Multi-Institutional Randomized Phase II Trial of Gefitinib for
`
`Previously Treated Patients With Advanced Non—Small-Cell
`Lung Cancer
`
`By Masahiro Fukuoka, Seiii Yano, Giuseppe Giaccone, Tomohide Tamura, Kazuhiko Nakagawo, Jean-Yves Douillard,
`Yutaka Nishiwaki, Johan Vansteenkiste, Shinzoh Kudoh, Danny Rischin, Richard Eek, Takeshi Horai, Kazumasa Noda,
`lchiro Takata, Egbert Smit, Steven Averbuch, Angela Macleod, Andrea Feyereislova, Rui-Ping Dong, and José Baselga
`
`Purpose: To evaluate the efficacy and tolerability of two
`doses of gefitinib (lressa [ZD1839]; AstraZeneca, Wilming-
`ton, DE), a novel epidermal growth factor receptor tyrosine
`kinase inhibitor, in patients with pretreated advanced non-
`small—cell lung cancer (NSCLC).
`Patients and Methods: This was a randomized, double-
`blind, parallel-group, multicenter phase II trial. Two hun-
`dred ten patients with advanced NSCLC who were previ-
`ously treated with one or two chemotherapy regimens (at
`least one containing platinum) were randomized to receive
`either 250-mg or 500-mg oral doses of gefitinib once daily.
`Results: Efficacy was similar for the 250- and 500-mg/d
`groups. Obiective tumor response rates were 18.4% (95%
`confidence interval [Cl], 11.5 to 27.3) and 19.0% (95% CI,
`12.1 to 27.9); among evaluable patients, symptom im-
`provement rates were 40.3% (95% CI, 28.5 to 53.0) and
`37.0% (95% CI, 26.0 to 49.1); median progression-free
`survival times were 2.7 and 2.8 months; and median over-
`
`all survival times were 7.6 and 8.0 months, respectively.
`Symptom improvements were recorded for 69.2% (250
`mg/d) and 85.7% (500 mg/d) of patients with a tumor
`response. Adverse events (AE5) at both dose levels were
`generally mild (grade 1 or 2) and consisted mainly of skin
`reactions and diarrhea. Drug-related toxicities were more
`frequent in the higher-dose group. Withdrawal due to drug-
`related AEs was 1.9% and 9.4% for patients receiving
`gefitinib 250 and 500 mg/d, respectively.
`Conclusion: Gefitinib showed clinically meaningful anti-
`tumor activity and provided symptom relief as second- and
`third-line treatment in these patients. At 250 mg/d, gefitinib
`had a favorable A! profile. Gefitinib 250 mg/d is an impor-
`tant, novel treatment option for patients with pretreated
`advanced NSCLC.
`
`J Clin Oncol 21:2237-2246. © 2003 by American
`Society of Clinical Oncology.
`
`UNG CANCER is the most common cause of cancer deaths
`
`in both men and women worldwide.1 Despite advances in
`treatment, such as combination chemotherapy and chemoradiation,
`survival has improved very little over the past few decades.2 A
`meta—analysis demonstrated that
`the median survival
`time for
`patients with advanced disease receiving cisplatin—based chemother-
`apy is around 6 months.3 The 5-year survival rate for all stages is
`less than 15%.4 Prognosis is particularly poor for patients who have
`progressive disease following chemotherapy; for non—small—cell
`lung cancer (NSCLC) patients receiving best supportive care (BSC)
`after 1 or more prior chemotherapy regimen, median survival time
`is just 16 weeks, with a 1-year survival rate of 16%.5
`
`Recently, it has become generally accepted that systemic
`chemotherapy is beneficial in terms of improved survival and
`quality of life (QoL) in those with advanced NSCLC.3’6 As
`more patients receive first—line chemotherapy, the need for
`effective second—line therapy is increasing. Currently, do-
`cetaxel, having demonstrated survival benefits over BSC, is
`the only approved treatment in the United States and the
`European Union for patients who have been failed by previ-
`ous platinum—based chemotherapy.7
`Patients with late-stage NSCLC are often symptomatic, with
`specific pulmonary problems (eg, cough, breathlessness, hemo—
`typsis) and general symptoms (eg, fatigue, weight loss) that can
`
`From the Kinki University School ofMedicine, Osaka City University School
`of Medicine, and AstraZeneca, Osaka, Tokushima University School ofMedi-
`cine, Tokushima, National Cancer Center, Central Hospital, and Japanese
`Foundation for Cancer Research, Tokyo, National Cancer Center, East Hospi-
`tal, Chiba, Kanagawa Cancer Center, Yokohama, and National Shikoku Cancer
`Center, Matsuyama, Japan; CR.L.C.C. Rene Gauducheau, Saint-Herblain,
`France; University Hospital Gasthuisberg, Leuven, Belgium," CentrefiJr Devel-
`opmental Cancer Therapeutics, Melbourne, Australia; Mary Potter Oncologz
`Centre, Pretoria, South Africa; Academic Hospital Free University, Amsterdam,
`the Netherlands; AstraZeneca, Wilmington, DE; AstraZeneca, Alderley Park,
`United Kingdom; and Vall d’Hebron University Hospital, Barcelona, Spain.
