1177PD - EGFR del19/L858R activating mutation (M+) subgroup in RADIANT: Baseline characteristics, prognostic role, and disease-free survival (DFS) by stage

Date 27 September 2014
Event ESMO 2014
Session NSCLC early stage, SCLC and other thoracic malignancies
Topics Non-Small Cell Lung Cancer
Pathology/Molecular Biology
Basic Scientific Principles
Presenter Nasser Altorki
Citation Annals of Oncology (2014) 25 (suppl_4): iv409-iv416. 10.1093/annonc/mdu347
Authors N.K. Altorki1, K. Kelly2, M.E.R. O'Brien3, W.E.E. Eberhardt4, J. Wang5, J.D. Horan6, M.A. Foley5, F.A. Shepherd7
  • 1Ford-isom Professor Of Thoracic Surgery, Chief Division Of Thoracic Surgery, New Presbyterian Hospital - Weill Cornell Medical College, 10065 - New York City/US
  • 2Comprehensive Cancer Center, UC Davis, 95817 - Sacramento/US
  • 3Medical Oncology, Royal Marsden Hospital, SW36JJ - London/GB
  • 4Department Of Medical Oncology, West German Cancer Centre, University Hospital Essen, 45122 - Essen/DE
  • 5Astellas Pharma Global Development, Astellas, 60062 - Northbrook/US
  • 6Novella Clinical, Novella Clinical, 80305 - Boulder/US
  • 7University Health Network, Princess Margaret Cancer Centre, M5G 2M9 - Toronto/CA




In patients (pts) with advanced NSCLC (stage IIIB/IV), EGFR M+ is predictive of response to EGFR-TKI; data on its prognostic role are mixed. RADIANT, a randomized (2:1) phase 3 trial of erlotinib (E) v placebo (P) as adjuvant therapy of EGFR-expressing (IHC+ or FISH+) stage IB-IIIA completely resected NSCLC, failed to meet the primary DFS endpoint (HR 0.9; 95% CI 0.741, 1.104; p = 0.3). Exploratory analyses suggest E may prolong DFS in EGFR M+ pts, a secondary endpoint, (HR 0.61; 95% CI 0.384, 0.981; p = 0.04, not statistically significant due to hierarchical testing); some prognostic factors were imbalanced between arms, limiting interpretation in this subgroup. Overall survival data were immature.


Baseline characteristics were compared between EGFR M+ and wild type (WT). The prognostic role of EGFR M+ was explored by comparing DFS between EGFR M+ and WT in the P arm via a Cox model. In EGFR M+ pts, DFS was compared between E and P by stage. EGFR M+ was tested by WAVE® HS and confirmed by Sanger sequencing.


Among 973 randomized pts, 161 (16.5%) were EGFR M+ (E: 102; P: 59) and 703 (72.3%) were WT (E: 458; P: 245). Compared to WT, the EGFR M+ group had significantly more Asians (47% v 10%), women (65% v 35%), never smokers (63% v 10%), adenocarcinoma (91% v 52%), EGFR FISH+ (84% v 70%), smaller tumors (<40mm, 70% vs 49%), but higher stage (IB: 47% v 52%, II: 29% v 34%, IIIA: 22% v 14%). In the P arm, worse DFS (p = 0.23) was observed in EGFR M+ compared to WT (HR 1.28; 95% CI 0.859, 1.899; median 28.5 v 55.1 m) but the difference was not statistically significant; a similar result was observed after adjusting for selected baseline variables (HR 1.20; 95% CI 0.72, 1.99). Among EGFR M+ pts, a non-significant treatment effect on DFS was observed in stages IB (n = 75; HR 0.52; p = 0.14) and II (n = 47; HR 0.51; p = 0.09) but not IIIA (n = 36; HR 1.12; p = 0.78).


Exploratory analyses suggest EGFR M+ may be associated with higher stage and other variables. A significant prognostic effect of EGFR M+ was not observed; interpretation is limited by small sample size and possible association of EGFR M+ with other baseline variables. Further study of adjuvant E in early stage EGFR M+ NSCLC is warranted.


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 a Novella Clinical employee (formerly an OSI Pharmaceuticals employee); M.A. Foley: Margaret A Foley is an employee of Astellas; 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. All other authors have declared no conflicts of interest.