
`
`N/A
`
`NS
`
`88
`
`N/A
`
`NS
`
`34
`
`NS
`
`N/A
`N/A
`105
`
`N/A
`N/A
`N/A
`
`N/A
`N/A
`N/A
`N/A
`N/A
`N/A
`9/10
`13/9
`101/23
`(overall)
`
`185
`156
`197
`6]
`116
`115
`
`92
`
`2
`4}
`47
`36
`32
`33
`27
`29
`53
`53
`23
`
`4}
`40
`77
`
`Hoeltgen?!
`
`livingstan?®
`
`Britell”
`
`Davis'®
`
`1974
`
`1976-77
`
`1976-78
`
`livingston?”
`
`1977-78
`
`Hoffman??
`
`1977-80
`
`Krook?¢
`
`Davis!
`
`1977-81
`
`1978-79
`
`Gralla*°
`
`1978-79
`
`Ruckdeschel4@
`
`1978-79
`
`Einhorn >
`
`1978-81
`
`Robert?®
`
`1978-81
`
`181
`N/A
`3.9
`0/23
`Not specified A(50)
`150
`N/A
`40
`1/17
`A(75)
`194
`N/A
`37
`2/16
`Cc
`55
`N/A
`Al
`0/6
`A(40)C
`N/A
`N/A
`37
`24
`BAC,ON
`N/A N/A
`37,
`18
`NAC,
`N/A
`N/A
`67
`1/4
`CAP
`N/A
`N/A
`57
`44
`P >AC
`9
`N/A
`48
`0/2
`Notspecified C
`2
`N/A
`3.9
`0/4
`cc,
`4
`N/A
`44
`0/2
`CCA
`N/A
`N/A
`43
`46
`0/3
`Notspecified ACh
`N/A
`N/A
`37
`0/5
`AT
`16/11
`23.
`73
`0/4
`CAMP, -L
`17/12
`10
`3.5
`o/t
`ML >AC
`37/16
`27
`79
`T/I2
`Not specified CAP
`36/17
`27
`75
`4/4
`MACC,
`46/4
`2
`3.5
`o/1
`Notspecified CAP(50)
`12
`50
`0/2
`CAP(100}
`{overall}
`23/18
`N/A
`N/A
`3/16
`VdP{60}
`25/15
`N/A
`N/A
`5/11
`VdP(120)
`53/24
`N/A
`N/A
`$1
`2/8
`~ HAM
`52/25
`N/A N/A
`44
`5/12
`CAMP,
`51
`N/A
`3105
`7A
`2/9
`Notspecified CAM
`53
`N/A
`30
`47
`57
`V/4
`C>-ASM
`185
`N/A
`21
`N/A
`6.3
`1/6
`AC
`163
`N/A
`24
`N/A
`5.1
`5/12
`CAP
`140
`N/A
`24
`N/A
`6.2
`2/16
`CAMF
`99
`67/32
`N/A
`194
`Al
`1/13
`CAE
`101
`68/33
`N/A
`Al
`2/it
`MiVb
`38
`53/21
`N/A
`8.0
`2/12
`PCVd
`36
`(overall)
`N/A
`8.0
`2/7
`PACVd
`109
`73/36
`N/A
`50
`0/18
`AFP(30)
`107
`80/27
`N/A
`54
`1/23
`CAP{40)
`112
`81/31
`N/A
`5.3
`4/18
`CBP{30)
`104
`82/22
`N/A
`54
`5/22
`MiVbP{40}
`147
`N/A
`13
`46
`4/29
`FOMi
`156
`N/A
`20
`5.3
`5/23
`FOMi/CAP
`149
`N/A
`13
`55
`1/21
`CAP
`48
`29/19
`N/A N/A
`8.5
`2/6
`VdP
`49
`36/13
`N/A
`12.3
`3/7
`VbP
`67
`46/21
`33.
