69P - External quality assessment of EGFR testing in circulating DNA: a french pilot study

Date 10 October 2016
Event ESMO 2016 Congress
Session Poster display
Topics Translational Research
Basic Principles in the Management and Treatment (of cancer)
Presenter Marc Denis
Citation Annals of Oncology (2016) 27 (6): 15-42. 10.1093/annonc/mdw363
Authors M.G. Denis1, I. Rouquette2, A. Morel3, C. Villalva4, A. Lespagnol5, C. Caumont6, P. Gueguen7, K. Durand8, C. Collin9
  • 1Biochemistry, CHU de Nantes, 44093 - Nantes/FR
  • 2Pathology, Institut Universitaire du Cancer -Toulouse- Oncopole, Toulouse/FR
  • 3Oncopharmacologie, Centre Paul Papin, Angers/FR
  • 4Oncologie Biologique, CHU Poitiers, Jean Bernard Hôpital, Poitiers/FR
  • 5Genetique, CHU de Pontchaillou, Rennes/FR
  • 6Biologie Des Tumeurs, CHU Hôpital Haut-Lévêque, Bordeaux/FR
  • 7Génétique, C.H.U. Brest - Hôpital Morvan, Brest/FR
  • 8Centre Recherche Biologie, CHU Hôpital Haut-Lévêque, Limoges/FR
  • 9Biochimie, CHU Hôpital Trousseau, Tours/FR



Tissue testing for NSCLC patients is routinely performed in France in the regional platforms certified by the French National Cancer Institute (INCa). All laboratories participate in an annual EQA scheme for tissue testing, but there is no EQA for circulating tumor DNA testing. Therefore, we set up a pilot study to assess quality testing in western France.


Artificial samples were prepared by supplementing normal plasma (Clinisciences) with DNA extracted from control FFPE sections (Horizon Diagnostics) or plasma from NSCLC patients. Aliquots (2 ml) of 8 different samples were sent in dry ice. DNA extraction and EGFR testing (exon 19 deletions, L858R, G719X and T790M mutations) were performed according to local practice. Data were collected and compared to the expected results.


We collected 10 complete sets of data from 9 labs. DNA was extracted from 1 ml (n = 4) or 2ml (n = 6) using the QIAmp circulating DNA kit (Qiagen; n = 3), the Maxwell system (Promega; n = 4) or the cfDNA sample prep (Roche; n = 3). Mutation testing was performed by NGS (n = 3), using the COBAS EGFRv2 (Roche; n = 3) or the Therascreen EGFR RGQ kit (Qiagen; n = 2), using droplet digital PCR (BioRad; n = 1) or pyrosequencing (Qiagen; n = 1). A single false positive result was observed (T790M detected by NGS). The sensitivity (number of mutations detected / number of mutations present in the set of samples) and the number of correct genotypes are presented on the table. This pilot study suggested that, under the specific conditions of this scheme, the COBAS kit was the most sensitive approach.

Lab N° Sensitivity Correct genotypes
NGS 3 66.7% (6/9) 62.5% (5/8)
5 66.7% (6/9) 62.5% (5/8)
7 77.8%(7/9) 62.5% (5/8)
COBAS EGFR v2 1 88.9% (8/9) 87.5% (7/8)
2 88.9% (8/9) 87.5% (7/8)
6 88.9% (8/9) 87.5% (7/8)
Therascreen RGQ v2 1 66.7% (6/9) 62.5% (5/8)
8 55.6% (5/9) 50.0% (4/8)
ddPCR 4 55.6% (5/9) 62.5% (5/8)
Pyrosequencing 9 22.2% (2/9) 25.0% (4/8)


This pilot EQA allowed each lab to evaluate its practice and could be used to improve their process. These information will be important for labs that have not yet decided which technique to use for ctDNA testing. Samples were relatively simple to prepare and it will be easy to scale-up this process. A similar approach using other genes (BRAF, KRAS and NRAS) will also be developed. Supported by a grant from Astra Zeneca.

Clinical trial identification

Legal entity responsible for the study



Astra Zeneca


M.G. Denis: Advisory board Qiagen. All other authors have declared no conflicts of interest.