J Clin Oncol 34: by American Society of Clinical Oncology INTRODUCTION

Transcription

1 VOLUME 34 NUMBER 27 SEPTEMBER, 16 JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T Randomized Phase III Study Comparing With in Patients With Previously Treated Advanced Lung Adenocarcinoma: WJOG 5108L Yoshiko Urata, Nobuyuki Katakami, Satoshi Morita, Reiko Kaji, Hiroshige Yoshioka, Takashi Seto, Miyako Satouchi, Yasuo Iwamoto, Masashi Kanehara, Daichi Fujimoto, Norihiko Ikeda, Haruyasu Murakami, Haruko Daga, Tetsuya Oguri, Isao Goto, Fumio Imamura, Shunichi Sugawara, Hideo Saka, Naoyuki Nogami, Shunichi Negoro, Kazuhiko Nakagawa, and Yoichi Nakanishi See accompanying editorial on page 3233 Author affiliations appear at the end of this article. Published online ahead of print at on March 28, 16. Supported by West Japan Oncology Group, a nonprofit organization supported by unrestricted donations from several pharmaceutical companies. Terms in blue are defined in the glossary, found at the end of this article and online at This study is registered with University Hospital Medical Information Network in Japan: No Authors disclosures of potential conflicts of interest are found in the article online at Author contributions are found at the end of this article. Corresponding author: Yoshiko Urata, MD, Hyogo Cancer Center, Department of Thoracic Oncology, Kitaoji-cho, Akashi, Japan; urata@ hp.pref.hyogo.jp. 16 by American Society of Clinical Oncology X/16/3427w-3248w/$.00 DOI: 10.10/JCO A B S T R A C T Purpose The epidermal growth factor receptor (EGFR) tyrosine kinase has been an important target for non small-cell lung cancer. Several EGFR tyrosine kinase inhibitors (TKIs) are currently approved, and both gefitinib and erlotinib are the most well-known first-generation EGFR-TKIs. This randomized phase III study was conducted to investigate the difference between these two EGFR-TKIs. Patients and Methods Previously treated patients with lung adenocarcinoma were randomly assigned to receive gefitinib or erlotinib. This study aimed to investigate the noninferiority of gefitinib compared with erlotinib. The primary end point was progression-free survival (PFS). Results Five hundred sixty-one patients were randomly assigned, including 1 patients (71.7%) with EGFR mutation. All baseline factors (except performance status) were balanced between the arms. Median PFS and overall survival times for gefitinib and erlotinib were 6.5 and 7.5 months (hazard ratio [HR], 1.125; 95% CI, 0.9 to 1.347; P =.257) and 22.8 and 24.5 months (HR, 1.038; 95% CI, to 1.294; P =.768), respectively. The response rates for gefitinib and erlotinib were 45.9% and 44.1%, respectively. Median PFS times in EGFR mutation positive patients receiving gefitinibversus erlotinib were 8.3 and 10.0 months, respectively (HR, 1.093; 95% CI, to 1.358; P =.424). The primary grade 3 or 4 toxicities were rash (2.2% for gefitinib v 18.1% for erlotinib) and alanine aminotransferase (ALT)/ aspartate aminotransferase (AST) elevation (6.1%/13.0% for gefitinib v 2.2%/3.3% for erlotinib). Conclusion The study did not demonstrate noninferiority of gefitinib compared with erlotinib in terms of PFS in patients with lung adenocarcinoma according to the predefined criteria. J Clin Oncol 34: by American Society of Clinical Oncology INTRODUCTION Lung cancer is the leading cause of cancer-related death worldwide. 1 Non small-cell lung cancer (NSCLC) accounts for almost 85% of all lung cancers, and adenocarcinoma is the most common type of NSCLC. 2 The development of therapeutics targeting epidermal growth factor receptor (EGFR) tyrosine kinase has been one of the great clinical advances in NSCLC, especially in adenocarcinoma. Both gefitinib and erlotinib have anilinoquinazoline structure, bind reversely to the ATP binding pocket of the EGFR, and are referred to as first-generation EGFR tyrosine kinase inhibitors (TKIs). The recommended and maximum tolerated dosages are the same for erlotinib (150 mg/d), whereas the maximum tolerated dose of gefitinib is almost three-fold that of its standard recommended dosage of (250 mg/d). Studies on the efficacy of EGFR-TKIs began nearly a decade ago. Gradually with time, correlations were reported between better sensitivity to EGFR-TKIs and clinical factors including adenocarcinoma histology, female sex, never-smoker status, and Asian ethnicity. 3,4 Studies also showed by American Society of Clinical Oncology

2 Comparing v for Previously Treated NSCLC that several types of somatic mutations in the tyrosine kinase domain of the EGFR gene display a response to EGFR-TKI agents, with some investigators reporting this mutation to be a predictor for response to EGFR-TKIs. 5,6 EGFR-TKIs are now a mainstay in the treatment of EGFR mutation positive NSCLC, although for a fairly long time, they were used according to the clinical background of the patient with NSCLC. AphaseIItrialofgefitinib IDEAL1 3 (Iressa Dose Evaluation in Advanced Lung Cancer) showed the response rate (RR) in the Japanese subset to be 27.5%, which was higher than the RR among non-japanese patients. As a result of these data, gefitinib was approved in 02 in Japan for use as second-line treatment and beyond. Four years later, gefitinib failed to display superiority over best supportive care (BSC) in previously treated patients with NSCLC in a large phase III trial ISEL (Iressa Survival Evaluation in Lung Cancer); the median overall survival (OS), 5.6 months for gefitinib v 5.1 months for BSC; hazard ratio (HR), 0.89; P = In contrast to gefitinib, in the BR.21 trial, 8 arandomizedphase III trial of erlotinib versus placebo