135O_PR - Plasma genotyping for predicting benefit from osimertinib in patients (pts) with advanced NSCLC

Date 15 April 2016
Event European Lung Cancer Conference 2016 (ELCC) 2016
Session NSCLC targeted therapy and circulating biomarkers
Topics Anticancer agents
Translational Research
Non-Small Cell Lung Cancer
Basic Principles in the Management and Treatment (of cancer)
Biological Therapy
Presenter Geoffrey Oxnard
Citation Journal of Thoracic Oncology (2016) 11 (supplement 4): S57-S166. S1556-0864(16)X0004-4
Authors G.R. Oxnard1, K.S. Thress2, R.S. Alden1, R. Lawrance3, C.P. Paweletz4, M. Cantarini5, C. Barrett2, J.C. Yang6, P. Jänne1
  • 1Lowe Center For Thoracic Oncology, Dana-Farber Cancer Institute, 02215 - Boston/US
  • 2Imed Oncology Translational Sciences, AstraZeneca, 02451 - Waltham/US
  • 3Biometrics And Informatics, AstraZeneca, SK10 5TL - Macclesfield/GB
  • 4Medical Oncology, Dana-Farber Cancer Institute, 02215 - Boston/US
  • 5Global Medicines Development, AstraZeneca, SK10 4TG - Macclesfield/GB
  • 6Department Of Oncology, National Taiwan University Hospital, 100 - Taipei/TW



Osimertinib (AZD9291) is a selective, irreversible EGFR-TKI recently approved by the FDA for the treatment of pts with EGFR T790M positive metastatic NSCLC who have progressed on/after EGFR-TKI therapy. In the Phase I AURA trial (NCT01802632), plasma was collected for analysis to determine whether genotyping of plasma DNA could identify pts who gain clinical benefit from osimertinib.


Pre-treated NSCLC pts from the trial (20–240 mg dosing cohorts) were included if they had a common EGFR-sensitising (sens) mutation and central-lab confirmed tumour and/or exploratory plasma genotyping (BEAMing) T790M result (n = 308). Objective response rate (ORR) and median progression-free survival in months (mPFS) were assessed, dividing pts based on either tumour or plasma genotyping. Data cut-off was 1 May 2015.


In 216 pts with both plasma and tumour genotyping results, concordance for T790M was 70%. Concordance improved to 80% limiting to 137 cases with a sens mutation detected in plasma. Outcomes were robust in 179 pts T790M+ in tumour (62% ORR, 9.7 mPFS) or in 167 pts T790M+ in plasma (63% ORR, 9.7 mPFS). Outcomes were unexpectedly favourable in 104 pts T790M− in plasma (46% ORR, 8.2 mPFS) compared with 58 pts T790M− in tumour (26% ORR, 3.4 mPFS). Using detection of the plasma sens mutation as a control, plasma T790M− cases could be differentiated into a ‘T790M undetected’ group (T790M−/sens+) with poorer outcomes (38% ORR, 4.4 mPFS) and a ‘plasma uninformative’ group (T790M−/sens−) with better outcomes (64% ORR, 15.2 mPFS).


Plasma genotyping can identify pts with T790M resistance, therefore avoiding an invasive biopsy for tumour T790M. In contrast, the ORR observed in pts with T790M− plasma genotyping likely reflects false negative results. For EGFR-TKI-resistant pts without detectable T790M in plasma, a tissue-based test is therefore advised to identify T790M+ candidates for osimertinib therapy. Testing for a plasma sens mutation may serve as a control to inform the likelihood of falsely negative plasma T790M results. These data support the investigation of a new paradigm for resistance management, with rapid plasma genotyping as a test option prior to undergoing a biopsy for T790M.

Clinical trial identification

NCT01802632 (Release date 23 February 2013)

Legal entity responsible for the study





G.R. Oxnard: Advisory board/consulting: AstraZeneca, Boehringer-Ingelheim, Clovis, Genentech, Sysmex. K.S. Thress, R. Lawrance, M. Cantarini, C. Barrett: Employee and shareholder: AstraZeneca. C.P. Paweletz: Honoraria: BioRad, Clovis Oncology. J. Yang: Advisory boards: Boehringer Ingelheim, Eli Lilly, Bayer, Roche/Genentech, Astrazeneca, Astellas, MSD, Merck Serono, Pfizer, Novartis, Clovis Oncology, Celgene. P. Jänne: Consultancy fees: AstraZeneca, Pfizer, Roche Research support: AstraZeneca, Astellas Pharmaceuticals Stock ownership: Gatekeeper Pharmaceuticals Other: Post marketing royalties on DFCI owned patent on EGFR mutations licensed to Lab Corp. All other authors have declared no conflicts of interest.