`Submitted October 4, 2002; accepted January 28, 2003.
`The authors wish to disclose the following conflicts of interest statement:
`José Baselga has been in receipt ofa research grantfrom AstraZeneca and
`
`honoraria to attend advisory boards and to give talks on ZDI839; Johan
`Vansteenkiste has received honoraria from AstraZeneca to attend advisory
`boards; Jean Yves Douillard has received honoraria for participating in
`advisory boards or symposia; Giuseppe Giaccone has received honoraria
`and research grants; and Danny Rischin has been in receipt of honoraria
`and travel grants from AstraZeneca. Steven Averbuch, Angela Macleod,
`Andrea Feyereislova, and Rui—Ping Dong were employed by AstraZeneca at
`the, time ofstudy completion, and as such, may hold stock in the company. All
`other authors have nothing to declare.
`Address reprintrequests to Masahiro Fukuoka, MD, Fourth Department
`of Internal Medicine, Kinki University School of Medicine, 3 77-2 Ohnohi—
`gashi Osakasayama, Osaka 589, Japan; email: mfukuoka@med. kindai.ac.jp.
`© 2003 by American Society of Clinical Oncology.
`0732-183X/03/21 12-223 7/$20. 00
`
`Journal of Clinical Oncology, Vol 21, No 12 (June 15), 2003: pp 2237-2246
`DOI: 10.1200/JCO.2003.10.038
`
` 2237
`
`Downloaded from ascopubs.org by Reprints Desk on Augu
`st 2, 2017 from 216.185.156.028
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`03' 2046
`APOTEX v. OSI
`|PR2016-01284
`
`OSI 2046
`APOTEX V. OSI
`IPR2016-01284
`
`

`

`2238
`
`cause extreme distress to the patient. Therefore, improvements in
`disease-related symptoms and QoL are the key desired outcomes
`of medical managements Effective, palliative,
`low-toxicity
`treatments for patients with advanced NSCLC are needed.
`The epidermal growth factor receptor (EGFR) is a promising
`target for anticancer therapy because it is expressed or highly
`expressed in a variety of tumors, including NSCLC.9’10 Further-
`more, high levels of EGFR expression have been associated with
`a poor prognosis in lung cancer patients in several studies.1 1'13
`EGFR—targeted cancer therapies are currently being developed;
`strategies include inhibition of the intracellular tyrosine kinase
`domain of the receptor by small molecules such as gefitinib
`(Iressa [ZD1839]; AstraZeneca, Wilmington, DE).14 Gefitinib is
`an orally active, selective EGFR tyrosine kinase inhibitor that
`blocks signal transduction pathways implicated in the prolifera—
`tion and survival of cancer cells.”’16
`
`Four phase 1 studies assessed gefitinib tolerability and phar-
`macokinetics in pretreated patients with solid tumors, including
`100 patients with heavily pretreated advanced NSCLC.17 Evi-
`dence of major tumor regression was seen in 10 patients with
`NSCLC; a number of other patients had nonprogressive disease
`lasting for 6 months or
`longer; and palliation of specific
`symptoms was also frequently observed.
`In these trials, re-
`sponses were seen across the dose range 150 to 800 mg/day,
`while the majority of dose interruptions and reductions due to
`toxicity were required in patients receiving more than 600
`mg/d. From these data, two doses (250 and 500 mg/d) were
`selected for investigation in phase II and phase III trials. The
`250 mg/d dose is higher than the lowest dose level at which
`objective tumor regression was seen, while 500 mg/d is the
`highest dose that was well
`tolerated when taken over an
`extended period in phase I trials.
`The aims of this Iressa Dose Evaluation in Advanced Lung
`Cancer (IDEAL 1) trial were to further investigate the efficacy
`and safety of oral gefitinib in patients with advanced NSCLC
`who had previously received one or two chemotherapy regimens,
`with at
`least one containing platinum. The population was
`prospectively stratified into Japanese and non-Japanese patients
`to investigate whether there were any differences between the
`two patient populations with respect to efficacy.