in previously treated patients with NSCLC, erlotinib extended survival (median OS, 6.7 months for erlotinib v 4.7 months for BSC; HR, 0.70; P =.001). was subsequently also approved in Japan in 07. When our study was being planned, both of these drugs were available in Japan in the normal clinical setting for use as second-line and beyond treatment, and EGFR mutation inspections were not widespread in Japan. There had been no prospective phase III studies directly comparing gefitinib to erlotinib in the treatment of advanced lung adenocarcinoma. Therefore, we designed a multicenter randomized phase III study to demonstrate the noninferiority of gefitinib compared with erlotinib. Treatment Plan Patients were randomly assigned in a 1:1 ratio to receive either erlotinib (150 mg/d) or gefitinib (250 mg/d) by a centralized dynamic random assignment method using the following factors: sex, stage, EGFR mutation status, ECOG PS, smoking history, line of chemotherapy, and institution. Treatment was continued until progression of disease or development of intolerable toxicities or a request by the patient or physician to discontinue treatment. The erlotinib dosage could be reduced to mg and 50 mg, and the gefitinib dosage could be reduced to 250 mg on alternate days and 250 mg once every 3 days to reduce toxic effects End Points The primary end point of this open-labeled, multicenter, randomized phase III trial was progression-free survival (PFS), which was assessed by investigators to establish the noninferiority of gefitinib to erlotinib. The secondary end points (Appendix, online only) included OS, investigatorassessed RR, disease control rate (DCR), safety, and time to treatment failure (TTF). Evaluation The baseline evaluation consisted of a complete medical history and physical examination, ECG, ECOG PS, CBC, blood chemistry, blood gas analysis, CT scan of the chest and abdomen, magnetic resonance imaging or CT of the brain, and bone scintigraphy or positron emission tomography scan. According to the study protocol, during treatment, patients underwent a CT scan of the chest and upper abdomen every 4 weeks and a CT or magnetic resonance imaging scan of the brain after 3 to 6 months and upon symptoms related to brain metastasis. All responses were defined according to Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. We used a random sampling method to sample one third of the patients. External reviews of their PFS and response were performed using CT film by experts who were not aware of the treatment assignments. Toxicity was evaluated according to the Common Terminology Criteria for Adverse Events (version 3.0). PATIENTS AND METHODS Patient Selection Patients were considered eligible if they met the following criteria: histologically or cytologically proven adenocarcinoma with stage IIIB or IV disease (American Joint Committee on Cancer version 6) or postoperative recurrence; previous treatment with at least one chemotherapy regimen; EGFR-TKI treatment naïve; evaluable disease lesions; age $ years; Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 to 2; and adequate organ function. The exclusion criteria were as follows: radiographic evidence of interstitial pneumonia or pulmonary fibrosis on chest computed tomography (CT), radiation therapy on the mediastinum or lung field within the past 2 weeks, massive or uncontrolled pleural/pericardial effusion or ascites, presence of active infection, watery diarrhea, intestinal paralysis or ileus, symptomatic brain metastasis, active double cancer, unable to swallow oral medication, serious cardiac disease, serious psychiatric disorder, pregnancy, or judged to be inappropriate by the attending doctor. Protocol amendments were made in December 11 because the Pharmaceuticals and Medical Devices Agency in Japan limited the use of gefitinib to patients harboring EGFR mutations. This resulted in only patients with EGFR mutations being enrolled in the study. All enrolled patients provided written informed consent to participate and the study protocol was approved by the West Japan Oncology Group Protocol Review Committee and the review board of each participating institution. Study Design and Statistical Analysis The primary end point of this study, PFS, was analyzed in the fullanalysis set (FAS) population by estimation of the HR and two-sided 95% CIs derived from a Cox regression model with adjustment for the random assignment factors except for investigator center. The FAS patient population included all enrolled patients and excluded patients who did not meet the inclusion criteria for this study after enrollment and were, therefore, ineligible for inclusion in the analysis. The safety population was defined as all patients who received study medication at least once. This trial was designed to demonstrate noninferiority of gefitinib compared with erlotinib with a noninferiority margin of 1.30 in terms of the adjusted HR. That is, the noninferiority was to be concluded if the upper limit of the 95% CI of the adjusted HR was less than Assuming a median PFS of 4 months in both treatment arms with 1 year of follow-up after 3.5 years of accrual, 268 patients were required in each treatment group to show noninferiority with an % statistical power at a two-sided significance level of P =.05. The sample size was set at 5 patients (2 patients per arm) in consideration for potential dropouts. No interim analysis was planned. Survival curves (PFS, OS, and TTF) were analyzed by the Kaplan Meier method and were compared between groups by the Cox regression model and log-rank test. The 95% CIs for median PFS, OS, and TTF were calculated using the Brookmeyer and Crowley method. Planned subgroup analyses for PFS were performed to examine the interaction effect of treatment arm with each of the following factors: sex, clinical stage, ECOG PS, smoking status, and EGFR mutation status. Patient characteristics (ie, age, ECOG PS, sex, clinical stage, prior treatment, EGFR mutation status, and smoking status) and tumor response were compared between the two treatment arms using the x 2 test by American Society of Clinical Oncology 3249

