`
`JOURNAL OF CLINICAL ONCOLOGY
`
`O R I G I N A L R E P O R T
`
`Fred R. Hirsch, Rafal Dziadziuszko,
`Wilbur A. Franklin, Marileila Varella-
`Garcia, Paul A. Bunn Jr, D. Ross Camidge,
`University of Colorado Cancer Center,
`Aurora, CO; Fairooz Kabbinavar, Univer-
`sity of California, Los Angeles, David
`Geffen School of Medicine, Los Ange-
`les, CA; Tim Eisen, Cambridge Biomedi-
`cal Research Center, Cambridge,
`United Kingdom; Renato Martins, Fred
`Hutchinson Cancer Research Center,
`Seattle, WA; Frederick M. Schnell,
`Central Georgia Cancer Center, Macon,
`GA; Katherine Richardson, Frank Rich-
`ardson, Bret Wacker, David W. Stern-
`berg, Jason Rusk, OSI Pharmaceuticals,
`Melville, NY.
`
`Submitted January 11, 2011; accepted
`May 26, 2011; published online ahead
`of print at www.jco.org on August 8,
`2011.
`
`Supported by an unrestricted grant
`from OSI Pharmaceuticals and from
`National Cancer Institute Lung SPORE
`Award No. P50 CA058187.
`
`Authors’ disclosures of potential con-
`flicts of interest and author contribu-
`tions are found at the end of this
`article.
`
`Clinical Trials repository link available on
`JCO.org.
`
`Corresponding author: Paul A. Bunn Jr,
`MD, University of Colorado Cancer Center,
`12801 E 17th Ave, Aurora, CO 80045;
`e-mail: Paul.Bunn@ucdenver.edu.
`
`© 2011 by American Society of Clinical
`Oncology
`
`0732-183X/11/2926-3567/$20.00
`
`DOI: 10.1200/JCO.2010.34.4929
`
`A Randomized, Phase II, Biomarker-Selected Study
`Comparing Erlotinib to Erlotinib Intercalated With
`Chemotherapy in First-Line Therapy for Advanced
`Non–Small-Cell Lung Cancer
`Fred R. Hirsch, Fairooz Kabbinavar, Tim Eisen, Renato Martins, Fredrick M. Schnell, Rafal Dziadziuszko,
`Katherine Richardson, Frank Richardson, Bret Wacker, David W. Sternberg, Jason Rusk, Wilbur A. Franklin,
`Marileila Varella-Garcia, Paul A. Bunn Jr, and D. Ross Camidge
`
`A
`
`B
`
`S
`
`T
`
`R
`
`A
`
`C
`
`T
`
`Purpose
`Erlotinib prolongs survival in patients with advanced non–small-cell lung cancer (NSCLC). We report the
`results of a randomized, phase II study of erlotinib alone or intercalated with chemotherapy (CT ⫹
`erlotinib) in chemotherapy-naïve patients with advanced NSCLC who were positive for epidermal
`growth factor receptor (EGFR) protein expression and/or with high EGFR gene copy number.
`Patients and Methods
`A total of 143 patients were randomly assigned to either erlotinib 150 mg daily orally until disease
`progression (PD) occurred or to chemotherapy with paclitaxel 200 mg/m2 intravenously (IV) and
`carboplatin dosed by creatinine clearance (AUC 6) IV on day 1 intercalated with erlotinib 150 mg orally
`on days 2 through 15 every 3 weeks for four cycles followed by erlotinib 150 mg orally until PD
`occurred (CT ⫹ erlotinib). The primary end point was 6-month progression-free survival (PFS);
`secondary end points included response rate, PFS, and survival. EGFR, KRAS mutation, EGFR
`fluorescent in situ hybridization and immunohistochemistry, and E-cadherin and vimentin protein levels
`were also assessed.
`Results
`Six-month PFS rates were 26% and 31% for the two arms (CT ⫹ erlotinib and erlotinib alone,
`respectively). Both were less than the historical control of 45% (P ⫽ .001 and P ⫽ .011,
`respectively). Median PFS times were 4.57 and 2.69 months, respectively. Patients with tumors
`harboring EGFR activating mutations fared better on erlotinib alone (median PFS, 18.2 months v
`4.9 months for CT ⫹ erlotinib).
`Conclusion
`The feasibility of a multicenter biomarker-driven study was demonstrated, but neither treatment
`arms exceeded historical controls. This study does not support combined chemotherapy and
`erlotinib in first-line treatment of EGFR-selected advanced NSCLC, and the patients with tumors
`harboring EGFR mutations had a better outcome on erlotinib alone.
