Absolute Quantification of Nucleotide Variants in Cell-Free DNA via Quantitative NGS: Clinical Application in Non-Small Cell Lung Cancer Patients

Abstract

Background/Objectives: Circulating tumor DNA (ctDNA) analysis is a powerful tool for non-invasive monitoring of tumor burden and treatment response. Reliable quantification methods are critical for the effective use of ctDNA as a tumor biomarker. Digital PCR (dPCR) offers high sensitivity and quantification, but requires the prior knowledge of tumor-specific genomic alterations. Next-generation sequencing (NGS) provides a more comprehensive approach but is semi-quantitative, relying on variant allelic fraction (VAF), which can be influenced by non-tumor cell-free DNA. Methods: We developed a novel quantitative NGS (qNGS) method for absolute quantification of nucleotide variants, utilizing unique molecular identifiers (UMIs) and of quantification standards (QSs), short synthetic DNA sequences modified to include characteristic mutations for unique identification in sequencing data. We evaluated the performance of this method using plasma samples spiked with mutated DNA and plasma pools from cancer patients. We further applied our technique to plasma samples from four non-small cell lung cancer (NSCLC) patients enrolled in the ELUCID trial. Results: Our qNGS approach demonstrated robust linearity and correlation with dPCR in both spiked and patient-derived plasma samples. Notably, the analysis of clinical samples from the ELUCID trial revealed the ability of our method to simultaneously quantify multiple variants in a single plasma sample. Significant differences in ctDNA levels were observed between baseline and post-treatment samples collected after three weeks of front-line therapy. Conclusions: We introduce a novel qNGS method that enables the absolute quantification of ctDNA, independent of non-tumor circulating DNA variations. This technique was applied for the first time to serial samples from NSCLC patients, demonstrating its ability to simultaneously monitor multiple variants, making it a robust and versatile tool for precision oncology.

Keywords: NGS; NSCLC; ctDNA; dPCR; quantification.

Figure 1

Schematic representation of quantification standards (QS). Each QS was designed based on a 103-bp reference locus. A unique 25-bp sequence was inserted adjacent to the region targeted by the NGS panel to distinguish QSs from endogenous DNA. Generics ends were introduced to allow quantification. For detailed sequences see Supplementary Figure S1.

Figure 2

Linearity of qNGS quantification of variants in synthetic samples relative to their expected values. Normal plasma was spiked with a standard containing 13 variants detected by our NGS panel. Eight samples were prepared, and each DNA extract was analyzed three times using the qNGS method. Data points are represented by crosses, with the average concentration of each variant shown as a white square. The linear regression is shown.

Figure 3

qNGS reproducibility is dependent on ctDNA concentration. Normal plasma was spiked with a standard containing 13 variants detected by our NGS panel. Eight samples were prepared, and each DNA extract was analyzed three times using the qNGS method. Coefficients of variation for each variant were measured and compared to its concentration.

Figure 4

Linearity of qNGS variant quantification in pooled NSCLC patient samples relative to dPCR quantification. Eight samples were prepared by mixing, in various proportions, three plasma samples from the cancer patients positive for EGFR exon 19 deletion (c.2235_2249del), EGFR p.L858R (c.2573T > G), and BRAF p.V600E (c.1799T > A) mutations. Somatic variants were quantified by qNGS and dPCR. The linear regression is shown.

Figure 5

Linearity of qNGS quantification of EGFR variants in NSCLC patient samples relative to dPCR quantification. Eight plasma samples from NSCLC patients with a known EGFR exon 19 deletion in their tumors were tested using qNGS and dPCR. The linear regression is shown.

Figure 6

Early kinetics of ctDNA in NSCLC patients. Plasma samples collected at baseline (w0) and after three weeks of treatment (w3) from four patients ((A–D), see details in the text) enrolled in the ELUCID trial were tested using qNGS. The concentrations of the detected variants are presented.