3 Urata et al Toxicity incidence was compared using Fisher s exact test. All P values were two-sided. Statistical analyses were performed using SAS for Windows, release 9.3 (SAS Institute, Cary, NC). RESULTS Patients From July 09 to October 12, 561 patients were enrolled from 63 hospitals in Japan (Fig 1). Two patients with no prior chemotherapy (one patient in each arm) were excluded from final analysis; thus, 2 and 279 patients were randomly assigned to gefitinib and erlotinib, respectively, in the FAS, with approximately 70% of patients being EGFR mutation-positive. Baseline characteristics (Table 1), except PS, were well balanced between the treatment arms. Five hundred fifty-three patients who received study treatment (gefitinib, n = 277; erlotinib, n = 276) were assessable for toxicity analysis. At the cutoff date for collection of data (October 28, 13), 10.4% of the patients in each arm were continuing to receive the study treatment. The most common reason for incompletion of the protocol treatment was disease progression, and no dissociation was seen in either arm. Thirtythree and 32 patients in the gefitinib and erlotinib arms, respectively, had treatment discontinued as a result of toxicity. Median follow-up times for gefitinib and erlotinib were 25.1 months (95% CI, 22.1 to 30.1 months) and 26.5 months (95% CI, 21.9 to 35.1 months), respectively. Efficacy Median PFS, TTF, and OS time for gefitinib versus erlotinib in the FAS were 6.5 and 7.5 months (adjusted HR, 1.125; 95% CI, 0.9 to 1.347; and unadjusted HR, 1.068; 95% CI, to 1.277), 5.6 and 5.3 months (HR, 1.032; 95% CI, to 1.231), and 22.8 and 24.5 months (HR, 1.038; 95% CI, to 1.294), respectively (Fig 2). Because the adjusted HR in PFS for gefitinib compared with erlotinib was not less than the noninferiority margin (HR, 1.30), this study did not meet the primary end point. The RRs for gefitinib versus erlotinib were 45.9% and 44.1%, respectively, and the DCRs were 70.9% and 75.3%, respectively (Table 2). Planned subgroup analyses for PFS in the FAS revealed no statistically significant interaction of treatment, although a noticeable interaction effect of treatment with age was suggested (Fig 3A). However, the P value of the interaction test between age (, and $ 65 years) and treatment was not statistically significant (P =.065). The PFS and response of randomly sampled patients were reviewed externally, and the results showed the same trends as the original data. At initiation of the study, EGFR mutation status was selected as one of the stratification factors. The EGFR mutation positive Patients randomly assigned (N = 561) Allocated to erlotinib (n = 281) Allocated to gefitinib (n = 2) Ineligible Received no previous (n = 1) treatment Ineligible Received no previous (n = 1) treatment Met criteria and included in full analysis set (n = 2) Did not start treatment Withdrew consent (n = 2) Experienced disease (n = 1) progression Mistreatment for (n = 1) gefitinib Met criteria and included in full analysis set (n = 279) Did not start treatment Experienced disease (n = 2) progression Confirmed double (n = 1) cancer Experienced progressed (n = 1) complication Fig 1. CONSORT diagram. Treated with erlotinib (n = 276) Completed study Disease progression Discontinued due to toxicity Discontinued due to patient refusal Discontinued due to physician's decision Death during protocol treatment Other (n = 251) (n = 172) (n = 32) (n = 25) (n = 18) (n = 2) (n = 2) Treated with gefitinib Completed study Disease progression Discontinued due to toxicity Discontinued due to patient refusal Discontinued due to physician's decision Death during protocol treatment Other (n = 275) (n = 250) (n = 185) (n = 33) (n = 9) (n = 15) (n = 1) (n = 7) by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

4 Comparing v for Previously Treated NSCLC Patient Demographics and Clinical Characteristics Table 1. Baseline Characteristics of Study Patients (n = 2) No. of Patients (%) (n = 279) All Patients (n = 559) Age, years, median (range) 67.0 (39-85) 68.0 (34-91) 67.0 (34-91).307, (41.8) 104 (37.3) 221 (39.5).275 $ (58.2) 175 (62.7) 338 (.5) Performance status (50.0) 112 (.1) 252 (45.1) (42.5) 150 (53.8) 269 (48.1) 2 21 (7.5) 17 (6.1) 38 (6.8) Sex.963 Male 128 (45.7) 127 (45.5) 255 (45.6) Female 152 (54.3) 152 (54.5) 304 (54.4) Stage.585 IIIB 26 (9.3) (7.2) 46 (8.2) IV 194 (69.3) 193 (69.2) 387 (69.2) Recurrence (21.4) 66 (23.7) 126 (22.5) Regimen.535 Second line 193 (68.9) 199 (71.3) 392 (70.1) Third line and later 87 (31.1) (28.7) 167 (29.9) Smoker.965 Never or light 141 (50.4) 141 (50.5) 282 (50.4) Ever 139 (49.6) 138 (49.5) 277 (49.6) EGFR mutation (predefined).966 Positive (L858R or Ex19del) 185 (65.0) 186 (66.7) 371 (66.4) Negative (wild type, uncommon, double) 64 (22.9) 64 (22.9) 128 (22.9) Unknown 31 (11.1) 29 (10.4) (10.7) EGFR mutation (post hoc) Positive 198 (70.7) 3 (72.8) 1 (71.7) Pure L858R or pure Ex19del 182 (65.0) 182 (65.2) 364 (65.1) Double mutation 6 (2.1) 6 (2.2) 12 (2.1) Uncommon mutation 10 (3.6) 15 (5.4) 25 (4.2) Wild type 51 (18.2) 47 (16.8) 98 (17.5) Unknown 31 (11.1) 29 (10.4) (10.7) P group consisted of patients with one of the major EGFR mutations (Ex19del or L858R). The EGFR mutation negative group consisted of patients with wild-type EGFR or EGFR mutations other than Ex19del or L858R. After FAS data analysis, we checked every EGFR mutation status by original Case Report Form. Several patients displayed uncommon mutations or double mutations including T790M. To clarify the EGFR status and effect from each EGFR-TKI, we recategorized EGFR mutation status. Four hundred one of the patients displayed at least one kind of EGFR mutation. One hundred seventy-two patients displayed a single L858R mutation (gefitinib, n = 92; erlotinib, n = ), and 192 patients displayed a single Ex19del mutation (gefitinib, n = 90; erlotinib, n = 102). Twelve patients displayed a double mutation (gefitinib, n = 6; erlotinib, n = 6), including three patients with L858R plus Ex19del, A Median (months) Events, No. (%) 7.5 (95% CI, 6.3 to 8.5) 243 (86.8) 6.5 (95% CI, 5.9 to 7.6) 243 (87.1) B Median (months) Events, No. (%) 24.5 (95% CI,.8 to 28.8) 159 (56.8) 22.8 (95% CI,.5 to 26.6) 162 (58.1) Progression-Free Survival (%) Adjusted HR (95% CI, 0.9 to 1.347) Unadjusted HR (95% CI, to 1.277) Overall Survival (%) Adjusted HR (95% CI, to 1.294) Unadjusted HR (95% CI, to 1.286) Fig 2. Kaplan-Meier graphs of (A) progression-free survival and (B) overall survival for full analysis set. HR, hazard ratio by American Society of Clinical Oncology 3251

5 Urata et al Table 2. Response to Treatment No. of Patients (%) FAS EGFR Mutation Positive Ex19del L858R Uncommon Mutation Wild Type Response (n = 227) (n = 244) (n = 1) (n = 175) (n = 83) (n = 78) (n = 67) (n = ) (n = 7) (n = 12) (n = 41) (n = 43) Complete response 3 (1.3) 1 (0.4) 2 (1.3) 1 (0.6) 2 (2.4) (1.3) Partial response 97 (42.7) 111 (45.5) 86 (53.8) 102 (58.3) 52 (62.7) 51 (65.4) 31 (46.3) 45 (56.3) 3 (42.9) 3 (25.0) 5 (12.2) 1 (2.3) Stable disease 71 (31.3) 61 (25.0) 47 (29.4) (22.9) 22 (26.5) 14 (17.9) 22 (32.8) 19 (23.8) 2 (28.6) 5 (41.7) 17 (41.5) 14 (31.6) Progressive 46 (.3) 61 (25.0) 17 (10.6) 25 (14.3) 4 (4.8) 10 (12.8) 10 (14.9) 12 (15.0) 1 (14.3) 3 (25.0) 17 (41.5) 25 (58.1) disease Not evaluable 10 (4.4) 10 (4.1) 8 (5.0) 7 (4.0) 3 (3.6) 3 (3.8) 4 (6.0) 3 (3.8) 1 (14.3) 1 (8.3) 2 (4.9) 3 (7.0) Objective response (44.1) 112 (45.9) 88 (55.0) 103 (58.9) 54 (65.1) 51 (65.4) 31 (46.3) 46 (57.5) 3 (42.9) 3 (25.0) 5 (12.2) 1 (2.3) P Disease control 171 (75.3) 173 (70.9) 135 (84.4) 143 (81.7) 76 (91.6) 65 (83.3) 53 (79.1) 65 (81.3) 5 (71.4) 8 (66.7) 22 (53.7) 15 (34.9) rate P Abbreviation: FAS, full-analysis set by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

6 Comparing v for Previously Treated NSCLC A Female (n = 304) Male (n = 255) Stage IIIB (n = 46) Stage IV (n = 387) Recurrence (n = 126) P S0 (n = 252) P S1 (n = 269) P S2 (n = 38) Never-smoker or light smoker (n = 306) Ever-smoker (n = 253) EGFR mutation positive (n = 1) EGFR wild-type (n = 98) Second line (n = 392) Third or latter line (n = 167) < 65 (n = 221) 65 (n = 338) All patients (n = 559) HR (95% CI) P (0.869 to 1.424) (0.799 to 1.342) (0.439 to 1.530) (0.876 to 1.3) (0.7 to 1.701) (0.709 to 1.227) (0.8 to 1.465) (0.785 to 3.110) (0.908 to 1.483) (0.748 to 1.261) (0.879 to 1.358) (0.847 to 1.904) (0.831 to 1.283) (0.893 to 1.677) (1.024 to 1.799) (0.741 to 1.178) (0.9 to 1.347) Favors Favors B HR (95% CI) P Female (n = 2) (0.885 to 1.571).257 Male (n = 161) (0.715 to 1.399).999 Stage IIIB (n = 31) (0.414 to 2.025).827 Stage IV (n = 274) (0.836 to 1.5).544 Recurrence (n = 96) (0.766 to 1.938).3 PS 0 (n = 191) (0.753 to 1.4).7 PS 1 (n = 182) (0.771 to 1.452).726 PS 2 (n = 28) (0.846 to 4.427).111 Never-smoker or light smoker (n = 2) 1.5 (0.907 to 1.3).197 Ever-smoker (n = 161) (0.685 to 1.349).817 EGFR Ex19del (n = 192) 1.1 (0.813 to 1.544).487 EGFR L858R (n = 172) (0.675 to 1.304).704 Uncommon mutation (n = 25) (0.464 to 2.653).772 Second line (n = 334) (0.847 to 1.369).544 Third or latter line (n = 67) (0.761 to 2.1).348 < 65 (n = 158) (1.002 to 1.983) (n = 243) (0.718 to 1.268).746 All EGFR mutation position (n = 1) (0.897 to 1.358).424 Fig 3. Forest plots of progression-free survival in (A) full analysis set and (B) subset analysis of EGFR mutation positive patients. PS, performance status Favors Favors four patients with L858R plus an uncommon mutation, three patients with L858R plus T790M, one patient with Ex19del plus an uncommon mutation, and one patient with Ex19del plus T790M. Twenty-five patients displayed uncommon mutations. In the subset analysis, the EGFR mutation positive group includes all patients with EGFR mutations, including common mutations, uncommon mutations, and double mutations. Median PFS, OS, RR, and DCR among patients in the gefitinib arm versus erlotinib arm with at least one kind of EGFR mutation were 8.3 and 10.0 months (HR, 1.093; 95% CI, to 1.358; P =.424; Fig 4A), 26.5 and 31.4 months (HR, 1.189; 95% CI, to 1.570; P =.221; Appendix Fig A1, online only), 58.9% and 55.0% (P =.476), and 81.7% and 84.4% (P =.517), respectively. Exploratory subset analysis of EGFR mutation restricted to Ex19del indicated that, for gefitinib versus erlotinib, the PFS, RR, and DCR were 11.1 and 11.5 months (HR, 1.1; 95% CI, to 1.544; P =.487), 65.4% and 65.1% (P =.965), and 83.3% and 91.6% (P =.113), respectively. When restricted to L858R mutation, PFS, RR, and DCR for gefitinib versus erlotinib were 8.1 and 8.5 months (HR, 0.938; 95% CI, to 1.304; P =.704), 57.5% and 46.3% (P =.174), and 81.3% and 79.1% (P =.744), respectively. In the subgroup with uncommon mutations (n = 25), median PFS, RR, and DCR for gefitinib versus erlotinib were 6.4 and 5.3 months (HR, 1.109; 95% CI, to 2.653; P =.815), 25.0% and 42.9% 16 by American Society of Clinical Oncology 3253