`
`J Clin Oncol 29:3567-3573. © 2011 by American Society of Clinical Oncology
`
`INTRODUCTION
`
`Erlotinib, an epidermal growth factor receptor
`(EGFR) –directed tyrosine kinase inhibitor (TKI),
`prolongs progression-free survival (PFS) and overall
`survival (OS) in unselected patients with non–
`small-cell lung cancer (NSCLC) in the first-line,
`second/third-line and first-line maintenance
`therapies.1-3 Randomized studies of chemotherapy
`in combination with erlotinib demonstrated no ad-
`vantage and possible antagonism among these
`therapies in an unselected population.4,5 Preclinical
`studies suggested that G1 cell cycle arrest induced by
`
`erlotinib could interfere with the G2/M cytotoxicity
`of taxanes and suggested that appropriate schedul-
`ing of erlotinib with taxanes produce additive or
`synergistic growth inhibition.6 We previously dem-
`onstrated that patients with advanced NSCLC who
`were negative for EGFR by both fluorescent in situ
`hybridization (FISH) and immunohistochemistry
`(IHC) had no benefit from gefitinib therapy in the
`second/third-line setting.7
`These studies led to the current randomized,
`phase II study evaluating erlotinib versus chemo-
`therapy intercalated with erlotinib in chemotherapy-
`naive patients with advanced NSCLC who were
`
`© 2011 by American Society of Clinical Oncology
`
`3567
`
`APOTEX EX. 1032-001
`
`
`
`Hirsch et al
`
`positive for EGFR protein expression and/or high EGFR gene copy
`number. This study was initiated before the results from the Iressa
`Pan-Asia Study (IPASS), which identified the significance of EGFR
`mutation testing before first-line therapy, was available.8 Other goals
`were to determine the feasibility of a prospective biomarker multi-
`center study and to select a treatment arm for a randomized, phase
`III trial.
`
`PATIENTS AND METHODS
`
`Study Design
`This was an international, randomized, phase II study of erlotinib as
`single-agent treatment or of carboplatin/paclitaxel chemotherapy intercalated
`with erlotinib in newly diagnosed patient with NSCLC who had EGFR-
`positive tumors assessed by IHC or FISH. Thirty-seven centers in the United
`States and five in the United Kingdom participated. The primary end point was
`the percentage of patients alive and without tumor progression at 6 months (ie,
`6-month PFS). Secondary end points included tumor response rate (RR), PFS,
`and OS as well as the exploration of the correlation between clinical outcome
`and biomarkers of interest. Key inclusion criteria were sufficient tumor tissue
`sample for EGFR testing; histologically or cytologically advanced (ie, stages
`IIIB or IV) NSCLC; radiologically measurable or evaluable disease; and ade-
`quate organ function. Patients who received any prior or concurrent antican-
`cer therapy for advanced NSCLC and patients who had uncontrolled brain
`metastases were excluded.
`Web-based, centralized random assignment was performed by IDDI
`(Brussels, Belgium) by using an adaptive random assignment method by
`Pocock and Simon.9 Patients were stratified by the number of positive tests
`for EGFR expression (by IHC, FISH: 1 or 2) smoking status (current,
`former, or never), ECOG performance status (0/1 or 2), and extent of
`disease (stage IIIB or IV).
`The study was approved by each institution’s institutional review board/
`ethics committee. Written informed consent was obtained from all patients for
`participation, including for tissue analyses and banking.
`Treatment
`Patients were randomly assigned (1:1) to receive erlotinib 150 mg daily
`orally until disease progression (PD) occurred or to receive chemotherapy
`(paclitaxel 200 mg/m2 intravenously [IV] and carboplatin dosed by creatinine
`clearance [AUC 6] according to local practice IV on day 1) alternating with
`erlotinib 150 mg orally on days 2 through 15 every 3 weeks for four cycles,
`followed by erlotinib 150 mg orally daily until PD occurred. Patients were
`evaluated every 6 weeks by chest x-ray or computed tomography (CT) scan for
`PD. After PD, patients were treated at physician’s discretion (Fig 1). Ongoing
`patient follow-up was conducted every 3 months.
`Biomarkers
`The University of Colorado Cancer Center (UCCC, Aurora, CO) re-
`ceived tumor samples from sites to assess EGFR IHC and FISH. UCCC per-
`formed quality-control assessments before the analyses to ensure sufficient
`tumor tissue. With consent, the remnant tissue was used for EGFR mutation
`testing by Genzyme Genetics (Westborough, MA) and KRAS mutation anal-
`ysis by OSI Pharmaceuticals (Boulder, CO). IHC was assessed for E-cadherin
`and vimentin by OSI Pharmaceuticals.
`EGFR IHC
`Protein expression for EGFR by IHC was assayed with the Dako (Car-
`pentaria, CA) EGFR PharmDX kit. For the purpose of eligibility, positive
`EGFR IHC was defined by greater than 10% positive cells assessed by two
`independent reviewers.10 In cases of discrepancies, the final score was
`based on a consensus meeting.
