A Review of Circulating Tumor DNA (ctDNA) and the Liquid Biopsy in Cancer Diagnosis, Screening, and Monitoring Treatment Response

Abstract

The concept of liquid biopsy is based on the knowledge that blood or secretions from the body contain tumor cells, nucleic acids, cellular components, and tumor metabolites. Detection of circulating DNA (ctDNA) in liquid biopsy material shows the most promise due to the advances in DNA technologies that have made detection and sample screening possible. Clinical trials have begun to evaluate ctDNA monitoring for response to cancer treatment in clinical settings for non-small cell lung cancer (NSCLC), breast cancer, and colorectal cancer (CRC). However, most liquid biopsy tests introduced into the clinic have only been able to identify one or two features of tumor DNA, which limits the specificity of the test. In early 2025, a study led by a research team at Oxford University identified a new blood test, TriOx, developed using machine learning to detect multiple types of cancer at an early stage. These new blood tests may revolutionize oncology and make early cancer detection as routine as other diagnostic blood tests, such as blood glucose testing. This article aims to review ctDNA and liquid biopsy in the diagnosis, early detection, and monitoring of treatment response in cancer.

Introduction

The gold standard in cancer diagnosis is tissue biopsy, which allows for the typing and grading of the tumor cells and the identification of target expression for targeted therapies [1,2]. There is a need for less invasive early cancer diagnosis and techniques that can evaluate prognosis and treatment response by detecting cancer recurrence [3,4]. The tissue biopsy is an invasive procedure that is not suitable for early detection or continuous monitoring of cancer progression and recurrence [5].

The concept of liquid biopsy is based on the knowledge that blood or secretions from the body contain tumor cells, nucleic acids, cellular components, and metabolites from tumors (Table 1) [6]. The liquid biopsy may contain circulating tumor cells (CTCs), tumor-derived extracellular vesicles (EVs), tumor-educated platelets (TEPs), circulating free RNA (cfRNA), and circulating tumor DNA (ctDNA) [4]. CTCs are tumor cells that have been shed from the primary tumor and EVs include nano-particles with a lipid bilayer membrane that have central roles in tumor invasion and metastasis [4]. However, ctDNA consists of small fragments of DNA released by tumor cells into the blood and tissue fluids [7].

Table 1.

Biofluids used for circulating DNA (ctDNA) analysis (the liquid biopsy) [6].

Plasma and serum from blood

Pleural fluid

Bronchoalveolar lavage (BAL) fluid

Peritoneal fluid

Cerebrospinal fluid (CSF)

Lymphatic fluid

Urine

Breast milk

Saliva

Tears, aqueous and vitreous fluid

Vaginal secretions and uterine lavage fluid

Seminal fluid

The detection of ctDNA in liquid biopsy material shows the most promise due to the advances in DNA technologies that have made detection and sample screening possible. In 1977, Leon and colleagues identified that plasma levels of free DNA increased in cancer patients [8]. In 2008, Diehl and colleagues studied ctDNA of 18 patients with colorectal cancer and identified mutations in several genes, including KRAS and PIK3CA [9]. They also showed that the rate of ctDNA mutations changed following cancer treatment and correlated with both carcinoembryonic antigen (CEA) concentrations and tumor volume [9]. In 2014, the use of ctDNA to identify EGFR mutations in non-small cell lung cancer (NSCLC) was authorized by the European Medicines Agency (EMA) [10]. In 2021, the International Association for the Study of Lung Cancer (IASLC) issued a consensus statement on using liquid biopsy in advanced NSCLC [11]. In 2022, the European Society for Medical Oncology (ESMO) Precision Medicine Working Group issued a report on the recommendations for circulating tumor DNA assays for patients with cancer [12].

Cell-free DNA (cfDNA) normally circulates in all people due to the effects of cell proliferation, apoptosis, and other normal cellular physiological functions [13]. Liquid biopsy analysis of ctDNA involves detecting a tumor-specific characteristic such as somatic mutation, viral sequence, or methylation profile, distinguishing it from cfDNA of non-tumor origin [13]. Currently, several molecular technologies are used to detect ctDNA in liquid biopsy, which mainly includes polymerase chain reaction (PCR) and next-generation sequencing (NGS) techniques (Table 2) [12]. The PCR-based approaches include digital droplet PCR (ddPCR) and BEAMing (beads, emulsion, amplification, magnetics) to identify single or few well-characterized mutations suitable for targeted therapy [12]. Also, because cfDNA is released following cell death, cfDNA can indicate cancer treatment response [13]. The role of ctDNA has been investigated for cancer diagnosis and prognosis. However, a further important clinical application is to evaluate cancer treatment response and minimal residual disease (MRD) across various tumor types [13]. This article aims to review ctDNA and liquid biopsy in the diagnosis, early detection, and monitoring of treatment response in cancer.