`67
`3/14
`EP(60-80)
`62
`46/16
`28
`67
`3/15
`VdP(120)
`62
`38/24
`36
`64
`W/V
`VdEP(50-80}
`115
`79/36
`N/A
`58
`1/19
`CAMP,
`121
`90/31
`N/A
`5.1
`6/31
`MiVbP
`126
`85/41
`N/A
`6.0
`6/26
`YdP(120)
`124
`87/37
`N/A
`6.1
`2/23
`EP(60)
`92
`66/26
`17
`Ag
`0/6
`MACC
`94
`56/38
`13
`56
`0/8
`FAM
`42
`N/A
`14
`Al
`0/6
`Vd
`4]
`N/A
`5
`6.0
`o/11
`VdP(120)
`Al
`N/A
`7
`39
`0/8
`VdP(60)Mi
`133
`97/36
`133
`N/A
`5.0
`1/12
`FOMi/CAP
`135
`90/45
`134
`N/A
`53
`3/19
`EP
`136
`
`105/31
`PE MBGB
`7/38
`49
`N/A
`135
`
`N/A
`N/A
`N/A
`440
`
`Yes
`
`No
`
`Yes
`
`Yes
`
`0-2
`
`0-2
`
`0-2
`
`0-3
`
`Yes
`
`WAY
`
`N W
`
`/V
`
`W/V
`
`W/V
`
`0-2
`
`W/N/r
`
`Yes
`
`Yes
`
`V/r
`
`0-2
`
`No
`
`W/V/r
`
`II/IV
`
`Yes
`
`Yes
`
`Yes
`
`0-2
`
`0-3
`
`Ruckdeschel*!
`
`1979-79
`
`Kelsen?4
`
`1979-80
`
`Ruckdeschel4?
`
`1979-81
`
`Miller®4
`
`1980-83
`
`Kris?5
`
`Dhiangra’?
`
`1981-82
`
`1981-83
`
`Ruckdeschel“#
`
`1981-83
`
`Krook?7
`
`1981-84
`
`Einhorn '¢
`
`1982-84
`
`Weick**
`
`1982-84
`
`61
`
`408
`
`399
`
`146
`
`111
`overall)
`
`NS
`
`NS
`
`NS
`
`04
`
`14
`176
`
`N/A
`
`NS
`
`73
`
`65
`
`61
`
`Downloaded from jco.ascopubs.org on October 24, 2010. For personaluse only. No other uses without permission.
`1 American Society of Clinical Oncology. Alrights reserved.
`Copyright © 2
`
`

`

`PHASE lil TRIALS FOR ADVANCED NON-SMALL-CELL LUNG CANCER
`
`1737
`
`Niell?°
`
`Rapp?”
`
`1983-85
`
`1983-86
`
`Yes
`
`No
`
`Bonomi®
`
`1984-85
`
`0-2
`
`No
`
`Veeder*4
`
`1985-90
`
`IN/IV/r
`
`(0-3
`
`
`Table 1.
`(Cont'd)
`No.of
`No.of
`I-Year
`Median
`No.of
`Patients Sex Male/
`Study Survival=Survival Deathsper Female Disease Brain Responders
`
`
`
`
`
`
`
`First Author(ref)
`Years
`Arm
`{no.)
`Stage
`PS
`Metastcises
`Regimens
`OR/PR
`{months)
`{%)
`{n=}
`PVb
`142,
`«111/31
`4/30
`59
`N/A
`141
`PVbMi
`134,
`99/35
`5/17
`5.90
`N/A
`133
`MiMVbP
`52
`N/A IW/N/r 0-2
`0/13
`6.3
`16
`94 >
`Mi > Vb > M - BSC
`53
`N/A
`0/10
`5.4
`6
`{overall}
`BSC
`50
`38/12 tIb/N 0-2
`0/0
`3.9
`N/A N/A
`CAP
`92
`71/21
`0/13
`5.2
`20
`N/A
`VdP
`91
`69/22
`1/21
`79
`20
`N/A
`MiVbP
`176
`125/51 W/r
`1/35
`5.2
`N/A
`168
`| — MiVbP (on progression}
`88
`62/26
`1/4
`- 6.0
`N/A
`88
`Pb — MiVbP{on progression)
`88
`59/29
`0/8
`7.3
`N/A
`87
`Vb iv D.1,2/P days 1-3 bolus.
`175
`115/60
`2/21
`5.8
`N/A
`168
`
`172=«428/44 MiVbP + CAMP 4/18 57 N/A 165
`
`
`
`
`Luedke??
