1250PEPIDERMAL GROWTH FACTOR RECEPTOR (EGFR) AND ABCG2 POLYMORPHISMS AND TREATMENT OUTCOME IN THE RANDOMIZED PHASE III TORCH TRIAL IN ADVANCED NON-SMALL CELL LUNG CANCER (NSCLC) PATIENTS.

ABSTRACT Aim: In the TORCH trial, molecularly unselected pts with advanced NSCLC were randomized to erlotinib or cisplatin + gemcitabine (CisGem) as first line within planned treatment sequences. To confirm previous data or explore new biomarkers, polymorphisms potentially affecting erlotinib activity (EGFR-216G > T [EGFR216], EGFR-191C > A [EGFR191], EGFR intron 1 CA-dinucleotide-repeat [CADR], ABCG2 + 421C > A [ABCG2]) were tested and the interaction with treatment was studied. Methods: Genotyping was by Sanger sequencing and Taqman. Interaction between each polymorphism and treatment outcome was tested by Cox model for progression-free survival (PFS) and by Zelen exact test for response rate (RR) and toxicity (skin and diarrhoea). Results: EGFR216 (known in 240 pts): G/G 32%, T/G 43%, T/T 25%. EGFR191 (240 pts): C/C 81%, A/C 17%, A/A 2%. ABCG2 (257 pts): C/C 85%, A/C 14%, A/A 1% (A/- 15%). CADR (262 pts): short/short (S/S) 28%, long/short 31%, long/long 40%. EGFR216 and EGFR191 had no significant interaction with treatment outcomes. ABCG2 polymorphism has no interaction with RR and toxicity but significantly affects treatment effect on PFS (interaction p = 0.039). HR (erlotinib vs CisGem) 1.69 (95% CI 1.28 – 2.23) in ABCG2 C/C and 0.71 (95% CI 0.34 – 1.47) in ABCG2 A/-. HR for ABCG2 A/- vs C/C was 0.68 (95% CI 0.40 – 1.14) and 1.47 (95%CI 0.90-2.44) in erlotinib and Cis/Gem arm respectively. CADR polymorphism was associated with skin toxicity but not diarrhoea: erlotinib produced more severe skin toxicity in CADR S/S pts (20% vs 6%, odds ratio 4.15, 95% CI 1.30-13.25, p = 0.013), while no severe skin toxicity was reported with CisGem. Conclusions: TORCH data: (a) do not confirm previous evidence on the value of EGFR216 as predictive of EGFR TKI effect; (b) confirm no value of EGFR191; (c) suggest that erlotinib efficacy might be higher in ABCG2 A/- cases; and (d) suggest that skin toxicity of erlotinib might be higher in CADR S/S cases. Disclosure: M. Di Maio and C. Gridelli: acted as consultant for Eli Lilli, Roche and received honoraria from Eli Lilli, Roche; M. Tsao: acted as consultant for AstraZeneca, Roche; he received honoraria from AstraZeneca, Roche and he received a research funding from Roche; C.A. Butts: acted as consultant for Roche and received honoraria from Roche; R. Feld: acted as consultant for Roche, received honoraria from Roche and received a research funding from Roche; F. Perrone: received honoraria from Roche and received a research funding from Roche. All other authors have declared no conflicts of interest.