Table 2.

Genomic regions and methods used for circulating DNA (ctDNA) analysis in cancer patients (tumor-informed or non-informed) [12].

Type of analysis	Genomic profiling Epigenetic profiling Fragmentomic profiling Multimodal analysis
Genomic coverage	Whole genome Whole exome
Analytical methods	Polymerase chain reaction (PCR)-based, including digital droplet PCR (ddPCR) and BEAMing (beads, emulsion, amplification, magnetics) Next-generation sequencing (NGS)-based Methylomics-based Fragmentomics-based

Methods Used for ctDNA Analysis

The detection methods for ctDNA require high sensitivity due to the low levels of tumor-specific DNA in the circulation. Tumor-specific somatic mutations and other genomic alterations are mainly analyzed. However, driver mutations in ctDNA are indicators of disease burden, which can also be assessed by tracking through time [13]. Therefore, treatment response analysis can also be done using ctDNA analysis and clearance after treatment [13].

Polymerase Chain Reaction (PCR)-Based Methods

Targeted ctDNA analysis includes PCR methods, which can rapidly detect mutations with high sensitivity and a rapid turnaround, including quantitative PCR (qPCR), digital (d)PCR, and BEAMing (beads, emulsion, amplification, and magnetics) [14]. There are also commonly mutated genes in specific cancers, including the BRAF gene in melanoma, KRAS in lung and colorectal cancer (CRC), PIK3CA in breast cancer, and androgen receptor (AR) in prostate cancer, which are tumor targets that can be monitored for each assay [14].

Next-Generation Sequencing (NGS) Methods

NGS methods have recently developed to detect a broad range of genomic alterations and include whole-exome sequencing (WES) and whole-genome sequencing (WGS), targeted methods including tagged-amplicon deep sequencing (TAm-Seq), CAncer Personalized Profiling by deep Sequencing (CAPP-Seq), and targeted error correction sequencing (TEC-Seq) [15]. Because heterogeneous cancers have high genomic instability, these advanced methods have shown some advantages, but sequencing artifacts can occur [13]. The development of more sensitive NGS methodologies has begun to overcome the limitations of low levels of ctDNA in early-stage cancers or low-shedding tumors [14]. In 2024, Martin-Alonso and colleagues proposed using priming agents to transiently reduce cfDNA clearance in vivo to overcome the low levels of ctDNA in the circulation, which may be a future approach to improving the detection of ctDNA [16]. Currently, sequencing methods still face challenges associated with cost and longer analysis times [15]. There are also non-mutation-based methods to undertake ctDNA monitoring and detection of viral DNA for human papillomavirus (HPV) in cases of oropharyngeal and cervical carcinoma and hepatitis B virus (HBV) in cases of hepatocellular carcinoma (HCC) [13].

Methylomics Methods

DNA methylation and fragmentome analysis are used for ctDNA analysis to overcome the limitations of genomic ctDNA studies. DNA methylation analysis has been done using bisulfite conversion and analytical methods that include whole genome bisulfite sequencing (WGBS) and targeted bisulfite sequencing for longitudinal monitoring of cancer patients [13]. Recently, bisulfite-free methods have been developed to overcome the limitations of DNA degradation caused by bisulfite conversion, including chromatin immunoprecipitation sequencing (ChIP-Seq) and methylated DNA immunoprecipitation sequencing (MeDIP-Seq) [14,17].

Fragmentomics Methods

Fragmentomics of cfDNA is a new field in liquid biopsy diagnosis, which includes fragmentation patterns, fragment sizes, and end characteristics [14,18]. Cancer patients show more diverse fragmentation patterns that can be used to distinguish cancer from non-cancer-derived cfDNA [9]. Novel bioinformatics methods have resulted in more detailed fragmentomic data [19]. In 2019, Cristiano and colleagues developed a new method for genome-wide analysis of cfDNA fragmentation patterns using a low-coverage WGS method called DELFI (DNA evaluation of fragments for early interception) [20]. This machine learning model incorporates genome-wide fragmentation profiles and can be combined with mutation-based cfDNA analyses, which results in a sensitivity of cancer detection of 91% [20].