`
`PATIENTS AND METHODS
`
`Study Design
`This randomized, double-blind, parallel group, phase II multicenter trial
`recruited patients at 43 centers across Europe, Australia, South Africa, and
`Japan. Primary objectives were to evaluate the objective tumor response rate
`(R) for gefitinib doses of 250 and 500 mg/d and to further characterize the
`safety profile of these doses. Secondary objectives were to estimate disease-
`related symptom improvement rate, disease control rate (response + stable
`disease), progression-free survival (PFS), and overall survival (OS);
`to
`evaluate changes in QoL; and to assess any differences between Japanese and
`non-Japanese patients with respect to efficacy and safety.
`
`Patient Eligibility
`
`Eligibility criteria were histologic or cytologic confirmation of locally
`advanced or metastatic NSCLC; stage III or stage IV disease not curable with
`surgery or radiotherapy at study entry;
`recurrent or refractory disease
`
`47
`
`FUKUOKA ET AL
`
`following one or two previous chemotherapy regimens (at least one contain-
`ing platinum); at least one bidimensionally measurable or radiographically
`assessable lesion; age of 18 years or older; World Health Organization
`performance status (PS) of 0 to 2; and life expectancy of 12 weeks or longer.
`Patients with stable brain metastases were eligible. Exclusion criteria were
`more than two previous chemotherapy regimens, systemic anticancer therapy
`within 21 days, or radiotherapy within 14 days before the start of treatment;
`unresolved chronic toxicity higher than the National Cancer Institute
`common toxicity criteria (NCI-CTC, version 2) grade (G) of 2 (excluding
`cases of alopecia); ALT or AST levels greater than 2.5 times the upper limit
`of reference range (ULRR; more than 5 times the ULRR in the presence of
`liver metastases); serum creatinine levels greater than 1.5 times the ULRR;
`serum bilirubin levels greater than 1.25 times the ULRR; and neutrophils less
`than 1.5X109/L or platelets less than 75X109/L. Patients gave informed
`consent, and trial document approval was obtained from the ethics committee
`or institutional review board at each trial center. The study was performed in
`accordance with the Declaration of Helsinki and Good Clinical Practice
`guidelines.
`
`Treatment
`
`Patients were randomly assigned to receive double-blind gefitinib doses at
`250 or 500 mg/d. Tablets were administered once daily, except on day 1
`when patients received two doses approximately 12 hours apart. Patients
`continued uninterrupted treatment until disease progression,
`intolerable
`toxicity, withdrawal of consent, or trial closure (4 months after the last
`patient was recruited). Patients without progression were permitted to
`continue gefitinib treatment in a further study.
`One dose reduction per patient was permitted in the event of unacceptable
`toxicity. New blinded treatment supplies, decreasing the dose from 500 mg
`to 250 mg or from 250 mg to 100 mg, were dispensed. Gefitinib adminis-
`tration could be interrupted for a maximum of 14 days.
`No systemic anticancer treatment was permitted during the trial, except for
`palliative radiotherapy in patients with isolated symptomatic bone metasta-
`ses, and as long as trial drug administration was not interrupted for longer
`than 14 days.
`
`Eflicacy
`
`We assessed objective tumor response as complete response (CR), partial
`response (PR), partial response in nonmeasurable disease (PRNM), stable
`disease (SD), or progressive disease (PD) in accordance with the Southwest
`Oncology Group modification of Union Internationale Centre le Cancer/
`WHO criteria.18 Baseline assessments were performed within 14 days before
`randomization. After the start of treatment, assessments were performed
`every 4 weeks,
`then every 8 weeks following the fourth month. An
`independent response evaluation committee consisting of three radiolo-
`gists/oncologists at each session reviewed images of patients with CR,
`PR, and SD; reviewers were blinded to the investigators’ assessment and
`dose of gefitinib. Duration of response was defined as the time from the
`first objective assessment of CR or PR to the first instance of progression
`or death.
`‘
`
`Disease Control
`
`Disease control was defined as the best tumor response of CR, PR, or SD
`that was confirmed and sustained for 4 weeks or longer.
