1174OPROGNOSTIC AND PREDICTIVE ROLES OF EGFR COPY NUMBER AND KRAS MUTATION STATUS FROM THE RADIANT TRIAL OF ADJUVANT ERLOTINIB (E) VERSUS PLACEBO (P)

ABSTRACT Aim: RADIANT was a randomised trial of post-operative E v P in patients (pts) with completely resected stage IB-IIIA NSCLC (ASCO-2014 #7501). All pts provided pre-study tumor tissue to confirm eligibility based on EGFR protein expression or EGFR copy number (CN) and to conduct exploratory biomarker analyses. Methods: EGFR CN was determined by FISH: high CN was defined as gene amplification (EGFR/CEP7 ≥2 or ≥15 EGFR copies per cell in ≥10% of tumor cells examined) and/or high polysomy (≥4 EGFR copies in ≥40% of cells). EGFR and KRAS wild type (WT) and mutations (M+) were determined by WAVE® HS and confirmed by Sanger sequencing. Association of baseline characteristics with EGFR CN or KRAS M+ status was evaluated. Cox models were used to assess the prognostic role (P arm only) for EGFR CN (high v low) in EGFR WT, KRAS (M+ v WT) in the full analysis set (FAS), and KRAS in adenocarcinoma EGFR WT, and their predictive roles in DFS and OS. Results: Among 973 eligible pts, EGFR CN was determined for 954, and EGFR and KRAS mutation status for 921 and 828 pts, respectively. High EGFR CN was found in 73.4% (700/954) pts. The only baseline characteristic associated with EGFR CN was EGFR M + , which occurred more frequently in pts with high CN v low (21.5% v 8.8%). KRAS M+ was seen in 17.3% (143/828) pts (codon 12 n = 128; codon 13 n = 9; codon 61 n = 6), and was more frequent in non-Asians, adenocarcinoma, smokers, women, Americas region, and low EGFR CN. EGFR CN (in EGFR WT) or KRAS status was not prognostic or predictive of benefit from E. Prognostic Role (P arm only) EGFR CN (high v low) in EGFR WT KRAS (M+ v WT) in FAS* KRAS (M+ v WT) in Adeno EGF WT DFS HR 0.73 p = 0.12 1.14 p = 0.57 0.84 p = 0.54 OS HR 0.86 p = 0.54 1.19 p = 0.56 0.84 p = 0.63 Predictive Role High, Low M + , WT M + , WT DFS HR (E/P) Interaction 1.07, 0.71 p = 0.12 0.81, 0.93 p = 0.64 0.80, 0.82 p = 0.95 OS HR (E/P) Interaction 1.15, 0.81 p = 0.24 1.05, 1.15 p = 0.78 0.96, 0.89 p = 0.89 *Results from analyses of KRAS M+ in Codon 12 only vs WT were similar. Conclusions: These exploratory analyses of KRAS mutation status and EGFR CN did not identify a subgroup that may benefit from adjuvant erlotinib treatment. Disclosure: F.A. Shepherd: Frances A Shepherd participated on advisory boards for Roche/Astellas/OSI 1+ years ago which were compensated, and gave a compensated lecture for Roche 2+ years ago; W.E.E. Eberhardt: Honoraria: Ad board -Roche (R), Pfizer (P), BI, Novartis (N), AstraZeneca (AZ), GSK, BMS, Amgen, Teva, MerckSerono (MS), Merck (M), Astellas, Eli Lilly (EL); educational lectures -R, P, BI, N, AZ, GSK, BMS, Amgen, MS, M, EL; funding clinical trial –EL; J. Wang: Joe Wang is an employee of Astellas; J.D. Horan: Julie D Horan is an employee of Novella Clinical (former OSI Pharmaceuticals employee); M.A. Foley: Margaret A Foley is an employee of Astellas; K.K. Iwata: Kenneth K Iwata is an employee of Astellas S.C. Gill: Stanley C Gill is an employee of Astellas; F.C. Richardson: Frank C Richardson is an employee of FC Richardson Consulting LTC. All other authors have declared no conflicts of interest.