`1984-86
`128
`N/A Ill/IV
`Q-2
`Yes
`Vd
`0/1
`3.4
`13
`N/A
`06
`122
`N/A
`VdMi
`1/32
`47
`13
`N/A
`125
`N/A
`VdP
`2/22
`5.7
`20
`N/A
`64
`51/13,
`Mi
`0/17
`3.7
`82
`N/A
`69
`«55/14
`MivP
`3/19
`5A
`79
`N/A
`54
`32/22 WI/IV/r 0-2
`EP(30) iv days 1-3 bolus
`O/11
`49
`17
`N/A
`54
`33/21
`E days 1-3/P(45} D.2-3 (civi}
`0/13
`5.2
`26
`=N/A
`105
`70/35 W
`P(50)
`0/13
`69
`21
`103
`108
`88/20
`P(100}
`3/12
`5.3
`19
`103
`110
`88/22
`P(100)Mi
`4/25
`7.2
`21
`107
`118
`78/40 Nib/N 0-2
`EP-X
`2/18
`8.0
`26
`88
`119
`82/37
`EP-H
`1/21
`5.8
`26
`93
`68
`48/20 N/r
`F
`0/2
`5.1
`1%
`N/A
`143°
`102/41
`Vn
`0/17
`69
`25
`N/A
`193
`127/66 IIlb/IV
`EP
`ORR 12%
`76
`32
`190
`118/72
`Pafi35) x 24h/P
`ORR 27%
`95
`37
`191
`120/71
`Pa(225} x 24h/P 1 G-CSF
`ORR 32%
`10.1
`40
`Wozniak’”
`1993-95
`209
`139/70 IIlb/V 0-1
`P
`0/25
`6.0
`20
`No
`187
`002
`
`206 170 140/66 PVn 4/5050 8.0 36
`
`
`
`
`
`NOTE.Patients had either unresectable locally-advanced disease, stage !V disease, or recurrent disease.
`Abbreviations: PS, performancestatus; CR, complete response; PR, partial response; N/A, data not available; NS, notstatistically significant, iv, intravenously;
`PO,orally; MTH/YR, month/year; M/F, male/female; A, doxorubicin; B, bleomycin; C, cyclophosphamide; C,, CCNU (lomustine); Ch, chlorambucil; E, etoposide;
`F, fluorouracil; G-CSF, granulocyte colony stimulating factor; H, hydrazine; HEM, hexymethylmelomine;|, iproplatin; L, leucovorin; M, methotrexate; Mi, mitomycin
`C; MBGB, methylglyoxal bisguanylhydrazone; N, mechlorethamine; O,vincristine; P, dsplatin; Pa, paclitaxel; Pb, carboplatin; P,, procarbazine; V, vinblastine;
`Vb, vinblastine: Vd, vindesine; Vn, vinorelbine; T, thiabendazole; X, placebo; BSC, best supportive care; —>,followed by; r, recurrent.
`
`Goldberg!”
`
`1987-88
`
`Gandara!®
`
`1988-90
`
`loprinzi®°
`
`Crawford?
`
`1990-92
`.
`1990-92
`
`Bonomi?
`
`1993-94
`
`0-2
`
`0-1
`
`0-1
`
`Yes
`
`Yes
`
`No
`
`No
`
`No
`
`No
`
`01
`
`09
`
`09
`
`1
`
`53
`
`14
`
`03
`
`05
`
`529
`
`trials increased from 77 (time period, 1973 to 1983) to 121
`(time period, 1984 to 1994) (P < .001). Twenty-three trials
`(70%) included a platinum-based regimen in at least one
`arm of the trial. Eight of the nine (89%) trials in the
`1984-to-1994 time period included a platinum-based regi-
`men, compared with 15 of the 24 (63%) im the 1973-to-1983
`time period.
`The overall response rate with chemotherapy ranged from
`1% (vindesine alone) to 46% (vindesine and cisplatin), with
`a median of 17% overall therapeutic regimens.”°*? ‘Ihree
`trials compared the response rate of low dose versus higher
`dose cisplatin. No significant difference was detected in the
`overall response rates betweenpatients treated with low (50
`
`mg/m? to 60 mg/m?) or high doses (100 mg/m’ to 120
`mg/m”) of cisplatin.'''8° The overall response rates in the
`128 patients treated on the two trials with single-agent
`cisplatin at 50 mg/m? were 4% (n = 1/23) and 12% (n =
`13/105) compared with 7% (nm = 2/27) and 14% (m =
`15/108) in the 135 patients treated with single-agent cispla-
`tin at 100 mg/m?, respectively.'!-'® The third trial compared
`cisplatin at two dose levels (60 mg/m? v 120 mg/m’) in
`association with a fixed dose of vindesine (3 mg/m?). The
`overall response rate in the 4] patients who received the
`60-mg/m? dose ofcisplatin and vindesinc was 46% com-
`pared with 40% in the 40 patients who received a 120-
`mg/m? dose of cisplatin plus vindesine.”°
`
`Downloaded from jco.ascopubs.org on October 24, 2010. For personal use ony. No other uses without permission.
`r
`ights reserved.