`Disease-related symptom improvement was measured using the Lung
`Cancer Subscale (LCS), a validated Subscale of the QoL instrument, the
`Functional Assessment of Cancer Therapy — Lung (FACT-L) questionnaire
`(Fig 1).” Patients completed a weekly diary card rating the severity of each
`of the following seven LCS items on a scale of 0 to 4: shortness of breath,
`weight loss, lack of clear thinking, coughing, loss of appetite, tightness in the
`chest, and difficulty breathing. On day 28, the LCS was completed as part of
`the entire FACT-L questionnaire. The maximum (asymptomatic) attainable
`score was 28. Patients with a baseline LCS score of 24 or lower were
`evaluable for symptom improvement. This information was used to deter-
`mine symptom improvement rate,
`time to symptom improvement, and
`
`Downloaded from ascopubs.org by Reprints Desk on August 2, 2017 from 216.185.156.028
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`

`GEFlTlNIB IN NSCLC: THE |DEAL l TRlAL
`
`2239
`
`
` 1. Lung Cancer
`
`Subscale
`
`5. Social
`
`well-being
`
`2. Physical
`well-being
`
`(LCS)
`
`
`well—being
`
`
`4. Emotional
`
`well—being
`
`f3. Functional _.
`
`Fig 1. The five components of the Functional Assessment of Cancer Therapy -
`Lung (FACT-L). Component 1
`is measured by the Lung Cancer Subscale itself;
`components 1 through 3, by the Trial Outcome Index; and components 1 through
`5,byFACFL
`
`duration of symptom improvement. Based on data showing that a 2-point
`change in LCS score is clinically meaningful for patients and is significantly
`associated with Eastern Cooperative Oncology Group performance status,
`weight loss, objective tumor response, and time to progression,20 symptom
`improvement was prospectively defined as a 2-point (or greater) improve-
`ment in LCS score sustained for 4 weeks or longer, with no worsening at any
`interim weekly time points. Duration of symptom improvement was defined
`as the interval between the first visit presenting with symptom improvement
`and a subsequent visit at which symptoms had worsened. Missing data points
`were counted as no change in symptoms.
`
`QoL Assessment
`
`Patients completed the FACT-L questionnaire to assess QoL. The
`FACT-L assessment has been validated with respect to its psychometric
`properties and sensitivity to clinical changes.19 FACT-L was completed at
`baseline and then every 28 days after the start oftreatment. The questionnaire
`was administered before clinical assessment and before patients heard news
`about their disease status. The Trial Outcome Index (T01) of FACT-L (Fig
`1) measures the more physical aspects of patient QoL that are shown to be
`sensitive to chemotherapy.19 T01 and FACT-L scores were derived in a
`similar manner to the LCS scores; the highest scores attainable for T01 and
`FACT-L were 84 and 136, respectively. T01 and FACT-L responses were
`prospectively defined as a 6-point (or greater) improvement (for 4 weeks or
`longer), a change that has been shown to be clinically meaningfirl.20
`
`PFS and 0S
`
`PFS was defined as the period from the randomization date to the date
`when disease progression (or death) was observed. OS was defined as the
`period from the randomization date to the date of death. Patients alive at data
`cutoff were censored at the last date the patient was known to be alive.
`
`Safety and Tolerabz’lity
`All adverse events (AEs) were reported, and severity was assessed by the
`NCI-CTC (version 2.0) grading system. Data were collected on therapy
`interruptions and withdrawals due to AEs. Routine clinical and laboratory
`assessments were performed. ECGs and complete ophthalmic evaluations,
`including slitlamp examination, were performed at baseline, at 4 months, and
`on completion of or withdrawal from the trial.
`
`Statistical Methods
`
`Patients were randomized to receive oral gefitinib at doses of 250 or 500
`mg/d, and were stratified by ethnicity as Japanese and non-Japanese.
`Randomization and allocation were performed by a centralized registration g
`or randomization center using dynamic balancing21 with factors for country
`and WHO-PS of 0 to 1 versus 2. Patients were categorized at randomization
`with respect to prior taxane therapy (docetaxel : paclitaxel v paclitaxel alone
`v no taxane) and number of prior regimens (one v two).
`The target sample population of200 patients (100 in each dose group and 100
`in each ethnic group) was chosen to enable the tumor lower limit for RR to be
`independently evaluated in the four strata defined by dose and ethnicity. Within
`each stratum, the goal was to have 90% power for a two-sided 5% significance
`test to show that the R was greater than 5% assuming that the actual R was
`20%, which required 45 or more evaluable patients per stratum.
`RRs and disease control rates were compared between strata using Fisher’s
`exact test. Logistic regression models were used to further explore observed
`differences and to identify baseline factors that may independently predict
`for tumor response and disease control. PFS and OS were compared between
`strata using the log-rank test. Further analyses were conducted on these data
`using Cox’s proportional hazard modeling.
`
`RESULTS
`
`Patients
`
`A total of 210 patients were randomized within 4 months
`(October 2, 2000 to January 30, 2001). Of these, 208 patients
`were evaluable for efficacy, and 209 patients were evaluable for
`safety (Fig 2). The two dose groups were well balanced for most
`baseline demographic factors, with the exception of sex (Table
`1). As planned, approximately half of the patients randomized
`were Japanese. There were some demographic imbalances between
`the Japanese and non-Japanese populations (62.7% v 77.8% male;
`20.6% v 15.7% PS of0; 8.8% v 16.7% PS of2; and 76.5% v 50.0%
`adenocarcinoma, respectively).
`
`Efifcacy
`
`The investigator assessments of the best overall tumor responses
`are shown in Table 2. RR was 18.4% for the 250-mg/d group, which
`was not statistically different from that of the 500-mg/d group (RR,
`19.0%; Table 2). The independent response evaluation committee
`reviewed 107 of the 110 patients whom the investigator considered
`to have CR, PR, PRNM or SD. These included 38 of the 39
`responders. There was a high concordance in tumor response
`evaluation between investigators and independent
`reviewers
`(73.8%; Table 3). In addition, the response evaluation committee
`evaluated an additional 25 patients who were assessed by the
`investigators as having a best response of PD. Of these 25 patients,
`the response evaluation committee considered 7 patients to have had
`a best response of SD.
`Of the patients who responded, most showed rapid tumor
`regression, with 68% meeting the criteria for objective response
`by the first postbaseline assessment. The remaining patients met
`the criteria in the second,
`third, or fourth month following
`randomization. Furthermore, across both doses, most responders
`(87.2%) still had a response at the data cutoff, with a median
`follow—up of 6.3 months (range, 4.0—7.9 months). For patients
`who responded, median duration of response was more than 3
`months (ranges: 250-mg/d group, 1—5 months; 500-mg/d group,
`1—5.5 months). RRs were similar
`irrespective of whether
`
`Downloaded from ascopubs.org by Reprints Desk on August 2, 2017 from 216.185.156.028
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`

`2240
`
`(.7
`
`FUKUOKA ET AL
`
`Randomized (n = 210)
`1 patient did not receive treatment (250 mg/day)
`
`A
`
`tntention-to-treat/evaluable for safety p0 ulation (n = 209)
`
`500 mg/day in = 106)
`250 mg/day (n = 103)
`non-Japanese (n = 55), Japanese (n = 51)
`non-Japanese (n = 52), Japanese (n = 51)
`
`1 patient excluded due to protocol violation' (500 mg/day)
`
`B
`
`Per-protocoI/evaluable for response p0 ulation (n = 208)
`
`250 mg/day (n = 103)
`500 mg/day (n = 105)
`non-Japanese (n = 54), Japanese (n = 51)
`non-Japanese (n = 52), Japanese (n = 51)
`
`C
`
`Evaluabie for symptom improvement
`
`(n = 140)
`
`250 mg/day (n = 67)
`non-Japanese (n = 34), Japanese (n = 33)
`
`500 mg/day (n = 73)
`non-Japanese (n = 36), Japanese (n = 37)
`
`Fig 2. Number of patients included in the analysis populations. (A) Patients who received 1 or more doses of trial treatment. (5) Patients who received 14 or more
`days of trials treatment in each 28-day treatment period before the first tumor assessment recorded their best tumor response. (C) Patients with 0 Lung Cancer Subscale
`score of 24 or lower. Asterisk indicates that this patient’s last dose of systemic anticancer therapy was received within 21 days prior to the start of trial treatment.
`
`gefitinib was used as second-line (17.5%, 250 mg/d; 18.3%,
`500 mg/d) or third-line treatment (19.6%, 250 mg/d; 20.0%,
`500 mg/d). A post hoc nonrandomized analysis showed that
`RRs for the subgroup of patients who had previously received
`a platinum and a taxane were 24.0% at 250 mg/d and 28.0%
`
`at 500 mg/d. Similarly, RRs for patients previously given
`platinum and docetaxel were 24.0% at 250 mg/d and 26.0% at
`500 mg/d. RRs for patients who had progressed on two prior
`chemotherapy regimens were 13.6% at 250 mg/d and 7.9%
`at 500 mg/d.
`
`Table 1. Baseline Demography of the Randomized Population
`Gefitinib
`
`No. of patients randomized
`Age
`Median
`Range
`Sex (male:female)
`Performance status
`0
`l
`2
`
`Disease stage at study entry
`”IA
`NIB
`IV
`
`Tumor histology
`Adenocarcinoma'
`Squamous
`Large-cell
`Undifferentiated
`Previous cancer treatment
`
`Failed 1 previous chemotherapy regimen
`Failed 2 previous chemotherapy regimens
`Radiotherapy
`Surgery
`lmmuno/hormonal therapy
`Symptomatic at entry
`
`No.
`
`250 mg/cl
`
`104
`
`61.0
`28 to 85
`
`No.
`
`500 mg/d
`
`106
`
`60.0
`37 to 78
`
`78:26
`
`75:25
`
`70:36
`
`66:34
`
`l 8
`73
`l 3
`
`4
`1 9
`81
`
`67
`25
`9
`3
`
`104
`46
`52
`32
`4
`67
`
`17.3
`70.2
`12.5
`
`3.8
`18.3
`77.9
`
`64.4
`24.0
`8.7
`2.9
`
`100.0
`44.2
`50.0
`30.8
`3.8
`64.4
`
`20
`72
`14
`
`16
`88
`
`71
`
`*0
`
`106
`46
`48
`25
`9
`73
`
`18.9
`67.9
`13.2
`
`1.9
`15.1
`83.0
`
`67.0
`17.0
`8.5
`7.5
`
`100.0
`43.4
`45.3
`23.6
`8.5
`68.9
`
`‘Bronchioloalveolar carcinomas were included in this group. Three patients in each dose group had adenosquamous histology.
`
`Downloaded from ascopubs.org by Reprints Desk on August 2, 2017 from 216.185.156.028
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`

`GEFITINIB IN NSCLC: THE lDEAL l TRIAL
`
`2241
`
`Table 2. Best Overall Obiective Response
`
`Getitinib
`
`250 mg/d
`500 mg/d
`No.
`%
`No.
`%
`
`
`No. oi patients evaluable
`Complete response
`Partial response
`Partial response in nonmeasurable disease
`Stable disease
`Progressive disease
`Unknown“
`Response rate
`%
`95% CI
`Disease control rate
`%
`95% CI
`
`103
`
`105
`
`0
`l8
`l
`37
`42
`5
`
`0
`17.5
`1.0
`35.9
`40.8
`4.9
`
`l 8.4
`11.5 to 27.3
`
`54.4
`44.3 to 64.2
`
`l
`19
`O
`34
`44
`7
`
`1.0
`18.1
`0.0
`32.4
`41 .9
`6.7
`
`l 9.0
`12.1 to 27.9
`
`51.4
`41.5 to 61.3
`
`Abbreviation: CI, confidence interval.
`'No conclusion was reached about the best overall tumor response (eg, because at missing scans or relevant x-ray Films).
`
`As expected, the mean number of days under treatment was
`higher for responders than for nonresponders (150 v 68 days,
`respectively); however, the number of days under treatment, as
`compared with the number of days under the trial was 95%
`versus 96% in both groups.
`
`Disease Control
`
`The disease control rate was 54.4% for the 250-mg/d group,
`which was not statistically different from that of the 500-mg/d
`group, 51.4% (P = .68; Table 2). Median duration of disease
`control for patients who responded or had stable disease was 3.2
`and 4.6 months, respectively. Disease control was similar for
`second-line (59.6%, 250 mg/d; 50.0%, 500 mg/d) and third-line
`treatment (47.8%, 250 mg/d; 53.3%, 500 mg/d). SD rate was
`35.9% at 250 mg/d and 32.4% at 500 mg/d.
`
`Disease-Related Symptom Improvement
`
`Evaluable baseline questionnaires were received from 80 and
`81 patients from the 250- and 500-mg/d groups, respectively. Of
`these, 67 and 73 patients,
`respectively, were evaluable for
`symptom improVement. Median baseline scores for LCS were
`18.0 (ranges: 250 mg/d, 4—24; 500 mg/d, 2—24) for each dose
`group, indicating that this was a symptomatic population. The
`symptom improvement rate was 40.3% (95% confidence interval
`
`Table 3. Tumor Response Evaluation by Investigators and Independent
`Response Evaluation Committee
`Investigator Evaluation
`
`REC Evaluation
`
`PR (n = 34)
`SD (n = 53)
`PD (n = 18)
`UK (n = 2)
`
`PR in = 38)
`
`SD (n = 69)
`
`31
`5
`l
`l
`
`3
`48
`l7
`l
`
`NOTE. Complete response and partial response in nonmeasurable disease are
`indicated by PR For the purpose of calculating concordance rates.