7 Urata et al A Progression-Free Survival (%) HR Median (months) 10.0 (95% CI, 8.5 to 11.2) 8.3 (95% CI, 7.2 to 9.7) (95% CI, to 1.358) No P =.424 B Progression-Free Survival (%) HR Median (months) 11.5 (95% CI, 9.5 to 13.3) 11.1 (95% CI, 8.1 to 14.7) 1.1 (95% CI, to 1.544) No P = C Progression-Free Survival (%) HR Median (months) No. 8.5 (95% CI, 6.4 to 10.6) 8.1 (95% CI, 6.4 to 9.4) (95% CI, to 1.304) P =.704 D Progression-Free Survival (%) HR Median (months) No. 1.7 (95% CI, 1.2 to 2.9) (95% CI, 1.2 to 2.5) (95% CI, to 1.904) P = Fig 4. Kaplan-Meier graphs of progression-free survival for the following patient groups: (A) EGFR mutation positive, (B) Ex19del mutation, (C) L858R mutation, and (D) EGFR wild type. HR, hazard ratio. (P =.419), and 66.7% and 71.4% (P =.829), respectively. In the subgroup with wild-type EGFR (n = 98), median PFS, RR, and DCR for gefitinib versus erlotinib were 2.0 and 1.7 months (HR, 1.270; 95% CI, to 1.904; P =.241), 2.3% and 12.2% (P =.079), and 34.9% and 53.7% (P =.083), respectively (Table 2, Figs 4B and 4D, Appendix Fig A2, online only). In exploratory subset analysis of subgroups restricted to the EGFR mutations, subgroup analyses for PFS in the FAS revealed no statistically significant interaction of treatment, although a noticeable interaction effect of treatment with age was suggested (Fig 3B). However, the P value of the interaction test between age (, and $ 65 years) and treatment was not statistically significant (P =.128) Toxicity Table 3 lists major toxicities. The incidence of grade 3 skin rash was lower in the gefitinib arm (2.2%) than the erlotinib arm (18.1%). The incidence of grade 3 or 4 increased aspartate aminotransferase (AST) and alanine aminotransferase (ALT) was 6.1% and 13.0% for gefitinib versus 2.2% and 3.3% for erlotinib, respectively. Liver function toxicity was more commonly observed in the gefitinib arm than the erlotinib arm. A 4% incidence of interstitial lung disease (ILD) was reported in each group, with three patients having grade 5 ILD in the erlotinib group. There were no significant differences between gefitinib and erlotinib in the incidence of any of the specific toxicity measures. Overall worst grade toxicity per patient (any toxicity) was computed, and the gefitinib arm showed a tendency toward significantly lower toxicity than the erlotinib arm (P,.001). Poststudy Treatment Appendix Table A1 (online only) lists the poststudy treatment. Overall, 76.4% of the erlotinib arm and 78.5% of the gefitinib arm received poststudy treatment. There was no significant difference in the percentage of patients who received either an anticancer agent or radiation therapy after protocol treatment between the erlotinib and gefitinib arms. In the gefitinib arm, 17.2% of patients received another EGFR-TKI and 21.5% received the same EGFR- TKI. In the erlotinib arm, 10.0% of patients received another EGFR-TKI and 27.5% received the same EGFR-TKI. DISCUSSION To our knowledge, this is the first prospective randomized phase III study to directly compare two first-generation EGFR- TKIs, gefitinib and erlotinib, in pretreated patients with lung by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

8 Comparing v for Previously Treated NSCLC Table 3. Adverse Events and Overall Worse Grade Adverse Events No. of Patients (%) (n = 276) (n = 277) Adverse Event All Grades Grades 1-2 Grade 3 Grade 4 Grade 5 All Grades Grades 1-2 Grade 3 Grade 4 Grade 5 P Nonhematologic toxicity Skin rash 255 (92.4) 5 (74.3) 50 (18.1) (74.7) 1 (72.6) 6 (2.2) 0 0,.001 Diarrhea 141 (51.1) 132 (47.8) 9 (3.3) (42.6) 112 (.4) 6 (2.2) General malaise 85 (30.8) 73 (26.4) 10 (3.6) 2 (0.7) 0 (28.9) 69 (24.9) 9 (3.2) 2 (0.7) Nausea 61 (22.1) 58 (21.0) 3 (1.1) (19.5) 51 (18.4) 3 (1.1) Vomiting 35 (12.7) 33 (12.0) 2 (0.7) (10.1) 25 (9.0) 3 (1.1) Infection (7.2) 14 (5.1) 6 (2.2) (9.7) 19 (6.9) 8 (2.9) Paronychia NA NA 12 (4.3) 0 0 N. A. N. A. 1 (0.4) 0 0 ILD 11 (4.0) 7 (2.5) 1 (0.4) 0 3 (1.1) 12 (4.3) 11 (4.0) 1 (0.4) T-bilirubin increased 105 (38.3) 103 (37.6) 2 (0.7) (9.8) 26 (9.4) 0 1 (0.4) 0,.001 ALT increased 105 (38.2) 96 (34.9) 8 (2.9) 1 (0.4) (50.9) 105 (37.9) 34 (12.3) 2 (0.7) 0,.001 AST increased 96 (34.9) 90 (32.7) 5 (1.8) 1 (0.4) (42.2) (36.1) 15 (5.4) 2 (0.7) 0,.001 Hematologic toxicity Leukopenia 77 (28.0) 72 (26.2) 4 (1.5) 1 (0.4) 0 75 (27.1) 75 (27.1) Neutropenia 49 (17.8) 46 (16.7) 2 (0.7) 1 (0.4) 0 58 (.9) 56 (.2) 2 (0.7) Anemia 74 (26.9) 66 (24.0) 6 (2.2) 2 (0.7) 0 65 (23.5) 56 (.2) 8 (2.9) 1 (0.4) Thrombocytopenia 7 (2.5) 5 (1.8) 2 (0.7) (2.9) 8 (2.9) Overall worst grade toxicity per patient 275 (*) 156 (56.7*) 106 (38.5*) 10 (3.6*) 3 (1.1*) 274 ( ) 178 (65.0 ) 82 (29.9 ) 14 (5.1 ) 0,.001 Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; IDL, interstitial lung disease; NA, not applicable. *The denominator of the percentage is 275, meaning the number of patients who suffered $ G1 toxicity treated with erlotinib. The denominator of the percentage is 274, meaning the number of patients who suffered $ G1 toxicity treated with gefitinib by American Society of Clinical Oncology 3255