`EGFR FISH
`FISH analysis was performed according to previously published meth-
`ods.11,12 Samples identified with EGFR high polysomy (⬎ four copies of the
`EGFR gene present in 40% to 100% cells) or with EGFR gene amplification
`
`Tumor tissue
`analyzed at
`UCCC
`(N = 214)
`
`IHC positive and/or FISH positive
`
`Screen Failure:
`IHC negative AND FISH negative
`(n = 12)
`UNK/UNK
`(n = 24)
`UNK/Negative
`(n = 3)
`
`Randomly
`assigned
`(n = 143)
`
`Arm A
`erlotinib 150
`mg/day
`(n = 72)
`
`Arm B
`CBDP (AUC6) +
`paclitaxel (200
`mg/m2) d1, x 4
`cycles;
`erlotinib 150 mg/d,
`d2-15
`(n = 71)
`
`Safety/efficacy
`analysis
`(n = 69)
`Discontinued
` PD
` AE
` Patient request
` Death
` Other reason
`
` (n = 50)
`(n = 7)
`(n = 6)
`(n = 4)
`(n = 2)
`
`Safety/efficacy analysis
`(n = 68)
`Discontinued
` PD
` AE
` Patient request
` Other reason
`
` (n = 49)
`(n = 14)
`(n = 3)
`(n = 4)
`
`Fig 1. CONSORT diagram. AUC, area under the curve; AE, adverse event;
`CBDP, carboplatin; FISH, fluorescent in situ hybridization; IHC, immunohisto-
`chemistry; PD, progressive disease; UCCC, University of Colorado Cancer
`Center; UNK, unknown.
`
`(gene/chromosome ratio ⬎ two or ⱖ 15 gene copies in ⱖ 10% cells) were
`considered positive for copy number gain (FISH positive). All other samples
`were considered FISH negative. The FISH assessment was performed by two
`independent reviewers, and discrepant assessments were solved by consen-
`sus discussion.
`EGFR Mutation
`EGFR exons 18 through 21 were amplified by polymerase chain reaction
`(PCR) at Genzyme Genetics according to their standard procedure for EGFR
`mutational analysis. The resultant PCR fragments were sequenced by using
`BigDye version 1 and 3130 Genetic Analyzer (Applied Biosystems, Foster City,
`CA). EGFR activating mutations were noted by deletions on exon 19 or L858R
`mutations on exon 21. Patients with other mutations or deletions were classi-
`fied wild type (WT) for analyses.
`KRAS Mutation, E- Cadherin, and Vimentin
`The DNA isolated for EGFR mutational analysis was used for KRAS
`mutational analysis in codons 12 and 13. The protein expression of E-cadherin
`was assessed by IHC with antibody H-108 (Santa Cruz Biotechnology No.
`7870, Santa Cruz, CA). The assessment was considered high when at least 40%
`of the cells stained with intensity 2 or 3. The vimentin status was determined by
`IHC with antibody V9 (Dako No. M0725). The results were considered high
`when there was at least 10% staining of any intensity.
`Statistical Analysis
`This was a pick-the-winner, phase II design that was not adequately
`powered to test for treatment differences, as proposed by Simon et al13 Both
`
`3568
`
`© 2011 by American Society of Clinical Oncology
`
`JOURNAL OF CLINICAL ONCOLOGY
`
`APOTEX EX. 1032-002
`
`
`
`EGFR-Selected Therapy in Advanced NSCLC
`
`treatment arms were considered experimental, and the treatment arm with the
`numerically superior PFS was to be considered for testing in future studies.
`The sample size was based on the ability to detect, with a one-sided ␣of
`.05, an improvement in the 6-month PFS rate from an historical 45% with
`standard first-line platinum-based therapy to a hypothesized 60%, which
`would be a clinically meaningful improvement.4,5 PFS was defined as the time
`from random assignment until occurrence of documented radiologic and/or
`symptomatic PD according to RECIST (Response Evaluation Criteria in Solid
`Tumors), version 1.0, or until death in the absence of progression.14 Patients
`who did not experience progression were censored on the last day known to be
`free of progression by objective tumor measurements. Patients who received
`other therapy before documented PD were censored on the day subsequent
`therapy started. Survival was defined as time from random assignment until
`documented death. Patients who were still alive were censored on the last day
`known to be alive.
`PFS and OS analyses included patients who received any study therapy.
`The 6-month PFS rates with 90% CIs were calculated for each treatment arm,
`and Kaplan-Meier estimates of PFS and OS were constructed for each treat-
`ment arm. In each arm, the 6-month PFS rate was compared with the historical
`control of 45%. Analyses of RR included patients who received any study
`therapy and had measurable disease.
`Kaplan-Meier estimates of PFS were calculated for each biomarker level
`(positive v negative or mutation v WT) within each treatment arm. Log-rank
`analyses were performed to test for significant difference between biomarker
`levels. All P values presented are for exploratory purposes. RR and disease
`control rates (DCRs) were compared between the two groups with two-sided
`Fisher’s exact tests. A P value ⱕ .05 was considered statistically significant.