Multimodal Methods

Multimodal combinations of analyses are used, including analysis of copy number alterations (CNA) and epigenetic, genomic, and fragmentomic analyses of cfDNA samples [13]. In 2021, Parikh and colleagues showed that the integration of epigenomic signatures increased the sensitivity for detection of recurrence by 25–36% when compared with genomic alterations alone [21]. Currently, the range of biofluids for liquid biopsy analysis and detection of ctDNA in response monitoring has gone beyond serum and plasma to include urine, saliva, and cerebrospinal fluid (CSF) (Table 1) [13].

Limitations of ctDNA Analysis in Clinical Practice

Currently, the main clinical role of ctDNA is as a tool for monitoring treatment response in patients with a history of cancer, including lung, colorectal, and breast cancer [13]. However, several limitations to this liquid biopsy approach have prevented its widespread clinical use. There is still a lack of liquid biopsy sample collection and analysis standardization, and clinical and laboratory guidelines for optimal time points for taking samples [22]. Also, it is still unclear what the optimal sampling times after cancer treatment best predict clinical relapse [22]. These limitations may also affect the results of ongoing clinical trials. Also, ctDNA analysis of liquid biopsies may be limited by low levels of DNA due to fragment degradation on sampling and storage [22]. Importantly, there will be potential confounding DNA data from patient comorbidities, particularly chronic diseases, inflammatory diseases, and diseases that involve cell proliferation [22]. These limitations highlight the importance of establishing standardized protocols for ctDNA analysis in patients with various cancer types [13,22].

However, clinical trials have begun to evaluate ctDNA monitoring for response to cancer treatment in clinical settings. Table 3 summarizes the use of ctDNA to diagnose and monitor treatment response in common solid cancers, including non-small cell lung cancer (NSCLC), breast cancer, and colorectal cancer (CRC) in some reported clinical trials [23–37]. Therefore, ctDNA analysis from liquid biopsies has the potential for cancer diagnosis, selection of targeted therapies, treatment modification, and response to treatment.

Table 3.

Completed clinical trials in 2025 of the role of circulating tumor DNA (ctDNA) in solid tumor diagnosis and monitoring [23–37].

Study/Clinical trial	Aims	No. of patients	ctDNA methods	Results
Non-Small Cell Lung Cancer (NSCLC)
BFAST (NCT03178552) [23,24]	Relationship between blood-based genetic changes and targeted therapies in advanced NSCLC	2219	NGS	The application of blood-based NGS informed clinical decision-making
Liquid-Lung-A (NCT0262952) [25]	Efficacy of afatinib in treatment-naïve NSCLC patients with EGFR exon 19 deletions or exon 21 point mutations	331	PCR (PANA Mutyper R EGFR assay)	Afatinib showed similar ORR and PFS in patients with lung cancer and EGFR mutations. Survival benefit of afatinib treatment achieved
Liquid-Lung-O (NCT0276928) [26]	Treatment efficacy of osimertinib in patients with NSCLC with activating EGFR mutations (cohort 1) or T790M EGFR mutations (cohort 2) detected by ctDNA	119	PCR (PANA Mutyper R EGFR assay and Cobas EGFR Mutation Test v2)	Cohort 1: Osimertinib had favorable outcomes in first-line treatment of metastatic NSCLC Cohort 2: Osimertinib had favorable outcomes in patients with NSCLC with T790M EGFR mutations
APPLE (NCT0285689) [27]	Feasibility of using longitudinal plasma EGFR T790M monitoring to determine treatment (gefitinib and osimertinib)	103	PCR (Cobas EGFR mutation test v2)	Serial monitoring of ctDNA T790M status was feasible and lead to the identification of molecular progression
ACCELERATE (NCT0486392) [28]	Association between ctDNA genotyping before tissue diagnosis and time to treatment	150	NGS (InVisionFirst-Lung)	Use of plasma ctDNA genotyping before tissue diagnosis was associated with accelerated time to treatment
LOCAL (NCT03046316) [29]	Feasibility of de-escalation of TKI treatment guided by ctDNA to achieve complete remission after local consolidative therapy	60	NGS (oncoMRD-B panel of 338 genes (GenePlus)	Overall, a ctDNA-guided adaptive de-escalation TKI treatment strategy was feasible for patients with advanced NSCLC
Breast Cancer
PlasmaMATCH (NCT03182634) [30]	To determine the ability of ctDNA testing to select patients for mutation-directed therapy	1034	ddPCR and NGS (Guardant360)	Accurate genotyping identified mutation-specific treatments for breast cancer, including targeted therapies for uncommon HER2 and AKT1 mutations
PADA-1 (NCT03079011) [31]	Effectiveness of early therapy change based on increasing ESR1 mutation in blood, and safety of combining fulvestrant and palbociclib	1017	Multiplex ddPCR	Early therapeutic targeting of ESR1 mutation in ER+/HER2-advanced breast cancer resulted in significant clinical benefit
ACTDNA (NCT05079074) [32]	Efficacy of re-subtyping and determining treatment strategy based on ctDNA alterations	223	NGS	Patients with druggable ctDNA alterations showed significant improvements in PFS and disease control rate
c-TRAK-TN (NCT03145961) [33]	Role of ctDNA in detecting residual disease following standard primary treatment for TNBC	208	ddPCR	High rate of metastatic disease on ctDNA detection, MRD detection, and personalized ctDNA assays clinically achievable
Colorectal Cancer (CRC)
ASPECCT (NCT01001377) [34]	To analyze EGFR mutations in cfDNA in plasma from patients with CRC treated with panitumumab	261	NGS	Genotyping identified that EGFR mutations were associated with poorer patient outcomes
CHRONOS (NCT0322792) [35]	To identify RAS/BRAF/EGFR mutations in ctDNA for a chemotherapy-free anti-EGFR re-challenge with panitumumab	52	NGS and ddPCR	ctDNA analysis was an effective, safe, and rapid method to guide anti-EGFR re-challenge therapy with panitumumab in patients with mCRC
DYNAMIC (ACTRN12615000381583) [36]	To assess whether a ctDNA-guided approach could reduce the use of ACT without compromising recurrence risk	455	NGS (Safe-SeqS)	A ctDNA-guided approach reduced ACT use without compromising recurrence-free survival
COBRA (NCT0406810) [37]	To evaluate whether positive ctDNA after resection could identify patients who will benefit from ACT	635	NGS (Guardant LUNAR assay)	No improvement in ctDNA clearance after 6 months of chemotherapy for patients with ctDNA detected following resection of stage IIA CRC

