Skip to main content

Table 2 Select Examples ctDNA studies in GI cancers

From: The current state of molecular profiling in gastrointestinal malignancies

Gene

Tumor

Study type and details

Sample #

Assay utilized

Findings

Reference

RAS

CRC

Prospective cohort

98 pts

BEAMing expanded RAS mutation panel

RAS mut status was evaluable in plasma

RAS plasma-tissue concordance was 91.8%

ctDNA MAF was associated with clinical stage

RAS testing using BEAMing was comparable to tissue testing

[125]

RAS

mCRC

Prospective

236 pts

OncoBEAM

RAS plasma-tissue concordance was 92%

Plasma false-negatives were more frequent in lung-metastases-only disease

[126]

RAS

mCRC

Prospective

280 pts

OncoBEAM

RAS plasma-tissue concordance was 86.4%, positive percent agreement was 82.1%, and negative percent agreement was 90.4%

Lung-metastases-only disease was associated with more discordance (concordance rate of 64.5%), but concordance improved with larger tumor burden

[127]

RAS/EGFR/BRAF

mCRC

Retrospective cohort

Plasma samples from RAS/EGFR/BRAF wt mCRC pts analyzed after PD on anti-EGFR therapy

135 pts

496 pts in the validation cohort

Guardant 360 NGS

RAS and EGFR mut clones decrease exponentially from the time of EGFRi discontinuation

Confirmed ECD EGFR mutation as a potential driver for EGFRi resistance

Identified half-life of RAS/EGFR clones to guide the timing of re-challenge to therapy

[128]

RAS/EGFR/BRAF

mCRC

Phase II

EGFRi therapy rechallenge after prior PD guided by RAS/BRAF/EGFR status in ctDNA

27 pts

ddPCR and NGS

69% of screened pts were wt

First interventional trial of liquid biopsy—ctDNA molecular selection—driving EGFRi rechallenge in mCRC

RAS/BRAF/EGFR wt pts rechallenged to EGFRi demonstrated 30% ORR, 40% SD, 59% DCR with SD > 4 months, mPFS 16 weeks

[129]

CHRONOS

RAS/BRAF

mCRC

Prospective

72 pts

Idylla Biocartis

RAS/BRAF plasma-tissue concordance was 81.94%, with higher concordance in liver metastases cases

Emerging KRAS mutations were identified in 33% of pts treated with EGFRi

[130]

RAS/BRAF

mCRC

Prospective

Plasma mutational testing prospective series

278 pts

OncoBEAM

RAS/BRAF plasma-tissue concordance in chemotherapy naïve pts with liver metastases was 91.8%

Supports ctDNA as a surrogate marker to tissue testing for RAS and BRAF status

[131]

ColoBEAM

RAS/BRAF

mCRC

Prospective, non-randomized

EGFRi therapy rechallenge after prior PD guided by RAS/BRAF status in ctDNA

22 pts

PyroMark Q24 MDx Workstation, Genetic Analyzer ABI3130, Idylla RT-PCR, QZ200 System ddPCR

70% of enrolled pts were wt

Wt pts who underwent rechallenge experienced 27% ORR, 55% DCR, 7-month mOS, and 3-month mPFS

Rechallenge strategy is feasible with molecular selection through ctDNA

[132]

RAS

mCRC

Prospective, phase II, single-arm

EGFRi therapy rechallenge in 3rd line setting after prior PD

28 pts

ddPCR and Ion Torrent S5 XL ultra-deep NGS

48% samples at rechallenge baseline were RAS mut

wt patients experienced 21% ORR, 32% SD, 54% DCR, mPFS 3.4 mo, and mOS 9.8 mo

ctDNA RAS status predicted for responses to EGFRi. RAS wt was associated with a longer mPFS compared to RAS mut (4.0 vs. 1.9 mo, p = 0.03), with a trend toward longer mOS (12.5 vs. 5.2 mo, p = 0.24)

[133]

CRICKET

RAS

mCRC

Retrospective post-hoc biomarker study (pts from JACCRO CC-08 and 09)

16 pts

OncoBEAM Ras CRC

38% pts at rechallenge were RAS mut

DCR was better in RAS wt compared to RAS mut pts (80% vs. 33%, respectively)

mPFS in RAS mut vs. RAS wt pts was 2.3 vs. 4.7 mo (p = 0.0042) and mOS 3.7 vs 16 mo (p = 0.0002), respectively

ctDNA RAS status was significantly associated with clinical outcomes in pts receiving EGFRi rechallenge

[134]

BRAF

mCRC

Retrospective

64 pts

MD Anderson/GuardantHealth LB70 NGS

BRAF V600E plasma-tissue concordance was 80%

Lower BRAF V600E VAF was associated with acquired resistance to EGFRi

ctDNA to detect BRAF V600E mut is feasible

[135]

Multiple genes

mGEC

Cohort study

26 pts

28 pts

Guardant 360

Demonstrated genomic heterogeneity within the primary tumor and in disseminated disease

Found discordance between primary tumor and metastases in 36% of patients and high concordance between metastases and ctDNA (85%)

ctDNA profiling may enhance selection of therapy by identifying heterogeneous mutation profiles

[124]

HER2

GC

Prospective

Biomarker study in pts treated with neoadjuvant capecitabine + oxaliplatin + lapatinib in HER2 + GC

32 pts

Guardant 360

Plasma ERBB2 amplification predicted for chemotherapy + lapatinib responses

Changes in plasma ERBB2 copy number were associated with responses to therapy

Plasma genomics at the time of PD revealed emergence of MYC, EGFR, FGFR2, and MET amplification

