Frequent post-operative monitoring of colorectal cancer using individualised ctDNA validated by multiregional molecular profiling

Abstract

Background Circulating tumour DNA (ctDNA) is known as a tumour-specific personalised biomarker, but the mutation-selection criteria from heterogeneous tumours remain a challenge. Methods We conducted multiregional sequencing of 42 specimens from 14 colorectal tumours of 12 patients, including two double-cancer cases, to identify mutational heterogeneity to develop personalised ctDNA assays using 175 plasma samples. Results “Founder” mutations, defined as a mutation that is present in all regions of the tumour in a binary manner (i.e., present or absent), were identified in 12/14 tumours. In contrast, “truncal” mutations, which are the first mutation that occurs prior to the divergence of branches in the phylogenetic tree using variant allele frequency (VAF) as continuous variables, were identified in 12/14 tumours. Two tumours without founder and truncal mutations were hypermutators. Most founder and truncal mutations exhibited higher VAFs than “non-founder” and “branch” mutations, resulting in a high chance to be detected in ctDNA. In post-operative long-term observation for 10/12 patients, early relapse prediction, treatment efficacy and non-relapse corroboration were achievable from frequent ctDNA monitoring. Conclusions A single biopsy is sufficient to develop custom dPCR probes for monitoring tumour burden in most CRC patients. However, it may not be effective for those with hypermutated tumours.

Impact of sensitive circulating tumor DNA monitoring on CT scan intervals during postoperative colorectal cancer surveillance

“…Supplementary Tables 3 and 4 and Supplementary Figure 1 summarize the mutation profile; the detailed mutation profile for Set 1 is available in our previous report. 33 A total of 109 mutations [3 mutations per sample (range, 1-24) on average] and 85 mutations [3 mutations per sample (range, 1-36) on average] were identified using the customized panel for CRC for Sets 2 and 3, respectively. Patient CC16024 in Set 2 and Patient CC16054 in Set 3 had hypermutations with 24 (212.8/Mb) and 36 (319.1/Mb) mutations, respectively.…”

“…The dPCR assay for quantitative monitoring of ctDNA levels was performed as described previously. [32][33][34][35] Briefly, specific primers and probes labeled for wild-type and mutant alleles were specifically designed for each mutation identified in a primary tumor, using Hypercool Primer & Probe™ technology (Nihon Gene Research Laboratories, Sendai, Japan). For frequently recurring missense mutations, commercially available primer/probe sets were used (Thermo Fisher Scientific, Waltham, MA and Quantdetect, Inc, Tokyo, Japan).…”

“…Briefly, Set 1 was analyzed using the ClearSeq Comprehensive Cancer Panel (Agilent Technologies, Inc., Santa Clara, CA) targeting 151 disease-associated genes on an Illumina Hiseq 2000 sequencer (Illumina, Inc., San Diego, CA). 33 Set 2 and Set 3 were analyzed by the Ion Proton™ and the Ion S5™ system (Thermo Fisher Scientific, Waltham, MA), respectively, using a customized All rights reserved. No reuse allowed without permission.…”

“…Comprehensive descriptions of methodologies used to identify somatic mutations are presented in the Supplementary Methods. 33,35 Monitoring ctDNA levels using dPCR…”

“…Supplementary Tables 3 and 4 and Supplementary Figure 1 summarize the mutation profile; the detailed mutation profile for Set 1 is available in our previous report. 33 A total of 109 mutations Among the three different sequence platforms, the most frequently mutated genes were TP53 (37/52, 71.2%), APC (28/52, 53.8%), KRAS (24/52, 46.2%), PIK3CA (13/52, 25.0%), and BRAF (8/52, 15.4%). The average VAFs for these five genes were: TP53, 44.8% (range, 7.78%-85.9%), APC, 33.3% (range, 12.0%-61.0%), KRAS, 31.1% (range, 3.6%-81.7%), PIK3CA, 28.5% (range, 8.0%-40.5%), and BRAF, 28.1% (range, 12.9%-48.2%).…”