`Abbreviations: REC, response evaluation committee; PR, partial response; SD,
`stable disease; PD, progressive disease; UK, unknown due to missing slides or scans.
`
`[CI], 28.5 to 53.0) for the 250-mg/d group and 37.0% (95% CI,
`26.0 to 49.1) for the 500-mg/d group. Most patients with a tumor
`response who were evaluable for symptom improvement also
`showed an improvement in their disease-related symptoms, and
`more than 50% of the patients with SD also had symptom
`improvement (Fig 3).
`The median of the maximum change in LCS score for the
`patients with symptom improvement was 7.0 points (range, 3—17
`points) during the first interval of improvement (time between
`the first visit response of improved, to the subsequent response of
`worsened). Importantly, median time to symptom improvement
`was only 8 days (the time of first postbaseline assessment) for
`both doses. Median duration of symptom improvement was 5.1
`month (range, 1.1—5.6+ months) at 500 mg/d. At 250 mg/d, the
`median duration of symptom improvement was not calculable
`because patients were still responding at the time of data cutoff;
`symptom improvement lasted for at least 3 months in 75% of
`patients and for 6 months in 65% of patients. Median time to
`
`100
`
`80
`
`60
`
`4O
`
`2O
`
`
`
`
`
`Symptomimprovementrate(%)
`
`85.7
`
`l 250 mg/day
`
`[:1 500 mg/day
`
`
`
`(n = 13)(,n,=_14)
`CR + PR
`
`
`(n = 20x5 = 25)
`(n = 34)(n ; 34)
`SD
`PD
`
`Fig 3. Symptom improvement benefits by tumor response. CR, complete
`response; PR, partial response; SD, stable disease; PD, progressive disease.
`
`st 2, 2017 from 216.185.156.028
`Downloaded from ascopubsorg by Reprints Desk on Augu
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`

`FUKUOKA ET AL
`
`—- 250 mg/day
`- - 500mglday
`
`Atrisk
`103
`105
`
`Failures
`74
`69
`
`(months)
`Median
`2.7
`2.8
`
`
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`0
`
`1
`
`10
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`0.0
`
`A
`
`wEEg
`
`to
`I:
`.9
`‘6
`D.
`9n.
`
`Progression-free survival (months)
`
`B
`
`1.0
`
`-.
`
`— 250 mg/day
`-- 500 mg/day
`
`Atrisk
`103
`106
`
`Failures
`73
`82
`
`(months)
`Median
`7.6
`8.0
`
`a)
`2
`‘Ea)>a)
`C
`.9
`5a.
`9a.
`
`0
`
`
`
`0
`
`2
`
`4
`
`6
`
`12
`10
`8
`Overall survival (months)
`
`14
`
`16
`
`18
`
`20
`
`At risk
`Failures
`(months)
`Median
`— 250 mglday
`19
`14
`13.3
`
`- - 500mglday
`20
`13
`10.6
`
`1 0
`
`0.8
`
`a,
`
`9 E5
`
`0.6
`
`C
`.9
`1:
`8 0.4
`9a.
`
`—————
`
`0.2
`
`0.0
`
`0
`
`2
`
`4
`
`6
`
`12
`10
`8
`Overall survival (months)
`
`14
`
`16
`
`18
`
`20
`
`(B) overall
`Fig 4. Kaplan-Meier plots showing (A) progression-free survival,
`survival and (C) overall survival in patients with a complete or partial tumor
`response by dose.
`
`Safety and Tolerability
`
`A total of 209 patients were evaluable for safety and tolera-
`bility. Most AEs seen in this trial were mild (CTC, G1 or G2),
`there was no evidence of cumulative toxicity, and most events
`were reversible. Drug—related AEs observed in more than 10% of
`patients on either dose are shown in Table 5. The first occurrence
`of most AEs was seen before the end of the first month of
`
`treatment. In the 250-mg/d group, 15.5% of patients had AEs
`requiring a short treatment interruption, and none required a dose
`
`2242
`
`symptom worsening, for all patients, was longer for the 250-mg/d
`group (4.1 months) than for the 500-mg/d group (2.8 months).
`Generally, more patients showed an improvement
`in the
`pulmonary items of the LCS than in the nonpulmonary items.
`Improvements in shortness of breath, cough, and breathing were
`each seen in 78% of LCS responders, while appetite, weight loss,
`and clear thinking improved in 54% to 57% of LCS responders.
