Genomic temporal heterogeneity of circulating tumour DNA in unresectable metastatic colorectal cancer under first-line treatment

Abstract

Objective: Circulating tumour DNA (ctDNA) sequencing is increasingly used in the clinical management of patients with colorectal cancer. However, the genomic heterogeneity in ctDNA during treatments and its impact on clinical outcomes remain largely unknown.

Design: We conducted a prospective cohort study (NCT04228614) of 171 patients with unresectable metastatic colorectal cancer (mCRC) who underwent first-line treatment and prospectively collected blood samples with or without tumour samples from patients at baseline and sequentially until disease progression or last follow-up.

Results: The RAS/BRAF alterations in paired baseline tissue and plasma samples from 63 patients displayed a favourable concordance (81.0%, 51/63). After a period of first-line treatment (median time between baseline and last liquid biopsy, 4.67 months), 42.6% (26/61) of RAS-mutant patients showed RAS clearance and 50.0% (5/10) of BRAF-mutant patients showed BRAF clearance, while 3.6% (3/84) and 0.7% (1/135) of patients showed new RAS or BRAF mutations in ctDNA. Patients with plasma RAS/BRAF clearance showed similar progression-free survival (PFS) and overall survival (OS) with patients who remained RAS/BRAF wild-type, while much better outcomes than those who remained RAS/BRAF mutant. Patients who gained new RAS/BRAF mutations showed similar prognosis as those who maintained RAS/BRAF mutations, and shorter PFS and OS than those who remained RAS/BRAF wild-type.

Conclusion: This prospective, serial and large-scale ctDNA profiling study reveals the temporal heterogeneity of mCRC-related somatic variants, which should be given special attention in clinical practice, as evidenced by the finding that the shift in plasma RAS/BRAF mutational status can yield a drastic change in survival outcomes.

Keywords: cancer genetics; colorectal cancer genes; colorectal carcinoma.

Figure 1

Flowchart of study design and patient selection. ctDNA, circulating tumour DNA; mCRC, metastatic colorectal cancer; NED, no evidence of disease; WES, whole-exome sequencing.

Figure 2

Concordance analysis of baseline mutations in paired plasma and tumour tissue samples among 63 patients. (A) Genomic profiling of some high-frequency mutations between baseline tissue samples and plasma samples (nonsynonymous single-nucleotide variants and indels). The top bar represents the number of mutations a patient carried; the side bar represents the number of patients who carried a certain mutation. (B) Correlation analysis between mutation frequencies of 378 genes from the NGS panel in circulating tumour DNA (ctDNA) samples versus tissue samples (Spearman’s rank correlation). (C) Comparison of RAS and BRAF^V600E mutations in tissue samples and plasma samples. MUT, mutant; P, plasma; T, tissue; WT, wild type.

Figure 3

Genomic temporal heterogeneity in plasma circulating tumour DNA under first-line treatment among 145 patients. (A) Evidence-based actionable targets in colorectal cancer (CRC) referring to the Oncology Knowledge Base database. The outer ring represents the different levels of evidence for the actionable targets, and the inner part represents the different treatments according to the corresponding targets. (B) Time from baseline plasma sample collection to the last liquid biopsy before progressive disease (PD). (C) Genomic profiling of the most commonly mutated genes and actionable targets in CRC between baseline and post-chemotherapy (postCT) plasma samples. The top bar represents the number of mutations a patient carried; the side bar represents the number of patients who carried a certain mutation; and the bottom bar represents patient characteristics, including age, sex, smoking history, tumour location, metastatic site at baseline, best response and first-line chemotherapy regimen. (D) The clearance and acquisition rates of standard-of-care targets and top mutant genes after treatment. (E) The shift of actionable targets for clinical trials after treatment. amp, amplification; BL, baseline; CR, complete response; MSI, microsatellite instability; NE, not evaluable; PR, partial response; SD, stable disease; TMB, tumour mutation burden.

Figure 4

Kaplan-Meier estimates of progression-free survival (PFS) (A) and overall survival (OS) (B) in patients stratified according to different changes in plasma RAS status under first-line treatment Statistical significance was determined by Wald test of the multivariable Cox models. The change in the circulating tumour DNA (ctDNA) fraction of cfDNA, estimated by maximum somatic allele frequency, was included as a variable. mPFS, median progression-free survival; mOS, median overall survival; MUT, mutant; ref, reference; WT, wild-type.

Figure 5

Genomic temporal heterogeneity under a period of first-line regimen after progressive disease among 20 patients. (A) Genomic profiling of the top mutant genes and actionable targets in colorectal cancer between baseline and progressive disease plasma samples. The top bar represents the number of mutations a patient carried; the side bar represents the number of patients who carried a certain mutation; and the bottom bar represents patient characteristics, including age, sex, smoking history, tumour location, metastatic site at baseline, best response and first-line chemotherapy regimen. (B) The prevalence of clearance and acquisition of standard-of-care targets and the top mutant genes after progression. (C) The shift of actionable targets for clinical trials after progression. CR, complete response; PD, progressive disease; PR, partial response; SD, stable disease.