`
`Table 1. Biomarker Results On the Basis of Treatment Arm and Key
`Patient Characteristics
`
`Biomarker Result
`
`IHC result
`Positive
`Negative
`FISH result
`Positive
`Negative
`EGFR mutation result
`Mutation
`Activating mutation
`Exon 19 deletion
`Exon 21 L858R mutation
`Other mutation
`No mutation
`KRAS mutation result
`Mutation
`No mutation
`E-cadherin
`High
`Low
`Vimentin
`High
`Low
`
`% of Total Patients by Treatment Arm
`
`Erlotinib
`(n ⫽ 72)
`
`CP ⫹ Intercalated
`Erlotinib (n ⫽ 71)
`
`Total
`(N ⫽ 143)
`
`93
`4
`
`54
`43
`
`18
`12
`11
`1
`6
`67
`
`18
`75
`
`36
`33
`
`29
`40
`
`92
`8
`
`54
`46
`
`17
`10
`4
`6
`7
`65
`
`23
`73
`
`30
`38
`
`17
`48
`
`92
`6
`
`54
`45
`
`17
`11
`8
`3
`6
`66
`
`20
`74
`
`33
`36
`
`23
`44
`
`RESULTS
`
`Abbreviations: CP, carboplatin/paclitaxel; IHC, immunohistochemistry; FISH,
`fluorescent in situ hybridization.
`
`Patient Characteristics and Tumor Samples
`Key patient characteristics and demographics were balanced be-
`tween arms (Appendix Table A1, online only). Two-hundred forty
`patients with advanced NSCLC were screened, and formalin-fixed,
`paraffin-embedded biopsies were obtained in 214 patients (Fig 1).
`EGFR IHC and/or FISH results were obtained for 190 samples (89%);
`24 (11%) failed the quality control analysis (eg, insufficient tissue for
`analysis) and were not evaluated. At least one of the two EGFR tests
`was positive in 175 samples (92%); 12 (6%) were negative for both
`assays; and three had combinations of negative and unknown results.
`Between March 2007 and December 2008, 143 patients were eligible
`and randomly assigned; 92% were positive by IHC, and 54% were
`positive by FISH (Table 1); 45% were positive by both IHC and FISH.
`Seventy-two patients were randomly assigned to erlotinib, and 71
`patients were randomly assigned to chemotherapy plus erlotinib; 137
`patients were included in the efficacy and safety analyses. Six patients
`did not receive study drug; three were in the erlotinib arm, and three
`were in the CT plus erlotinib arm.
`The 214 tumor tissue samples consisted of primary lung lesions
`(n ⫽ 145 [67%][), metastatic sites (n ⫽ 55 [26%]), and tumor from an
`unknown location (n ⫽ 14 [7%]). Biomarker results are listed in Table
`1. The average time from receipt of tissue at the central lab to bio-
`marker results being provided to the treatment site was 4 working days
`(range, 1 to 9 days).
`EGFR mutation results were obtained from 119 patients (83%),
`and activating EGFR mutations were found in 16 patients (11%;
`n ⫽ 11, exon-19 deletions; n ⫽ 5, exon-21 L858R). No difference in
`distribution between the treatment arms was seen. Two patients had
`concurrent L858R activating mutation and T790M-acquired resis-
`tance mutation. EGFR activating mutations were higher among
`women (16% v 6% in men), adenocarcinoma histology (15% v 0% in
`
`others), Asian ethnicity (38% v 8% in non-Asians), and never smokers
`(28% v 5% in former smokers and 4% in current smokers).
`KRAS mutation analysis was performed in 135 patients, and 29
`(20%) had mutations. No patient had both EGFR and KRAS muta-
`tion. KRAS mutation rates were highest in current smokers (40% v
`22% in former and 8% in never smokers).
`EGFR FISH was performed in 141 patients and was positive in 77
`patients (54%). No difference in the distribution of EGFR FISH posi-
`tivity was seen regarding sex, histology or smoking status. EGFR IHC
`was positive in 132 (92%) of 141 patients; no difference was associated
`with sex, histology or smoking status. E-cadherin expression was high
`in 47 (48%) of 98 patients, and vimentin was high in 33 (24%) of
`96 patients.
`The associations among EGFR mutation, KRAS mutations, and
`EGFR FISH are shown in Figure 2 for the 119 patients evaluable for
`FISH, EGFR mutation, and KRAS mutation. Of the 66 EGFR FISH-
`positive tumors, 10 had KRAS mutations. Among 16 tumors with
`EGFR activating mutations, 13 were EGFR FISH positive.
`
`Treatment Administration
`Patients in the erlotinib arm received a median of 10.3 weeks of
`treatment (range, 1.1 to 125.7 weeks). Patients in the chemotherapy
`plus erlotinib arm received a median of 9.8 weeks (range, 0.1 to
`95.6 weeks).
