Author of

• 14490

  +

  P2.04 - Poster Session/ Biology, Pathology, and Molecular Testing (ID 234)

  • Event: WCLC 2015
  • Type: Poster
  • Track: Biology, Pathology, and Molecular Testing
  • Presentations: 1
  • Moderators:
  • Coordinates: 9/08/2015, 09:30 - 17:00, Exhibit Hall (Hall B+C)

  • 14574

    +

    P2.04-084 - Next-Generation Sequencing with Digital Droplet PCR for Circulating Tumor DNA Quantification in Non-Small-Cell Lung Cancer Patients (ID 1499)

    09:30 - 09:30  |  Author(s): P. Combe
    • Abstract

    Background:
    We aimed to compare different methods to assess the dynamics of circulating tumor DNA (ctDNA) in metastatic NSCLC pts.
    
    Methods:
    A cohort of advanced NSCLC pts was followed by serial plasma blood samples and RECIST assessments every 2 months. Plasmatic cell-free DNA was extracted with the QIAsymphony DSP Virus/Pathogen kit (Qiagen). EGFR and KRAS mutations were assessed in plasmatic DNA using Q-PCR and picoliter droplet based digital PCR (ddPCR, Raindance) with CAST probes. Nexte-Generation Sequencing (NGS) was performed using the Lung and Colon Cancer Panel v2 (Iontorrent, AmpliSeq, Lifetechnologies) with a target of 10,000X (Proton). For NGS and dPCR, the limit of blank was determined for each mutation.
    
    Results:
    We included 37 pts treated by TKI (n=21) or chemotherapy (n=16), in 1[rst] (n=31) or 2[nd] line (n=6). Median follow-up, progression-free survival (PFS) and overall survival were 12, 6 and 31 months, respectively. Tumor mutations were: EGFR [L858R (n=9), Del19 (n=9), L861Q (n=1), Ins20 (n=2)], KRAS (n=2), TP53 (n=4). No mutation was found in 8 pts. At baseline, ctDNA was positive by one of the 3 technics in 22/29 (76%) pts, with 2 pts being positive by dPCR only. ctDNA quantification was highly correlated between NGS and ddPCR r[2]=0.74), but not with Q-PCR. The fraction of ctDNA (‰) was associated with radiological outcome (ROC AUC 0.87 and 0.81 for NGS and dPCR). The relative change of ctDNA between 2 consecutive samples did not improve the prediction of tumor evolution (AUC 0.74 and 0.76). RECIST tumor progression was best predicted by pDNA >1.1‰ in NGS (sensitivity 0.85, specificity 0.80), and >1.4‰ in dPCR (sensitivity 0.82, specificity 0.74). Persistence of >1‰ ctDNA under treatment was associated with a short PFS (2.2 months vs. 8 months, log-rank P = .0002).
    
    Conclusion:
    In this cohort, the persistence of plasmatic tumor DNA using NGS (10,000 X) or dPCR was associated with treatment failure in NSCLC patients. A threshold of 1/1000 mutation ratio was clinically meaningful using both technics. Plasmatic tumor DNA normalization could be evaluated in addition to standard RECIST criteria as clinical endpoints for clinical trials.