Molecular profiling from liquid biopsy, in particular cell-free DNA (cfDNA), represents an attractive alternative to tissue biopsies for the detection of actionable targets and tumor monitoring. In addition to PCR-based assays, Next Generation Sequencing (NGS)-based cfDNA assays are now commercially available and are being increasingly adopted in clinical practice. However, the validity of these products as well as the clinical utility of cfDNA in the management of patients with cancer has yet to be proven. Within framework of the Innovative Medicines Initiative (IMI) program CANCER-ID we evaluated the use of commercially available reference materials designed for ctDNA testing and cfDNA derived from Diagnostic Leukaphereses (DLA) for inter- and intra-assay as well as intra- and inter-laboratory comparisons. In three experimental setups, a broad range of assays including ddPCR, MassARRAY and various NGS-based assays were tested. We demonstrate that both reference materials with predetermined VAFs and DLA samples are extremely useful for the performance assessment of mutation analysis platforms. Moreover, our data indicate a substantial variability of NGS assays with respect to sensitivity and specificity highlighting the importance of extensive validation of the test performance before offering these tests in clinical routine practice.
Analysis of circulating tumor DNA (ctDNA) is a potential minimally invasive molecular tool to guide treatment decision-making and disease monitoring. A suitable diagnostic-grade platform is required for the detection of tumor-specific mutations with high sensitivity in the circulating cell-free DNA (ccfDNA) of cancer patients. In this multicenter study, the ccfDNA of 72 patients treated for advanced-stage non-small cell lung cancer (NSCLC) was evaluated using the UltraSEEK® Lung Panel on the MassARRAY® System, covering 73 hotspot mutations in EGFR, KRAS, BRAF, ERBB2, and PIK3CA against mutation-specific droplet digital PCR (ddPCR) and routine tumor tissue NGS. Variant detection accuracy at primary diagnosis and during disease progression, and ctDNA dynamics as a marker of treatment efficacy, were analyzed. A multicenter evaluation using reference material demonstrated an overall detection rate of over 90% for variant allele frequencies (VAFs) > 0.5%, irrespective of ccfDNA input. A comparison of UltraSEEK® and ddPCR analyses revealed a 90% concordance. An 80% concordance between therapeutically targetable mutations detected in tumor tissue NGS and ccfDNA UltraSEEK® analysis at baseline was observed. Nine of 84 (11%) tumor tissue mutations were not covered by UltraSEEK®. A decrease in ctDNA levels at 4–6 weeks after treatment initiation detected with UltraSEEK® correlated with prolonged median PFS (46 vs. 6 weeks; p < 0.05) and OS (145 vs. 30 weeks; p < 0.01). Using plasma-derived ccfDNA, the UltraSEEK® Lung Panel with a mid-density set of the most common predictive markers for NSCLC is an alternative tool to detect mutations both at diagnosis and during disease progression and to monitor treatment response.