9 Urata et al adenocarcinoma. The outcome in our study did not satisfy the primary end point, which was noninferiority of gefitinib compared with erlotinib in terms of PFS among patients with lung adenocarcinoma. This study contains seven stratification factors. Many stratification factors may undermine the random assignment in a clinical trial (eg, loss of statistical power and validity of the statistical test). It may be more important, however, to assure the comparability between the treatment arms in terms of the critical prognostic or predictive factors, including institutions. Therefore, our study used a centralized dynamic random assignment method to minimize overstratification issues and achieve the betweengroup comparability in those many factors. In Japan, gefitinib was approved in 02 and had been widely and commonly used at the time of erlotinib s approval in 07. In regard to adverse effects, gefitinib seemed to display milder adverse events than erlotinib; the cost of erlotinib is 1.5-fold the cost of gefitinib in Japan. On the basis of these factors, we established our trial as a noninferiority clinical trial. One reason for the negative primary end point may be the underestimation in the trial design (ie, we based the estimation of PFS on the phase III ISEL 7 and BR.21 8 trials). The PFS for both EGFR-TKIs was, thus, estimated to be 2 to 4 months. However, the outcomes showed PFS times of 6.5 months (gefitinib arm) and 7.5 months (erlotinib arm). The estimation of sample size may have been too small to create a firm conclusion. In 08, when this study was planned, testing for EGFR mutation was neither common nor consistently conducted in Japan. 12 Numerous recent reports indicate a strong relationship between the effect of EGFR-TKIs and EGFR mutation, which lead us to focus our attention on the data for the EGFR mutation positive population. Our study included 1 EGFR mutation positive patients (71.7% of FAS). Although only part of the subset analysis, we were able to conduct the largest prospective head-to-head trial comparing two EGFR-TKIs on EGFR mutation positive patients to date. We divided the patients into the following subgroups: L858R only (n = 172), Ex19del only (n = 192), and uncommon mutations (n = 25). Regardless of the subgroup, analysis did not show any statistically significant differences in PFS between the gefitinib and erlotinib arms. The Ex19del subgroup showed numerically better PFS and response to EGFR-TKIs than the L858R mutation subgroup. It has already reported that erlotinib and afatinib tend to be more effective for patients with Ex19del than L858R Previous reports concerning gefitinib treatment 9,10 cited no significant difference in PFS between Ex19del and L858R. In contrast, our data for gefitinib suggest that gefitinib also showed numerically better PFS for patients with Ex19del than L858R. This may indicate the need to distinguish between Ex19del and L858R populations when treating with EGFR-TKIs. Although the sample size was small for the EGFR wild-type population, the erlotinib arm indicated an RR of at least 10% even in second-line or later treatment. However, there was little response to gefitinib among the wild-type population. At data cutoff, no difference was observed in terms of OS between gefitinib and erlotinib in both the FAS and the EGFR mutation positive subgroup (Fig 2B, Appendix Fig A1). Although these data were immature and this clinical trial was lacking in power to detect a significant difference in terms of OS, up-to-date OS results will be warranted. Both drugs were well tolerated. The most common grade 3 or 4 toxicities in the erlotinib and gefitinib arms were rash and elevated ALT/AST, respectively. There was no significant difference in the incidence of ILD regardless of grade. The results were based on direct comparison between erlotinib and gefitinib, and the data are consistent with previous data from evaluations of each EGFR- TKI. 9,10,14,15 In conclusion, our study was unable to demonstrate statistical noninferiority in PFS between erlotinib and gefitinib. Subset analysis including mutation status did not reveal any populations with noteworthy differences in clinical efficacy between gefitinib and erlotinib. This suggests that, although not demonstrated to be statistically noninferior, gefitinib could be therapeutic option, considering the milder adverse effects in patients with EGFR mutation positive adenocarcinoma. AUTHORS DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST Disclosures provided by the authors are available with this article at AUTHOR CONTRIBUTIONS Conception and design: Yoshiko Urata, Nobuyuki Katakami, Satoshi Morita, Hiroshige Yoshioka, Miyako Satouchi, Shunichi Negoro, Kazuhiko Nakagawa, Yoichi Nakanishi Administrative support: Yoichi Nakanishi Provision of study materials or patients: Yoshiko Urata, Nobuyuki Katakami, Hiroshige Yoshioka, Miyako Satouchi, Daichi Fujimoto, Haruyasu Murakami, Haruko Daga, Tetsuya Oguri, Isao Goto, Shunichi Sugawara Collection and assembly of data: Yoshiko Urata, Nobuyuki Katakami, Reiko Kaji, Hiroshige Yoshioka, Takashi Seto, Miyako Satouchi, Yasuo Iwamoto, Masashi Kanehara, Daichi Fujimoto, Norihiko Ikeda, Haruyasu Murakami, Haruko Daga, Tetsuya Oguri, Isao Goto, Fumio Imamura, Shunichi Sugawara, Hideo Saka, Naoyuki Nogami Data analysis and interpretation: Yoshiko Urata, Nobuyuki Katakami, Satoshi Morita, Hiroshige Yoshioka, Miyako Satouchi Manuscript writing: All authors Final approval of manuscript: All authors REFERENCES 1. Siegel R, Naishadham D, Jemal A: Cancer statistics, 13. CA Cancer J Clin 63:11-30, Travis WD, Brambilla E, Noguchi M, et al: International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary classification of lung adenocarcinoma. J Thorac Oncol 6: , Fukuoka M, Yano S, Giaccone G, et al: Multiinstitutional randomized phase II trial of gefitinib for previously treated patients with advanced non-small-cell lung cancer (the IDEAL 1 trial). J Clin Oncol 21: , Kris MG, Natale RB, Herbst RS, et al: Efficacy of gefitinib, an inhibitor of the epidermal growth by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

