2343P - Evolution of circulating tumour DNA analysis in France over 5 years

Background

The detection of oncogenic mutations in cell-free DNA (cfDNA) may identify therapeutic targets in many cancers, but it requires maximum sensitivity and specificity to avoid any loss of chance. Since 2017, the national Gen&Tiss external quality assessment (EQA) program has been assessing the performance of French laboratories for the detection of EGFR mutations in lung cancer patients by providing each year 5 reconstituted plasma samples of known composition. Since 2019, the EQA also provides an evaluation for the genes of interest in colorectal cancer and melanoma, with 5 additional samples for KRAS and NRAS mutations, and 5 others for BRAF mutations.

Methods

The performance of the methods used in France was assessed based on the results of Gen&tiss EQA campaigns conducted from 2017 to 2021.

Results

A total of 1303 EQA samples were analysed over 5 years, including 720, 142, 71 and 103 EGFR, KRAS, NRAS and BRAF-mutated samples respectively and 267 WT samples. An average of 39 laboratories were evaluated each year, with no significant change over the period. cfDNA extraction protocols varied widely between laboratories, with a trend towards a diversification of methods: 5 different protocols were used in 2017 compared to 11 in 2021. Analytical methods have also changed over the period: the use of quantitative PCR (qPCR), which predominated in 2017, has decreased over the period (from 43% to 29% of laboratories) to the benefit of NGS (from 39% to 59%). The use of digital PCR (dPCR) has remained stable (from 15% to 12%). The true positive (TP) rate averaged 84% over the period, with no significant change from one campaign to the next one (Wilcoxon signed ranks test: p = 0.776). The TP rate was significantly associated with the extraction kit (Fisher: p = 0.0002) and the analytical method (p = 0.0003) but was independent of the mutated gene (p = 0.075). qPCR was the most sensitive (TP rate = 92%), ahead of NGS (83%) and dPCR (77%). The true negative (TN) rate averaged 99% with no significant change over the period (p = 0.763). The TN rate was independent of extraction kit (p = 0.273), analytical method (p = 0.677) and mutated gene (p = 0.455).

Conclusions

Despite better analytical performance, the use of qPCR for cfDNA analysis is tending to decrease in France in favour of NGS, as the number of target mutations increases.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Gen&tiss - AFAQAP - GFCO.

Funding

Amgen.

Disclosure

All authors have declared no conflicts of interest.