`
`PFS and 0S
`
`Median PFS was 2.7 months (95% CI, 2.0 to 2.8) for the
`250-mg/d group and 2.8 months (95% CI, 1.9 to 3.8) for the
`500-mg/d group (Fig 4a), with 29% and 39% of patients,
`respectively, progression—free after 4 months of therapy. Median
`overall survival times were 7.6 months (95% CI, 5.3 to 10.1) and
`8.0 months (95% CI 6.7 to 9.9) for the.250- and 500-mg/d
`groups, respectively (Fig 4b); 1—year survival rates were 35%
`and 29%, respectively.
`Tumor response was associated with improved OS. For patients
`with either CR or PR,
`the median OS was 13.3 months for
`250—mg/d group and 10.6 months for the 500—mg/d group (Fig 4c).
`
`QoL Assessed by T01 and FACT-L
`
`Median baseline scores were 53 for TOI (range, 15—75) and 85
`(range, 32—125) for FACT-L, indicating that this population had
`compromised QoL.
`QoL improvement rate measured by TOI was 20.9% (95% CI,
`11.9 to 32.6) for the 250-mg/d group and 17.8% (95% CI, 9.8 to
`28.5) for the 500—mg/d group. QoL improvement rate measured
`by FACT-L was 23.9% (95% CI, 14.3 to 35.9) and 21.9% (95%
`CI, 13.1 to 33.1) at 250 and 500 mg/d, respectively. The median
`time to improvement (measured by T01 and FACT-L) for both
`doses was 29 days, the time of the first postbaseline assessment.
`
`Efi‘icacy in Japanese and Non-Japanese Patients
`
`the RR was higher for Japanese patients than
`In this trial,
`non—Japanese patients (27.5%‘v 10.4%; odds ratio = 3.27; P =
`.0023). A population pharmacokinetic analysis of steady-state
`trough gefitinib plasma concentrations did not reveal any difi‘er-
`ences between Japanese and non-Japanese patients that might
`explain the difference in R (data not shown). To further investigate
`this difference in R observed between Japanese and non—Japanese
`patients, a planned multivariate logistic analysis was performed.
`Twenty—two baseline factors were evaluated independently to assess
`their value in predicting response. Using a 10% significance level,
`only seven factors were found to be predictive of response (baseline
`LCS, body mass index, PS, prior radiotherapy, histology, prior
`irnmuno/hormonal therapy, and sex). To ensure that only relevant
`baseline factors were retained in the multivariate model,
`the
`backward regression technique was employed at the 10% signifi-
`cance level. This resulted in only four factors being retained in the
`model: PS, sex, histology, and prior immuno/hormonal therapy
`(Table 4).
`The final multivariate model, including all four significant base—
`line prognostic factors and the factor for ethnicity, resulted in a
`Japanese:non—Japanese odds ratio of 1.64 (95% CI, 0.71 to 3.93;
`P = .25), which is not considered to be statistically significant.
`
`Downloaded from ascopubs.org by Reprints Desk on August 2, 2017 from 216.185.156.028
`Copyright © 2017 American Society of Clinical Oncology. All rights reserved.
`
`

`

`GEFITlNIB IN NSCLC: THE IDEAL 1 TRIAL
`
`2243
`
`
`
`Table 4. Final Adjusted Model of Prognostic Factors Associated With an Obiective Response According to a Multivariate Analysis
`Odds
`
`Parameter
`Ratio
`95% Cl
`P
`Interpretation
`
`Performance status (PS)
`
`Received prior immune/hormonal treatment‘
`
`Histology
`
`Sex
`
`Ethnicity
`
`The odds of responding is over 6 times higher For PS 0 or 1
`patients compared with P5 2 patients.
`The odds of responding is 6 times higher For patients who
`received immuno/hormonal treatments prior to entry
`compared with those who did not.
`The odds of responding is almost 3.5 times higher for patients
`with adenocarcinoma than tor patients with other tumor
`histologies.
`The odds of responding is over 2.5 times higher for females
`than males.
`_
`After accounting tor all baseline imbalances, the odds ratio
`indicates that the chance at responding is just over 1.5 times
`higher tor Japanese patients compared with non-Japanese
`patients.
`Abbreviation: CI, confidence interval.
`’Immuno/hormonal treatments include picibanil, investigational drugs, minomycin, marimastat, and tamoxifen citrate.
`
`6.26
`
`6.01
`
`3.45
`
`2.65
`
`1.64
`
`1.20 to
`l 15.36
`1 .58 to 26.15
`
`.081
`
`.01 l
`
`1.29 to 1 1.02
`
`.021
`
`1.19 to 5.91
`
`0.71 to 3.93
`
`.017
`
`.25
`
`reduction. At 500 mg/d,