10 Comparing v for Previously Treated NSCLC factor receptor tyrosine kinase, in symptomatic patients with non-small cell lung cancer: A randomized trial. JAMA 290: , Lynch TJ, Bell DW, Sordella R, et al: Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350: , Paez JG, Jänne PA, Lee JC, et al: EGFR mutations in lung cancer: Correlation with clinical response to gefitinib therapy. Science 304: , Thatcher N, Chang A, Parikh P, et al: plus best supportive care in previously treated patients with refractory advanced non-small-cell lung cancer: Results from a randomised, placebocontrolled, multicentre study (Iressa Survival Evaluation in Lung Cancer). Lancet 366: , Shepherd FA, Rodrigues Pereira J, Ciuleanu T, et al; in previously treated non-small-cell lung cancer. N Engl J Med 353: , Maemondo M, Inoue A, Kobayashi K, et al: or chemotherapy for non-small-cell lung cancer with mutated EGFR. N Engl J Med 362: , Mitsudomi T, Morita S, Yatabe Y, et al; versus cisplatin plus docetaxel in patients with non-small-cell lung cancer harbouring mutations of the epidermal growth factor receptor (WJTOG35): An open label, randomised phase 3 trial. Lancet Oncol 11: , Satoh H, Inoue A, Kobayashi K, et al: Low-dose gefitinib treatment for patients with advanced nonsmall cell lung cancer harboring sensitive epidermal growth factor receptor mutations. J Thorac Oncol 6: , Hagiwara K, Kobayashi K: Importance of the cytological sampled for the epidermal growth factor receptor gene mutation for non-small cell lung cancer. Cancer Sci 104: , Rosell R, Moran T, Queralt C, et al; Screening for epidermal growth factor receptor mutations in lung cancer. N Engl J Med 361: , Zhou C, Wu YL, Chen G, et al: versus chemotherapyas first-line treatment for patients with advanced EGFR mutation-positive non-small-cell lung cancer (OPTIMAL, CTONG-02): A multicentre, open-label, randomised, phase 3 study. Lancet Oncol 12: , Rosell R, Carcereny E, Gervais R, et al; versus standard chemotherapy as first-line treatment for European patients with advanced EGFR mutation-positive non-small-cell lung cancer (EURTAC): A multicentre, open-label, randomised phase 3 trial. Lancet Oncol 13: , Yang JC, Wu YL, Schuler M, et al: Afatinib versus cisplatin-based chemotherapy for EGFR mutation-positive lung adenocarcinoma (LUX-Lung 3 and LUX-Lung 6): Analysis of overall survival data from two randomised, phase 3 trials. Lancet Oncol 16: , 15 Affiliations Yoshiko Urata, Miyako Satouchi, and Shunichi Negoro, Hyogo Cancer Center, Akashi; Nobuyuki Katakami and Reiko Kaji, Institute of Biomedical Research and Innovation; Daichi Fujimoto, Kobe City Medical Center General Hospital, Kobe; Satoshi Morita, Kyoto University Graduate School of Medicine, Kyoto; Hiroshige Yoshioka, Kurashiki Central Hospital, Kurashiki; Takashi Seto, National Kyushu Cancer Center; Yoichi Nakanishi, Kyushu University, Graduate School of Medical Sciences, Fukuoka; Yasuo Iwamoto and Masashi Kanehara, Hiroshima City Hospital, Hiroshima; Norihiko Ikeda, Tokyo Medical University, Tokyo; Haruyasu Murakami, Shizuoka Cancer Center, Shizuoka; Haruko Daga, Osaka City General Hospital; Fumio Imamura, Osaka Medical Center for Cancer and Cardiovascular Diseases, Osaka; Isao Goto, Osaka Medical College, Takatsuki; Hideo Saka, National Hospital Organization Nagoya Medical Center; Tetsuya Oguri, Nagoya City University Graduate School of Medical Sciences, Nagoya; Shunichi Sugawara, Sendai Kousei Hospital, Sendai; Naoyuki Nogami, National Hospital Organization Shikoku Cancer Center, Matsuyama; and Kazuhiko Nakagawa, Kinki University Faculty of Medicine, Osaka-Sayama, Japan. nnn GLOSSARY TERMS epidermal growth factor receptor (EGFR): a member of a family of receptors (HER2, HER3, HER4 are other members of the family) that binds to the EGF, TGF-a, and other related proteins, leading to the generation of proliferative and survival signals within the cell. EGFR (also known as HER1) also belongs to the larger family of tyrosine kinase receptors and is generally overexpressed in several solid tumors of epithelial origin. erlotinib: also known as Tarceva (Genentech, South San Francisco, CA). is a small molecule that inhibits the tyrosine kinase activity of epidermal growth factor receptor/her1 and has been evaluated extensively in clinical trials in patients with non small-cell lung cancer, pancreatic cancer, and glioblastoma multiforme. gefitinib: a member of the class of tyrosine kinase inhibitors. (also known as Iressa, AstraZeneca, London, United Kingdom) binds to the cytoplasmic region of the epidermal growth factor receptor (EGFR) that also binds ATP. By competing with ATP binding that is essential for tyrosine phosphorylation, gefitinib inhibits activation of EGFR and blocks the cascade of reactions leading to cellular proliferation. noninferiority clinical trial: a clinical trial that aims to demonstrate that the outcome is not worse than the standard or comparator by more than a prespecified amount by American Society of Clinical Oncology 3257