`
`Efficacy
`PFS. Kaplan- Meiers curves of PFS are shown in Figure 3. For
`the overall population, the curves favored the chemotherapy plus
`
`www.jco.org
`
`© 2011 by American Society of Clinical Oncology
`
`3569
`
`APOTEX EX. 1032-003
`
`
`
`CP + intercalated erlotinib (n = 68)
`Median: 4.57 months
`6-month rate: 26.4%
`
`Erlotinib (n = 69)
`Median: 2.69 months
`6-month rate: 30.7%
`
`12
`15
`18
`Time (months)
`
`21
`
`24
`
`27
`
`30
`
`CP + intercalated erlotinib (n = 50)
`Median: 5.29 months
`6-month rate: 28.9%
`
`Erlotinib (n = 49)
`Median: 2.10 months
`6-month rate: 24.6%
`
`Hirsch et al
`
`1.00
`
`0.75
`
`0.50
`
`0.25
`
`A
`
`Survival (probability)
`Progression-Free
`
`KRAS
`mutation
`positive
`
`0
`
`3
`
`6
`
`9
`
`EGFR by FISH
`
`EGFR-activating
`mutation
`positive
`
`3
`
`13
`
`0
`
`0
`
`43
`
`10
`
`1.00
`
`0.75
`
`0.50
`
`0.25
`
`B
`
`Survival (probability)
`Progression-Free
`
`0
`
`3
`
`6
`
`9
`
`12
`15
`18
`Time (months)
`
`21
`
`24
`
`27
`
`30
`
`CP + intercalated erlotinib (n = 6)
`Median: 4.90 months
`6-month rate: 41.7%
`
`Erlotinib (n = 9)
`Median: 18.20 months
`6-month rate: 88.9%
`
`1.00
`
`0.75
`
`0.50
`
`0.25
`
`C
`
`Survival (probability)
`Progression-Free
`
`0
`
`3
`
`6
`
`9
`
`12
`15
`18
`Time (months)
`
`21
`
`24
`
`27
`
`30
`
`Fig 3. Kaplan-Meier plots of progression-free survival. (A) All patients; (B) EGFR
`wild-type patients; (C) EGFR mutant patients. CP, carboplatin/paclitaxel.
`
`therapy plus erlotinib arm (P⫽ 1.0). EGFR FISH-positive patients had
`a numerically higher RR (18.9% in the erlotinib arm and 26% in the
`chemotherapy plus erlotinib arm) compared with the EGFR FISH-
`negative patients (3% in the erlotinib arm and 19% in the chemother-
`apy plus erlotinib arm; Table 3). KRAS mutations had a trend toward
`a negative effect on RR and DCR when these patients were com-
`pared with patients who were KRAS WT (erlotinib: RR 0% v 16%
`[P ⫽ .1908]; DCR 31% v 53% [P ⫽ .2168]; and chemotherapy plus
`
`Fig 2. Thirty-six patients were fluorescent in situ hybridization negative, KRAS
`wild type (WT), and EGFR WT.
`
`erlotinib arm during the first 6 months and then crossed to favor the
`erlotinib arm. The 6-month PFS rate was 31% (90% CI, 22% to 40%)
`in the erlotinib arm, and it was 26% (90% CI, 17% to 36%) with
`chemotherapy plus erlotinib. The 6-month PFS rate in each arm was
`less than the historical control of 45% (erlotinib arm, P ⫽ .011;
`chemotherapy plus erlotinib arm, P ⫽ .001). The median PFS times
`were 2.69 months and 4.57 months within the two groups, respectively
`(Table 2). The 6-month PFS rate for patients with EGFR activation
`mutations was considerably better in the erlotinib arm than in the
`chemotherapy plus erlotinib arm (89% v 42%, respectively), as was the
`median PFS (18.2 months v 4.9 months, respectively).
`Within the erlotinib arm, patients with EGFR activating muta-
`tions had a 6-month PFS rate of 89% compared with 24% for the
`EGFR WT patients (P ⬍ .001). In the chemotherapy plus erlotinib
`arm, patients with EGFR mutations had a 6-month PFS rate of 42%
`compared with 28% for the EGFR WT patients (P ⫽ .502).
`For the EGFR FISH-positive patients, the 6-month PFS rate was
`39% in the erlotinib arm, and it was 23% in the chemotherapy plus
`erlotinib arm; the PFS rates were and 22% and 30%, respectively, for
`the FISH-negative group. In the erlotinib arm, among the EGFR WT
`patients, the 6-month PFS rate for the FISH-positive group was 27%,
`and it was 21% for the FISH-negative group (P ⫽ .520).
`EGFR IHC did not confer any difference in the 6-month PFS rate
`(Table 2). KRAS mutation appeared to have a negative effect on
`6-month PFS rate for patients in both arms, although the results were
`not statistically significant (Table 2). For E-cadherin or vimentin ex-
`pression, no differential association was seen (Table 2).
`
`Tumor Response
`The overall response rate (RR; ie, CR ⫹ PR) was 11.6% in the
`erlotinib arm, and it was 22.4% in the chemotherapy plus erlotinib
`arm (Table 3). For patients with activating EGFR mutation, the RR
`was 67% in the erlotinib arm and it was 33% in the chemotherapy plus
`erlotinib arm. For EGFR WT patients, the RRs were 0% and 23% in
`the two arms, respectively. The DCR was 100% in patients who were
`EGFR activating mutation positive, and it was 36% in EGFR WT
`patients in the erlotinib arm (P ⫽ .0004) compared with 67% for
`mutation-positive patients and 68% for WT patients in the chemo-
`
`3570
`
`© 2011 by American Society of Clinical Oncology
`
`JOURNAL OF CLINICAL ONCOLOGY
`
`APOTEX EX. 1032-004
`
`
`
`EGFR-Selected Therapy in Advanced NSCLC
`
`Table 2. Analyses of PFS and OS for Biomarker Subsets
`
`Table 3. Response by Biomarker Status
`
`All Patients
`
`EGFR WT Only
`
`% of Patients by Response and Treatment Arm
`
`Biomarker Subset
`
`Erlotinib
`
`CP ⫹
`Erlotinib Erlotinib
`
`CP ⫹
`Erlotinib
`
`EGFR FISH positive
`No.