Targeting MET kinase alongside HER2 in PDX model tumor that progressed on afatinib and had MET amplification resulted in tumor regression

[136]

HER2

mGEC

Biomarker analysis from phase II (NCT02954536)

HER2 positive mGEC treated with trastuzumab and chemotherapy in first-line setting

25 pts

Guardant 360 NGS

Baseline ctDNA ERBB2 amp and decreasing VAF on treatment were predictive biomarkers for response to HER2-directed therapy

[137]

HER2

mGC

Biomarker analysis from DESTINY-Gastric01

mGC pts treated with T-DXd

151 pts

GuardantOMNI

ORR in pts with baseline plasma ERBB2 amp was 60.6% and in pts without amp was 34.2%

ORR in pts with baseline plasma ERBB2 copy number above 6 had ORR 75.8% compared to 40.8% in those below 6.0

[138]

HER2

mCRC

Prospective, phase II

Pertuzumab plus trastuzumab in mCRC (refractory/intolerant to chemotherapy and EGFRi) with RAS wt and HER2 amp by tumor or ctDNA analysis

30 pts

Guardant 360 NGS

ORR was 30% in tissue HER2 amp pts and 28% in ctDNA HER2 amp pts

Pts with tissue + /ctDNA– HER2 amp had significantly lower ctDNA fraction compared to tissue + /ctDNA + HER2 amp pts, likely due to low tumor shedding

Baseline alterations in resistance pathways RTK/RAS/PI3K were enriched in non-responders and more frequently identified by ctDNA (67%) compared to tissue (19%) testing

Baseline ctDNA profiles predicted those who would benefit from pertuzumab plus trastuzumab

Decreasing ctDNA fraction on tx was associated with superior PFS and radiographic response

ctDNA identified an actionable new alteration in 62% of pts after PD

[139]

TRIUMPH

FGFR2

mGC

Retrospective

365 pts

Guardant 360 NGS and Illumina NextSeq 550

FGFR2 amp were detected more frequently with ctDNA than with tissue analysis (7.7% vs. 2.6–4.4%, respectively)

2 pts with FGFR2 amp by ctDNA after PD but not on pretreatment tissue analysis responded to FGFR inhibitors

[140]

FGFR2

CCA

Biomarker analysis from 3 pts enrolled in phase II study with infigratinib in FGFR mut CCA

3 pts

Guardant 360 NGS

ctDNA testing at the time of progression on infigratinib revealed new FGFR2 point mutations (resistance mechanisms)

ctDNA can reveal heterogeneous concurrent resistance mutations, unlike individual tissue biopsies

[141]

FGFR2

CCA

Retrospective

137 pts

Tempus xF liquid biopsy

ctDNA identified more actionable mutations in liquid biopsies (33.1%) compared to tissue biopsies (23.2%)

Prevalence of FGFR2 fusions was higher in liquid biopsies (11.3%) than tissue biopsies (3.4%)

ctDNA may be used to guide therapy selection

[142]

Multiple genes

CCA

Prospective

24 pts

15-gene and 710 gene oncopanel

Plasma-tissue concordance was 74% (higher at 94% in intrahepatic CCA and lower at 55% in extrahepatic CCA)

Baseline ctDNA VAF correlated with initial tumor loads

Baseline ctDNA VAF correlated with PFS in intrahepatic CCA

[143]

IDH1

CCA

Biomarker analysis from ClarIDHy

210 samples

BEAMing digital PCR test

IDH1 mut plasma-tissue concordance was 92%

A subset of pts with longer PFS on treatment with ivosidenib had plasma IDH1 mut clearance

[144]

KRAS

PDAC

Prospective

78 pts

ddPCR

Longitudinal ctDNA KRAS status was prognostic and predictive of responses to chemotherapy

[145]

KRAS

PDAC

Prospective

194 pts

ddPCR

KRAS plasma-tissue concordance was over 95%

ctDNA detection was prognostic for survival

[146]

Multiple genes

PDAC

Prospective

77 pts

Guardant 360 NGS

Baseline ctDNA mutations included TP53 (12.6%), KRAS (9.7%), MET (6.8%), ARID1A (4.8%), NF1 (4.8%), and others < 3%

ctDNA levels of TP53 and KRAS were associated with radiographic responses

New TP53 subclonal variant mutations were the most common resistance mutations in progressions (75%)

[147]

Multiple genes (mostly KRAS)

PDAC

Systematic review and meta-analysis

2326 pts

various

ctDNA muts and high concentrations of ctDNA are prognostic for survival (PFS and OS)

[148]

  1. Amp, amplification; BEAM, beads, emulsion, amplification, and magnetics; CCA, cholangiocarcinoma; CRC, colorectal cancer; ctDNA, circulating tumor DNA; DCR, disease control rate; ddPCR, digital droplet polymerase chain reaction; EGFRi, EGFR inhibitor; MAF, mean allele frequency; mCRC, metastatic colorectal cancer; mGC, metastatic gastric cancer; mGEC, metastatic gastric and esophageal cancer; mo, months; mut, mutated; NGS, next-generation sequencing; ORR, overall response rate; (m)OS, (median) overall survival; PDAC, pancreatic ductal adenocarcinoma; PDX, patient-derived xenograft; (m)PFS, (median) progression-free survival; PD, progression of disease; SD, stable disease; T-DXd, trastuzumab deruxtecan-nxki; VAF, variant allele frequency; WT, wild-type