`Copyright © 2001 American Society of Clinical Oncology. All
`
`

`

`
`
`Year
`
`_
`
`
`
`
`
`regimen and the year of study initiation (more recent years)
`were significantly related to median survival (P = .0318 and
`P = 0214, respectively).
`Five (15%) of the 33 phase III trials showeda statistically
`significant difference im survival time, andall were in favor of
`the patient cohort
`that received the experimental
`therapy
`comparedwith the control group (median, 2 months; range, 0.9
`to 2.7).5°343747 These five studies**477” included a signif-
`icantly larger number of patients per treatment arm (median,
`149 patients per arm; range, 50 to 209) than did the studies that
`did not show a significant difference in median survival time
`(median, 88 patients per arm; range, 19 to 197; Wilcoxontest
`P = .026). Four of the five trials mvolved platinum-based
`regimens. Wozniak et al compared cisplatin versus cisplatin
`and vinorelbine (median survival, 6 v 8 months, respective-
`Fig 1. Median survivaltimeofall patients treated on the phaseIll trials.
`ly).4” Bonomi et al compared etoposide and cisplatin to
`Each study is represented byacircle, the size of each circle is proportional
`24-hour infusional paclitaxel at two dose levels in association
`to the sample size of the treated patients (univariate analysis, P < .0001).
`with bolus cisplatin (median survival, 7.6 v 9.9 months,
`respectively).* Crawford el al comparedfluorouracil andleu-
`covorin versus vinorelbine (median survival, 5.1 v 6.9 months,
`respectively).° Rapp et al compared cyclophosphamide, doxo-
`rubicin, and cisplatin (CAP) versus vindesine and cisplatin
`(median survival, 5.2 v 7.9, respectively).°” The study also
`included an arm that evaluated patients treated with best
`supportive care alone (median survival, 3.9 months), which
`was not includedin this current analysis; however, for the sake
`of completeness, the results of the best supportive care arm are
`included in Table 1. Miller et al compared fluorouracil,
`vincristine, and nutomycin C (FOMi) versus CAP versus
`alternatmg FOMi/CAP (median survival, 4.6 v 5.5 v 53
`months, respectively)** Nineteen of the other 28 trials that
`showed nosignificant difference in survivalincluded platinum-
`based regimens.
`
`1738
`
`BREATHNACHET AL
`
`41
`
`©
`
`MedianSurvival{months} ~
`
`Survival Time of Patienis
`
`Eighteen trials included data on the numberof patients
`who haddied at the time of analysis (median percentage of
`patients deceased per trial at time of publication, 90%;
`range, 66% to 99%). ‘the median of the median survival
`times of all paticnts treated’ on phascIII trials initiated
`between 1973 to 1983 and between 1984 to 1994 were 5.2
`months (range, 3.5 to 12.3 months) and 5.8 months(range,
`3.4 to 10.1 months), respectively (P < .001; Fig 1). The
`median of median survivals of patients treated with plati-
`num-based regimens (n = 5,317) compared with those
`treated with nonplatinum-containing regimens (n = 2,976)
`during the 33 tials were 5.8 months and 4.6 months,
`respectively (P < .01). The median of median survivals in
`patients treated on trials initiated between 1973 and 1983
`were 5.6 months and 4.5 months for patients treated with a
`platinum-based regimens and nonplatinum-containing regi-
`mens, respectively (P < .001). The median of median
`survivals in patients treated on trials initiated between 1984
`and 1994 were 6.0 months and 4.7 months for patients
`treated with a platinum-based regimens and nonplatinum-
`containing regimens, respectively (P < .037). Platinum-
`containing regimens were not introducedinto phaseIIItrials
`until 1976.’ Analysis of the data following exclusion of the
`2 trials initiated before 1975 did not impact on the time
`trend for improved survival over the subsequenttime period
`(P < .0001; regression analysis).
`A multivariate regression analysis that included the year
`ofstudy initiation, platinum-bascd regimen indicator, trcat-
`ment group indicator, brain mctastascs indicator, and the
`maximum performance status showedthat platinum-based
`
`SEER Database
`
`Information on the median survival times of patients with
`distant non-small-cell lung canccris available in the SEER
`population database. SEER survival data is recorded from
`the time of diagnosis, rather than initiation of therapy under
`review, and therefore the data on median survival is not
`directly comparable to the survival data from the phase III
`trials. In the SEER database, the median survival time of
`patients with distant non-small-cell lung cancer increased
`from a median of 6.9 months between 1973 and 1974 to 7.3
`monthsfor those treated between 1993 and 1994 (P = .001;
`Fig 2). There was also a significant
`increase in 3-year
`survival from 2.3% to 3.6% between 1973 through 1974 and
`1993 through 1994, respectively (P value < .0009).
`lung
`The survival of all patients with non-small-ccll
`cancer in the SEER database was also examined to deter-
`mine potential change. There wasa statistically significant
`
`Downloaded from jco.ascopubs.org on October 24, 2010: For personaluse onl
`No other uses without permission.
`Copyright © 2001 American Society of Clinicat Oncology. Alrights reserved.
`
`

`

`(months)
`
`
`
`MedianSurvival
`
`73 7 77
`
`#79
`
`81
`
`85
`88
`Year
`
`87
`
`88
`
`91
`
`93
`
`95
`
`PHASElil TRIALS FOR ADVANCED NON-SMALL-CELL LUNG CANCER
`
`1739
`
`were only included in phaseIIItrials starting after 1994 and
`so are not included in this analysis.
`Of the 33 phase ID trials evaluated, only 5 (15%)trials
`showed a statistically significant prolongation in surviv-
`al.>°343747 Tn two of these five trials the control arm
`consisted of single-agent therapy. The choice of the control
`arm of fluorouracil in the trial by Crawford et al and the
`choice of a cisplatin-alone control arm in the trial by
`Wozniak ct al might be perecived as attempts to maximize
`the potential for statistically significant differences in sur-
`vival in these respective trials.**” Data from a meta-analysis
`(hal comparedsingle agent versus combination chemother-
`apy in patients with advanced-stage NSCLC reported in-
`creased objective response rates and toxicity with combina-
`tion therapy, however, when a platinum analog or
`vinorelbine was used as a single agent (10 trials), survival
`was not significantly increased with combination therapy.*®
`More recent phase III trials that compare the single-agent
`therapy of cisplatin with a combination therapy of cisplatin
`plus new agents, such as vinorelbine or gemcitabine, have
`favored the combination regimens, with statistically signif-
`icant prolongation of survival being reported.4”“? The
`therapeutic potential of these newer agents in isolation
`continues to be evaluated.
`The analyses of the phaseIII therapeutic trials for patients
`with advanced lung cancer show some similarities as well as
`some differences between patients with NSCLC (8,436
`patients on 33 trials) and those with SCLC (5,746 patients
`on 21 trials).* Fifteen percent of the phase II] trials in
`patients with advanced-stage NSCLC hadstatistically sig-
`nificant differences in survival compared with 5 (24%)of 21
`phaseIII trials in patients with extensive-stage SCLC during
`a similar time period.* The difference in median survival
`between patients treated with the control and experimental
`arm was 2 months or longer in only 11% of the phase III
`trials for both patients with NSCLC (8 of 33 trials) and
`SCLC (2 of 21 trials). Only one phase II] trial 3%) in
`paticnts with advanecd-stage NSCLC demonstrated a me-
`dian survival in excess of 11 months,?° compared with four
`(19%). of the 21 trials in patients with extensive-stage
`SCLC.‘ This finding mayrelate to the fact that SCLCis a
`more chemoresponsive disease, and that prolongation of
`
`Fig 2. The median survival of patients with advanced-stage non-small-
`cell lung cancer as recorded in the SEER database. Each dot represents the
`median survival for a 2-year period beginning with years 1973 and 1974"
`{linear regression, P = .001).
`
`linear trend in 5-year survival of all patients with non—
`small-cell lung cancer evaluated over 2-year intervals in
`1973 to 1974 through 1991 to 1992 (P = .012). The median
`survival
`time for all patients with non-small-cell
`lung
`cancer in the SEER database incrcascd from 9.1 months in
`the 1973-to-1974 time period to 10 months in the 1993-to-
`1994 time period, respectively. The corresponding 3-year
`and 5-year survival data increased by 2% to 4% (P < .0001)
`(Table 2).
`,
`
`DISCUSSION
`
`AnalysesofphaseIIItrials of systemic chemotherapy for
`advanced-stage non-small-cell lung cancer during the past
`25 years show perceptible ‘but modest advances in patient
`outcome. ‘he increase in median survival in the phase III
`trials from the early time period (1973 to 1983)to the latter
`period (1984 to 1994) was 2.6 weeks. This closely matches
`the increase in median survival of 2 weeks in the SEER data
`overa similar period. Duringthe late 1970s there was a shift
`from alkylating agent-based therapy to platinum-based
`regimens, with more recently available agents such as
`vinorelbine and paclitaxel includedin trials only since 1990.
`However, other agents such as gemcitabine or docetaxel
`
`Table 2. Overall SurvivalofAll Patients With Non-Small Cell Lung Cancer in the SEER Database.
`1972-1973
`199T-F992
`1993-1994
`%
`SE
`%
`SE
`S
`SE
`
`3-year survival (SE)
`14.4
`0.322
`18.5
`0.269
`18.1
`0.279
`< .0001
`
`5-year survival (SE) < 0001 10.6 0.281 127 0.239 n/a n/a
`
`
`
`
`
`
`
`Abbreviation: n/a, not available.
`
`P
`
`Downloaded from jco.ascopubs.org on October 24, 2010. For personal use only. No other uses without permission.
`Mri
`Copyright © 2001. American Society of Clinical Oncology. Al
`ights reserved.
`
`

`

`1740
`
`BREATHNACHET AL
`
`survivalis, therefore, more likely in its patients compared to
`those with NSCLC.
`
`The number ofphase IT trials for paticnts with advanced-
`stage NSCLC in North America initiated between 1973 and
`1994 decreased nearly three-fold over the latter half of the
`period assessed. In patients with extensive-stage SCLC a
`similar decrease in the number of phase IIf trials performed
`was not noted, with 11 trials that included 2,578 patients
`initiated from 1972 to 1981 and another 10 trials with 3,168
`patients from 1982 to 1990.” Potential explanations for the
`decrease in the number of phase III trials in patients with
`NSCLC may be the lack of innovative treatments intro-
`duced during this time period, and/or a declining interest on
`the part of physicians in the design and initiation ofclinical
`trials in these patients. Despite the decrease in the number of
`phase IIItrials, the average numberofpatients pertrial rose
`from 223 (5,359 patients on 24 trials) to 342 (3,075 patients
`on 9 trials) between the two time periods. Women with
`advanced-stage NSCLC represented ~40% of persons en-
`tered on the phase III trials. This percentage remained
`constant between the periods 1973 to 1983 and 1984 to
`1994. In the SEER database, women represent 34.5% of
`patients diagnosed with advanced-stage non-small-cell lung
`cancer between 1973 and 1994, with an increase in inci-
`dence from 29% to 37% for the time periods 1973 to 1983
`and 1984 to 1994, respectively. This indicates a propor-
`tional representation of womenin phaseIII trials of chemo-
`therapy in persons with advanced-stage NSCLC.
`A comparison of the median survival time of patients
`with advanced-stage NSCT.C in the SEER database showsa
`prolongation of median survival times ofslightlyless than 2
`weeks in the group treated between 1973 and 1974 (6.9
`months) compared with the group treated between 1993 and
`1994 (7.3 months). This result is significantly less than the
`=2-month prolongation in survival of patients with exten-
`sive-stage SCLC, as recorded within the SEER database
`from 1973 to 1994.* This difference in prolongation of
`median survival between patients with advanced-stage
`NSCLC and extensive-stage SCLC probably reflects the
`greater chemosensitivity of SCLC, but
`it also serves to
`emphasize the great need to develop better therapeutic
`strategies in both forms of lung cancer. A statistical model
`has been developed to aid clinical
`investigators in the
`selection of precisely which pilot regimens to bring forward
`to phaseIII trials in patients with extensive-stage SCLC.*°
`This model may also have implications on the selection of
`promising agents in the treatment of NSCLC.
`‘The quality of the trials performed during the time period
`under analysis also nccds to be considered. Given that we
`propose that phase III trials should include at least 200
`patients per treatment arm to detecta statistically sigmificant
`
`differences in survival of 2 months between control and
`
`experimental arms, only one ofthe 33 trials that we included
`was actually cligible.4”7 Comorbid conditions of included
`patients are not
`included in the descriptions of patient
`characteristics in the individual
`trials,
`though exclusion
`criteria are fully detailed inall the 33 trials included. Onthe
`other band,
`these eligibility criteria have become more
`strict, particularly since 1990, with the inchision of only
`patients with a performancestatus of 0-1 and no evidence of
`brain metastases. Patients with performancestatus ofeither
`0-3 (4 trials)”!*"!?6 or 0-4 (4 trials)?)???9 were included
`on trials initiated before 1978. Of the patients included on
`these trials, 29% to 63% had a performance status of 0-1]
`(median, 39%). In comparison, from 1990 to 1994,three of
`the four trials performed inclhided only patients with a
`performance status of 0-1, with all three