`Median PFS
`6-month PFS rate
`12-month OS rate
`EGFR FISH negative
`No.
`Median PFS
`6-month PFS rate
`12-month OS rate
`P for PFS of positive v negativeⴱ
`KRAS mutation
`No.
`Median PFS
`6-month PFS rate
`12-month OS rate
`KRAS WT
`No.
`Median PFS
`6-month PFS rate
`12-month OS rate
`P for PFS of mutated v WTⴱ
`E-cadherin positive: high
`No.
`Median PFS
`6-month PFS rate
`12-month OS rate
`E-cadherin negative: low
`No.
`Median PFS
`6-month PFS rate
`12-month OS rate
`P for PFS of high v lowⴱ
`Vimentin positive: high
`No.
`Median PFS
`6-month PFS rate
`12-month OS rate
`Vimentin negative: low
`No.
`Median PFS
`6-month PFS rate
`12-month OS rate
`P for PFS of high v lowⴱ
`
`37
`2.76
`39.2
`62.2
`
`30
`2.27
`22.2
`58.3
`.075
`
`13
`2.23
`11.5
`40.4
`
`51
`3.15
`38.2
`63.1
`.078
`
`25
`2.76
`28.0
`58.8
`
`22
`1.54
`33.2
`75.6
`.794
`
`20
`1.48
`27.1
`58.2
`
`28
`2.50
`27.5
`66.5
`.757
`
`35
`5.06
`23.4
`54.7
`
`33
`4.17
`29.8
`38.4
`.778
`
`15
`2.96
`8.6
`53.3
`
`51
`4.90
`31.7
`44.3
`.078
`
`21
`4.90
`21.4
`53.1
`
`26
`5.06
`31.0
`31.4
`.725
`
`12
`5.78
`41.6
`50.3
`
`33
`4.90
`22.0
`38.7
`.163
`
`23
`2.10
`27.6
`57.3
`
`26
`1.91
`22.0
`55.5
`.492
`
`11
`2.23
`NC
`47.7
`
`38
`1.97
`27.3
`57.9
`.550
`
`19
`2.69
`21.1
`55.9
`
`18
`1.45
`24.2
`69.6
`.495
`
`15
`1.41
`7.5
`50.3
`
`22
`2.27
`26.0
`66.1
`.287
`
`27
`5.29
`21.7
`60.1
`
`23
`4.24
`38.1
`49.5
`.652
`
`12
`2.30
`11.4
`58.3
`
`38
`5.36
`33.9
`53.5
`.092
`
`15
`5.52
`31.4
`65.2
`
`21
`5.06
`27.3
`34.4
`.836
`
`8
`6.01
`53.6
`75.0
`
`27
`5.29
`23.1
`41.6
`.058
`
`Abbreviations: PFS, progression-free survival; OS, overall survival; WT,
`wild type; CP, carboplatin/paclitaxel; FISH, fluorescent in situ hybridization;
`NC, not calculated.
`ⴱP values are from the log-rank test comparing the erlotinib and CP ⫹
`erlotinib curves.
`
`erlotinib: RR 20% v 24% [P⫽ 1.0]; DCR, 53% v 78% [P⫽ .0977]). For
`EGFR IHC, E-cadherin status, and vimentin status, no statistically
`significant differences were seen for RR.
`
`OS
`
`Among 137 patients assessed for survival, the median survival
`time was 16.7 months in the erlotinib arm, and it was 11.43 months in
`the chemotherapy plus erlotinib arm. The 12-month survival rates
`
`CR ⫹ PR
`
`CR ⫹ PR ⫹ SD
`
`CP ⫹
`Intercalated
`Erlotinib
`(n ⫽ 67)ⴱ
`
`Erlotinib
`(n ⫽ 69)
`
`CP ⫹
`Intercalated
`Erlotinib
`(n ⫽ 67)ⴱ
`
`Erlotinib
`(n ⫽ 69)
`
`11.6
`
`9.4
`33.3
`
`18.9
`3.3
`
`53.8
`66.7
`25.0
`0
`
`0
`15.7
`
`8.0
`18.2
`
`15.0
`10.7
`
`22.4
`
`21.3
`33.3
`
`25.7
`18.8
`
`36.4
`33.3
`40.0
`22.7
`
`20.0
`24.0
`
`28.6
`16.0
`
`27.3
`21.2
`
`46.4
`
`45.3
`66.7
`
`54.1
`36.7
`
`84.6
`100.0
`50.0
`35.6
`
`30.8
`52.9
`
`52.0
`36.4
`
`40.0
`42.9
`
`71.6
`
`72.1
`66.7
`
`74.3
`68.8
`
`81.8
`66.7
`100.0
`68.2
`
`53.3
`78.0
`
`76.2
`56.0
`
`100.0
`57.6
`
`Factor
`
`Overall
`EGFR by IHC status
`Positive
`Negative
`EGFR by FISH status
`Positive
`Negative
`EGFR mutation status
`Mutation
`Activating mutation
`Other mutation
`No mutation
`KRAS mutation status
`Mutation
`No mutation
`E-cadherin
`High
`Low
`Vimentin
`High
`Low
`
`Abbreviations: CR, complete response; PR, partial response; SD, stable
`disease; CP, carboplatin/paclitaxel; IHC, immunohistochemistry; FISH, fluores-
`cent in situ hybridization.