ACT – adjuvant chemotherapy; ACTRN – Australian New Zealand Clinical Trials Registry number; ctDNA – circulating tumor DNA; ddPCR – droplet digital PCR; ER – estrogen receptor; MBC – metastatic breast cancer; mCRC – metastatic colorectal cancer; MRD – minimal residual disease; NCT – National Clinical Trial (identifier number); NGS – next-generation sequencing; NSCLC – non-small cell lung cancer; OS – overall survival; PCR – polymerase chain reaction; PFS – progression-free survival; Safe-SeqS – Safe-Sequencing System; TNBC – triple-negative breast cancer; TKI – tyrosine kinase inhibitor; UMIN-CTR – University Hospital Medical Information Network Clinical Trials Registry.

Future Directions: Development of a Routine ctDNA Diagnostic for Cancer

Hopes for the application of liquid biopsy technologies as less invasive methods for cancer diagnosis and early detection (screening) have driven recent research. However, most liquid biopsy tests introduced into the clinic have only been able to identify one or two features of tumor DNA, which limits the specificity of the test. In early 2025, a study led by a research team at Oxford University identified a new blood test, developed using machine learning, to detect multiple types of cancer at an early stage [38]. This innovative test, named TriOx, analyses multiple features of DNA in the peripheral blood across six types of cancer (colorectal, oesophageal, pancreatic, renal, ovarian, and breast) and can distinguish between individuals with and without cancer [38]. The research team developed a new methodology for ctDNA detection using deep (80x) whole-genome TET-Assisted Pyridine Borane Sequencing (TAPS), which is a less destructive method than bisulfite sequencing [36]. TAPS, combined with machine learning, permits the simultaneous analysis of genomic and methylation modification data to analyze and combine multiple features from the ctDNA circulating in the blood to improve the detection rate for small fractions of DNA [38]. The research team tested diagnostic accuracy across multiple cancer types in patients with symptoms and matched tumor biopsies, which showed 94.9% sensitivity and 88.8% specificity [38]. Further in silico validation showed strong discrimination at ctDNA fractions, which were as low as 0.7% [38]. An important part of this study showed that tumor burden and ctDNA shedding could be tracked from early or pre-malignant lesions following treatment and without requiring matched tumor biopsies [38]. TriOx is still in the development phase, but this study has shown that blood-based early cancer detection is possible [38].

Conclusions

Because there are still many gaps in understanding the fundamental origins of and dynamics of tumor-derived molecules and DNA, further research must address these gaps before standardizing ctDNA analysis and interpretation can lead to improved clinical utility in different types of cancer as part of a non-invasive method for diagnosis, prognosis, and treatment response in a personalized medicine approach in oncology. Recent research, as shown by TriOx, has highlighted the possibility that blood tests for ctDNA may revolutionize cancer screening, monitoring, and diagnosis to make early cancer detection as routine as other diagnostic blood tests, such as blood glucose testing.