Application of circulating tumor DNA in prospective clinical oncology trials - standardization of preanalytical conditions

Abstract

Circulating tumor DNA (ctDNA) has emerged as a potential new biomarker with diagnostic, predictive, and prognostic applications for various solid tumor types. Before beginning large prospective clinical trials to prove the added value of utilizing ctDNA in clinical practice, it is essential to investigate the effects of various preanalytical conditions on the quality of cell-free DNA (cfDNA) in general and of ctDNA in particular in order to optimize and standardize these conditions. Whole blood samples were collected from patients with metastatic cancer bearing a known somatic variant. The following preanalytical conditions were investigated: (a) different time intervals to plasma isolation (1, 24, and 96 h) and (b) different preservatives in blood collection tubes (EDTA, CellSave, and BCT). The quality of cfDNA/ctDNA was assessed by DNA quantification, digital polymerase chain reaction (dPCR) for somatic variant detection and a β-actin fragmentation assay for DNA contamination from lysed leukocytes. In 11 (69%) of our 16 patients, we were able to detect the known somatic variant in ctDNA. We observed a time-dependent increase in cfDNA concentrations in EDTA tubes, which was positively correlated with an increase in wild-type copy numbers and large DNA fragments (> 420 bp). Using different preservatives did not affect somatic variant detection ability, but did stabilize cfDNA concentrations over time. Variant allele frequency was affected by fluctuations in cfDNA concentration only in EDTA tubes at 96 h. Both CellSave and BCT tubes ensured optimal ctDNA quality in plasma processed within 96 h after blood collection for downstream somatic variant detection by dPCR.

Keywords: DNA contamination; blood collection tube; cell-free DNA; circulating tumor DNA; preanalytical condition.

Figure 1

cfDNA concentrations for different preanalytical conditions. Boxes [interquartile ranges (IQR)] and whiskers (1.5 × IQR) are shown together with the median (black horizontal line) of the log cfDNA concentrations in copies per mL plasma of 16 patients for the different preanalytical conditions. Outliers are displayed as black dots. The Wilcoxon signed rank test was used to compare the difference between matched 1‐h and 24‐h samples relative to the difference between matched 1‐h and 96‐h samples. *P < 0.001.

Figure 2

Correlation between wild‐type or variant copy numbers and cfDNA concentration. The log number of wild‐type copies (A) or variant copies (B) in copies per mL plasma on the x‐axis is plotted against the log cfDNA concentrations in copies per mL plasma on the y‐axis. Data points correspond to single sample measurements from each time interval and each type of preservative. Correlations were tested by Spearman's rank correlation coefficient. *P < 0.001. Five patients with undetectable variant copy numbers in ctDNA are removed from plot B.

Figure 3

β‐Actin fragmentation assay for different preanalytical conditions. (A) Principle of β‐actin fragmentation assay. dPCR wells containing only 136‐bp signal are indicative of fragmented DNA (fragments < 200 bp), whereas the 420‐bp primer set will only bind to intact DNA (> 420 bp). When a large intact DNA fragment (> 2000 bp) is present in one of the wells, both primer sets can bind, resulting in a mixed signal. In theory, this can also occur when a small (< 200 bp) and large (> 420 bp) DNA fragment is present together in one well.(B) Results of β‐actin fragmentation assay. Boxes [interquartile ranges (IQR)] and whiskers (1.5 × IQR) are shown together with the median (black horizontal line) of the number of β‐actin copies for the different preanalytical conditions. Outliers are displayed as black points. The Wilcoxon signed rank test was used to compare the difference between matched 1‐h and 24‐h samples relative to the difference between matched 1‐h and 96‐h samples for the different fragment sizes. *P = 0.002; **P < 0.001.

Figure 4

VAF of 11 patients for different preanalytical conditions. Data points correspond to VAF for each individual patient and assay. The Wilcoxon signed rank test was used to compare the difference between matched 1‐h and 24‐h samples relative to the difference between matched 1‐h and 96‐h samples. *P = 0.003.

Figure 5

Variant copy numbers of 11 patients for different preanalytical conditions. Data points correspond to variant copy numbers for each individual patient and assay. The Wilcoxon signed rank test was used to compare the difference between matched 1‐h and 24‐h samples relative to the difference between matched 1‐h and 96‐h samples.