`ⴱOne patient on the CP ⫹ intercalated erlotinib arm had no measurable
`disease at baseline and was nonevaluable for response.
`
`were 59% and 46%, respectively (Table 2; Fig 4). For patients with
`activating EGFR mutations, the 12-month OS rate was 100% in the
`erlotinib arm, and it was 41.7% in the chemotherapy plus erlotinib
`arm. However, in the EGFR WT patients, survival curves were nearly
`overlapping, and the median survival times were 15.6 months and 13.3
`months in the erlotinib and the chemotherapy plus erlotinib arms,
`respectively (Fig 4).
`For the EGFR FISH-positive patients no statistically difference
`was seen between the two treatment arms (12-month OS rates of 62%
`for erlotinib and 55% for chemotherapy plus erlotinib). In the FISH-
`negative group, the 12-months OS rates were 56.5% with erlotinib and
`38.4% with chemotherapy plus erlotinib (Table 2). For the EGFR
`IHC-positive patients, the 12-month OS rates were 56.8% in the
`erlotinib arm and 40.1% in the chemotherapy plus erlotinib arm. No
`difference in survival was seen among patients within the same treat-
`ment arm for KRAS mutation versus WT, E-cadherin high versus low
`expression, or vimentin high versus low expression (Table 2).
`Toxicity. The most common adverse event was skin rash (81%
`[grades 3 to 4, 9%] in erlotinib arm and 76% [grades 3 to 4, 4%] in the
`chemotherapy plus erlotinib arm; Table A2). In the chemotherapy
`plus erlotinib arm, 10 patients (15%) had chemotherapy adjustments
`as a result of hematologic toxicity, and 29 patients (43%) had them as
`a result of nonhematologic toxicity. There was at least one dose
`interruption of erlotinib in 17 patients (25%) in the erlotinib arm and
`in 23 patients (34%) in the chemotherapy plus erlotinib arm.
`
`www.jco.org
`
`© 2011 by American Society of Clinical Oncology
`
`3571
`
`APOTEX EX. 1032-005
`
`
`
`Hirsch et al
`
`control evaluation. Thus, molecular results were available on 80% of
`the patients, demonstrating that molecular phenotyping can be done
`in the majority of patients in a reasonable time frame to select therapy.
`The primary goal was to evaluate treatment outcomes from interca-
`lating erlotinib with chemotherapy and erlotinib alone. In this EGFR-
`selected population, the intercalated therapy provided similar
`outcomes to erlotinib alone on the basis of the primary end point of
`6-month PFS. Neither treatment arm exceeded the historical 6-month
`PFS rate of 45%.4,5
`An exploratory goal was to determine whether erlotinib alone
`could be superior to intercalated therapy in any of the biomarker-
`selected patients. Patients with activating EGFR mutations treated
`with erlotinib alone had superior RRs, superior PFS, and superior OS
`compared with the intercalated therapy arm. The favorable response
`rates and outcome for patients with advanced NSCLC who had EGFR
`mutations and who were treated with EGFR TKIs alone are consis-
`tent with studies evaluating gefitinib in untreated patients with
`advanced-stage NSCLC7,15,16 and with the OPTIMAL study, which
`demonstrated superior PFS with erlotinib alone compared with
`chemotherapy in patients with EGFR mutations.16 In patients with
`EGFR mutations, the lower RR and shorter PFS with intercalated
`therapy suggest antagonism between the treatments.3,8 In patients
`with EGFR WT, the results with intercalated therapy were similar to
`those reported with chemotherapy alone.4,5
`In studies of erlotinib in later lines of therapy, patients with EGFR
`WT treated with erlotinib had a superior survival compared with
`placebo.1,2 Thus, a remaining question is whether a single biomarker
`or combination of biomarkers will help to select patients with EGFR
`WT tumors for EGFR TKI therapy in the first line and beyond. This
`trial does not support use of erlotinib or chemotherapy plus erlotinib
`in the first-line setting in EGFR WT patients who could be selected
`by any other analyzed biomarker, including EGFR FISH or KRAS
`mutations. When the EGFR-mutated patients are excluded from
`the FISH-positive analysis and from the KRAS WT analysis, there
`remains no striking difference in outcome between those treated
`with erlotinib or intercalated therapy on the basis of FISH or KRAS
`status. Our finding is supported by the results from Cancer and
`Leukemia Group B study CALGB 30406, in which no difference
`between erlotinib alone or chemotherapy plus erlotinib was found
`in never smokers or light smokers who had lung adenocarcinomas
`with high response and by the positive outcome in patients with
`EGFR mutations treated with erlotinib alone, which support the
`use of an EGFR TKI alone as first-line therapy in patients with
`NSCLC who have EGFR mutations.17
`In summary, this study could not demonstrate any benefit of
`combining chemotherapy and intercalated EGFR TKI in patients with
`advanced NSCLC. The study demonstrated the feasibility of a pro-
`spective, multi-institutional biomarker study in advanced NSCLC
`and supports the importance of determining the EGFR mutation
`status of patients with advanced NSCLC before initial therapy.
