Monitoring of somatic mutations in circulating cell-free DNA by digital PCR and next-generation sequencing during afatinib treatment in patients with lung adenocarcinoma positive for EGFR activating mutations.

Background: Analysis of circulating cell-free DNA (cfDNA) is under intensive investigation for its potential to identify tumor somatic mutations. We have now explored the usefulness of such liquid biopsy testing with both the digital polymerase chain reaction (dPCR) and next-generation sequencing (NGS) during treatment of patients with the epidermal growth factor receptor (EGFR) inhibitor afatinib.

Patients And Methods: Eligible patients had advanced lung adenocarcinoma with EGFR activating mutations and were treated with afatinib. Plasma samples were collected before and during (4 and 24 weeks) afatinib treatment as well as at disease progression. Tumor and plasma DNA were analyzed by dPCR and NGS.

Results: Thirty-five patients were enrolled. The objective response rate and median progression-free survival (PFS) were 77.1% and 13.8 months, respectively. Tumor and plasma DNA were available for 32 patients. dPCR and NGS detected EGFR activating mutations in 81.3% and 71.9% of baseline cfDNA samples, respectively. In 19 patients treated with afatinib for ≥24 weeks, the number of EGFR mutant alleles detected in cfDNA by dPCR declined rapidly and markedly after treatment onset, becoming undetectable or detectable at only a low copy number (<10 copies per milliliter) at 4 weeks. Median PFS was slightly longer for patients with undetectable EGFR mutant alleles in cfDNA at 4 weeks than for those in whom such alleles were detectable (14.3 versus 10.0 months). A total of 45 somatic mutations was identified in baseline tumor DNA, and 30 (66.7%) of these mutations were identified in cfDNA by NGS. Allele frequency for somatic mutations in cfDNA determined by NGS changed concordantly during afatinib treatment with the number of EGFR mutant alleles determined by dPCR.

Conclusions: Monitoring of cfDNA by dPCR is informative for prediction of afatinib efficacy, whereas that by NGS is reliable and has the potential to identify mechanisms of treatment resistance.