The effect of surgical trauma on circulating free DNA levels in cancer patients-implications for studies of circulating tumor DNA

Abstract

Detection of circulating tumor DNA (ctDNA) post-treatment is an emerging marker of residual disease. ctDNA constitutes only a minor fraction of the cell-free DNA (cfDNA) circulating in cancer patients, complicating ctDNA detection. This is exacerbated by trauma-induced cfDNA. To guide optimal blood sample timing, we investigated the duration and magnitude of surgical trauma-induced cfDNA in patients with colorectal or bladder cancer. DNA levels were quantified in paired plasma samples collected before and up to 6 weeks after surgery from 436 patients with colorectal cancer and 47 patients with muscle-invasive bladder cancer. To assess whether trauma-induced cfDNA fragments are longer than ordinary cfDNA fragments, the concentration of short (< 1 kb) and long (> 1 kb) fragments was determined for 91 patients. Previously reported ctDNA data from 91 patients with colorectal cancer and 47 patients with bladder cancer were used to assess how trauma-induced DNA affects ctDNA detection. The total cfDNA level increased postoperatively-both in patients with colorectal cancer (mean threefold) and bladder cancer (mean eightfold). The DNA levels were significantly increased up to 4 weeks after surgery in both patient cohorts (P = 0.0005 and P ≤ 0.0001). The concentration of short, but not long, cfDNA fragments increased postoperatively. Of 25 patients with radiological relapse, eight were ctDNA-positive and 17 were ctDNA-negative in the period with trauma-induced DNA. Analysis of longitudinal samples revealed that five of the negative patients became positive shortly after the release of trauma-induced cfDNA had ceased. In conclusion, surgery was associated with elevated cfDNA levels, persisting up to 4 weeks, which may have masked ctDNA in relapse patients. Trauma-induced cfDNA was of similar size to ordinary cfDNA. To mitigate the impact of trauma-induced cfDNA on ctDNA detection, it is recommended that a second blood sample collected after week 4 is analyzed for patients initially ctDNA negative.

Keywords: bladder cancer; cell-free DNA; circulating tumor DNA; colorectal cancer; trauma.

Fig. 1

Changes in cfDNA concentration after surgery. The median changes in cfDNA concentration relative to preoperative levels are shown for: (A) total cfDNA in samples from CRC patients (N = 436); (B) total cfDNA in samples from MIBC patients (N = 47); (C) short cfDNA fragments (N = 91); (D) long cfDNA fragments (N = 91) at different times after surgery. Plasma samples were organized in weekly bins based on the postoperative day the blood was collected. To account for varying interpatient baseline cfDNA concentrations, the postoperative cfDNA levels were normalized to the preoperative level for each patient. Error bars indicate median absolute deviation. Time after surgery depicts the mean time for all samples within that time bin. Lines drawn between data points are meant as guides and do not signify longitudinal samples. Statistical significance (P ≤ 0.05) in fold concentration change calculated by the Wilcoxon signed rank test is indicated by ‘*’. OP, time of operation.

Fig. 2

ctDNA detection in samples with elevated cfDNA levels. (A) Theoretical likelihood of ctDNA detection in CRC samples (N = 436), assuming three ctDNA molecules per millilitre of plasma. The likelihood of ctDNA detection is simulated based on different minimum detectable VAF cutoffs. Plasma samples were sorted in weekly bins based on the postoperative day the blood was collected. (B) Fold increase in cfDNA concentration and absolute ctDNA VAF in CRC (N = 1) and MIBC (N = 4) patients over time. All analyzed postoperative plasma samples until radiological detection of disease recurrence is shown.