`
`CP + intercalated erlotinib (n = 68)
`Median: 11.43 months
`12-month rate: 46.4%
`
`Erlotinib (n = 69)
`Median: 16.72 months
`12-month rate: 58.6%
`
`1.00
`
`0.75
`
`0.50
`
`0.25
`
`A
`
`Overall Survival
`
`(probability)
`
`0
`
`3
`
`6
`
`9
`
`18
`15
`12
`Time (months)
`
`21
`
`24
`
`27
`
`30
`
`CP + intercalated erlotinib (n = 50)
`Median: 13.34 months
`12-month rate: 55.1%
`
`Erlotinib (n = 49)
`Median: 15.61 months
`12-month rate: 56.4%
`
`1.00
`
`0.75
`
`0.50
`
`0.25
`
`B
`
`Overall Survival
`
`(probability)
`
`0
`
`3
`
`6
`
`9
`
`18
`15
`12
`Time (months)
`
`21
`
`24
`
`27
`
`30
`
`CP + intercalated erlotinib (n = 6)
`Median: 11.43 months
`12-month rate: 41.7%
`Erlotinib (n = 9)
`Median: . months
`12-month rate: 100.0%
`
`1.00
`
`0.75
`
`0.50
`
`0.25
`
`C
`
`Overall Survival
`
`(probability)
`
`0
`
`3
`
`6
`
`9
`
`18
`15
`12
`Time (months)
`
`21
`
`24
`
`27
`
`30
`
`Fig 4. Kaplan-Meier plots of overall survival. (A) All patients; (B) EGFR wild-type
`patients; (C) EGFR mutant patients. CP, carboplatin/paclitaxel.
`
`DISCUSSION
`
`Personalized medicine requires molecular analyses of tumor tissue
`obtained before therapy to select the best treatment. One goal of this
`study was to determine whether molecular tests could be performed
`on lung cancer samples from untreated patients with advanced-stage
`NSCLC in a clinically relevant time frame (defined in this study as ⬍ 5
`working days). We obtained tissue from 215 (89.6%) of 240 screened
`patients without additional rebiopsy. Of these, 11% failed quality
`
`AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS
`OF INTEREST
`
`Although all authors completed the disclosure declaration, the following
`author(s) indicated a financial or other interest that is relevant to the subject
`matter under consideration in this article. Certain relationships marked
`with a “U” are those for which no compensation was received; those
`
`3572
`
`© 2011 by American Society of Clinical Oncology
`
`JOURNAL OF CLINICAL ONCOLOGY
`
`APOTEX EX. 1032-006
`
`
`
`EGFR-Selected Therapy in Advanced NSCLC
`
`relationships marked with a “C” were compensated. For a detailed
`description of the disclosure categories, or for more information about
`ASCO’s conflict of interest policy, please refer to the Author Disclosure
`Declaration and the Disclosures of Potential Conflicts of Interest section in
`Information for Contributors.
`Employment or Leadership Position: Frank Richardson, OSI
`Pharmaceuticals (C); Bret Wacker, OSI Pharmaceuticals (C); David W.
`Sternberg, OSI Pharmaceuticals (C) Consultant or Advisory Role: Fred
`R. Hirsch, Genentech (C), Celgene (C), GlaxoSmithKline (C), Eli Lilly
`(C), OSI Pharmaceuticals (C), Pfizer (C), Boehringer Ingelheim (C),
`Ventana (C); Tim Eisen, Roche (C), AstraZeneca (C); Rafal
`Dziadziuszko, AstraZeneca (C); Katherine Richardson, OSI
`Pharmaceuticals (C); Jason Rusk, OSI Pharmaceuticals (C); Paul A.
`Bunn Jr, AstraZeneca (C), Bayer Pharmaceuticals (C), Boehringer
`Ingelheim (C), Bristol-Myers Squibb (C), Eli Lilly (C), GlaxoSmithKline
`(C), Merck (C), Novartis (C), OSI Pharmaceuticals (C), Genentech (C)
`Stock Ownership: Tim Eisen, AstraZeneca; David W. Sternberg,