11 Urata et al AUTHORS DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST Randomized Phase III Study Comparing With in Patients With Previously Treated Advanced Lung Adenocarcinoma: WJOG 5108L The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO s conflict of interest policy, please refer to or jco.ascopubs.org/site/ifc. Yoshiko Urata Honoraria: AstraZeneca, Chugai Pharma, Eli Lilly, Astellas Pharma, Novartis Nobuyuki Katakami Speakers Bureau: Dainippon Sumitomo Pharma, Chugai Pharma., Boeringer Ingelheim, AstraZeneca, Eli Lilly, Taiho Pharma, Janssen Pharma, Novartis, Pfizer, Ono Pharmaceutical, Daiichi-Sankyo Research Funding: AstraZeneca (Inst), Eisai (Inst), Ono Pharmaceutical (Inst), Kyowa Kirin Pharma (Inst), Shionogi Pharma (Inst), Daiichi- Sankyo (Inst), Taiho Pharmaceutical (Inst), Chugai Pharma (Inst), Eli Lilly (Inst), Boeringer Ingelheim (Inst), Merck Serono (Inst) Satoshi Morita Honoraria: AstraZeneca, Chugai Pharma Reiko Kaji No relationship to disclose Hiroshige Yoshioka Honoraria: Chugai Pharma, AstraZeneca, Boehringer Ingelheim Research Funding: Chugai Pharma (Inst) Takashi Seto Honoraria: AstraZeneca, Chugai Pharmaceutical, Daiichi Sankyo, Eisai, Eli Lilly, Fuji Pharma, Hisamitsu Pharmaceutical, Kyowa Hakko Kirin, Mochida Pharmaceutical, Nippon Boehringer Ingelheim, Nippon Kayaku, Ono Pharmaceutical, Pfizer, Roche Diagnostics, Sanofi, Showa Yakuhin Kako, Sumitomo Dainippon Pharma, Taiho Pharmaceutical, Takeda Pharmaceutical Research Funding: Chugai Pharmaceutical (Inst), Astellas Pharma (Inst), AstraZeneca (Inst), Novartis(Inst), Pfizer (Inst), MSD (Inst), Bayer Yakuhin (Inst), Yakult Honsha (Inst), Eisai (Inst), Nippon Boehringer Ingelheim (Inst), Eli Lilly (Inst), Merck Serono (Inst), Verastem (Inst) Miyako Satouchi Honoraria: AstraZeneca, Chugai Pharma, Taiho Pharmaceutical, Eli Lilly, Pfizer, Novartis, Boehringer Ingelheim, Bristol-Myers Squibb, Ono Pharmaceutical Research Funding: AstraZeneca (Inst), Chugai Pharma (Inst), Novartis (Inst), Ono Pharmaceutical (Inst), Bristol-Myers Squibb (Inst), Eli Lilly (Inst), Taiho Pharmaceutical (Inst), Pfizer (Inst), MSD (Inst) Yasuo Iwamoto No relationship to disclose Masashi Kanehara No relationship to disclose Daichi Fujimoto No relationship to disclose Norihiko Ikeda Honoraria: Eli Lilly, AstraZeneca, Chugai Pharma, Taiho Pharmaceutical, Olympus, Pfizer Research Funding: Eli Lilly (Inst), Chugai Pharma (Inst), Pfizer (Inst), Taiho Pharmaceutical (Inst), AstraZeneca (Inst), Eizai (Inst) Haruyasu Murakami Honoraria: Boehringer Ingelheim, Pfizer, Chugai Pharma, Taiho Pharmaceutical, AstraZeneca, Eli Lilly, Ono Pharmaceutical, Bristol-Myers Squibb, Novartis, Clovis Oncology Travel, Accommodations, Expenses: Astellas Pharma Haruko Daga Honoraria: Boehringer Ingelheim, Chugai Pharma Tetsuya Oguri Honoraria: Chugai Pharma, AstraZeneca Research Funding: Chugai Pharma, AstraZeneca Isao Goto Research Funding: Chugai Pharma (Inst) Fumio Imamura Honoraria: AstraZeneca, Pfizer Shunichi Sugawara Honoraria: Chugai Pharma, AstraZeneca Hideo Saka Stock or Other Ownership: Takeda Honoraria: Novelpharm, Eli Lilly, AstraZeneca, Olympus Research Funding: Daiichi Sankyo, AstraZeneca, Ono Pharmaceutical, Bayer, Eli Lilly Naoyuki Nogami Honoraria: AstraZeneca, Pfizer, Taiho Pharmaceutical, Daiichi Sankyo, Chugai Pharma, Eli Lilly, Nippon Boehringer Ingelheim Shunichi Negoro No relationship to disclose Kazuhiko Nakagawa Honoraria: Chugai Pharma, AstraZeneca Research Funding: Chugai Pharma (Inst), AstraZeneca (Inst) Yoichi Nakanishi Honoraria: Pfizer, Chugai Pharma, Boehringer Ingelheim Consulting or Advisory Role: MSD Research Funding: Pfizer, Boehringer Ingelheim, Chugai Pharma, Daiichi Sankyo 16 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY

12 Comparing v for Previously Treated NSCLC Acknowledgment We thank the data managers and other support staff of West Japan Oncology Group, especially Dr Shinichiro Nakamura, MD, and Kaori Mori. Appendix Progression-free survival was defined as the time from the date of random assignment to the earliest sign of disease progression, determined according to RECIST or death as a result of any cause. Patients without documented death or progression or for whom the exact date of the event was unknown at the time of analysis were censored on the last date known to be alive. Overall survival was defined as the time from date of enrollment until death from any cause. Patients alive at the time of analysis were censored on the last confirmed date of survival. Time to treatment failure was defined as the time from the date of enrollment to the date of treatment termination. Response rate was defined as the proportion of eligible patients with measurable lesions who achieved an overall response of confirmed complete response/remission or partial response/remission. The disease control rate was defined as the proportion of patients who achieved a confirmed complete response, partial response, or stable disease. Overall Survival (%) HR Median (months) 31.4 (95% CI, 25.9 to 37.9) 26.5 (95% CI, 22.6 to 32.1) (95% CI, to 1.570) No P = Fig A1. Overall survival in the EGFR-positive population. HR, hazard ratio. Progression-Free Survival (%) HR Median (months) 5.3 (95% CI, 1.7 to 9.3) 6.4 (95% CI, 1.8 to 7.2) (95% CI, to 2.653) No P = Fig A2. Progression-free survival in patients with uncommon mutation. HR, hazard ratio by American Society of Clinical Oncology

13 Urata et al Table A1. Poststudy Treatment No. of Patients (%) Treatment Total (N = 559) (n = 2) (n = 279) Poststudy treatment No 119 (21.3) 64 (22.9) 55 (19.7) Yes 433 (77.5) 214 (76.4) 219 (78.5) Anticancer agent No 148 (26.5) 73 (26.1) 75 (26.9) Yes 411 (73.5) 7 (73.9) 4 (73.1) Radiation therapy No 361(64.6) 189 (67.5) 172 (61.6) Yes 198 (35.4) 91 (32.5) 107 (38.4) Operation No 549 (98.2) 275 (98.2) 274 (98.2) Yes 10 (1.8) 5 (1.8) 5 (1.8) EGFR-TKI (rechallenge) 125 (22.4) 77 (27.5) 48 (17.2) 88 (15.7) 28 (10.0) (21.5) Both 32 (5.7) 9 (3.2) 23 (8.2) Abbreviation: EGFR-TKI, epidermal growth factor receptor tyrosine kinase inhibitor. 16 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY