Can a Blood Test Detect Cancer? Circulating Tumour DNA and Liquid Biopsy Explained
Can a blood test really reveal the presence of cancer? In some situations, yes — but not because a laboratory can somehow “see” a tumour in a tube of blood.
Instead, certain tests look for molecular traces released by cancer cells. One of the most important is circulating tumour DNA, usually shortened to ctDNA.
This approach is central to what is known as a liquid biopsy. It can already provide clinically useful information about some cancers, particularly by identifying genetic alterations that may influence treatment. Researchers are also studying whether ctDNA can help monitor treatment, identify molecular signs of recurrence and, eventually, detect cancer before symptoms appear.
That last possibility has attracted considerable attention. Multi-cancer blood tests are being developed to look for signs of several cancers at once in people who feel well.
However, detecting a cancer-related signal is not the same as having a universal cancer screening test. Evidence available in England in 2026 does not support replacing established NHS screening programmes with a single ctDNA blood test.
To understand why, it helps to start with two terms that are often confused: cfDNA and ctDNA.
What is circulating tumour DNA?
DNA is not permanently confined inside the nucleus of our cells. As cells die and are broken down, small fragments of their genetic material can enter the bloodstream.
These fragments are collectively called cell-free DNA, or cfDNA.
In healthy people, much of the cfDNA found in plasma originates from blood-forming cells. The concentration is usually low, and many fragments cluster around approximately 167 base pairs in length, reflecting the way DNA is packaged around proteins called histones.
Analysing cell-free DNA is not unique to cancer medicine. For example, non-invasive prenatal testing (NIPT/NIPS) examines fragments of placental DNA circulating in the pregnant woman's blood to estimate the risk of certain chromosomal abnormalities.
From cfDNA to ctDNA
When a person has cancer, some of the DNA circulating in their blood may come from tumour cells.
This tumour-derived fraction is called circulating tumour DNA, or ctDNA.
A primary tumour, metastases and different subpopulations of cancer cells may all release DNA fragments into the bloodstream. These fragments can contain molecular characteristics of the cancer, including:
genetic mutations;
copy-number alterations;
epigenetic changes;
abnormal methylation patterns;
other genomic features that can be detected by laboratory analysis.
Tumour-derived DNA also tends to include a greater proportion of shorter fragments than non-tumour cfDNA, with many ctDNA fragments found around 145–150 base pairs.
The more important problem, however, is quantity.
In some patients, ctDNA represents only a tiny fraction of all the cell-free DNA in a blood sample, sometimes well below 1%.
Finding that signal among a much larger background of normal DNA is one of the fundamental technical challenges of liquid biopsy.
Why do some cancers release more ctDNA than others?
The amount of tumour DNA circulating in the blood varies considerably between patients and cancers.
It may depend on factors such as:
tumour size;
cancer type;
tumour location;
vascularisation;
biological activity;
the presence of metastases;
treatment;
how quickly cancer cells are dying and being replaced;
when the blood sample is collected.
As a general rule, advanced and metastatic cancers are more likely to produce measurable amounts of ctDNA than very small, localised tumours.
This creates an important paradox.
The earlier a cancer is detected, the more useful early diagnosis may be — yet early cancers are often precisely the cancers that release the least ctDNA.
A small tumour may produce so few fragments that its molecular signal becomes lost among DNA released by normal cells.
This biological limitation is one of the main reasons why developing a reliable universal blood test for early cancer remains difficult.
How is circulating tumour DNA detected in a blood sample?
After blood is collected, the laboratory normally separates the plasma from the blood cells. Cell-free DNA can then be extracted from the plasma.
Because the amount of ctDNA may be extremely small, laboratories need highly sensitive techniques to distinguish genuine tumour-derived abnormalities from normal genetic material and technical noise.
Two of the most important methods are next-generation sequencing (NGS) and digital PCR.
Next-generation sequencing: searching for many alterations at once
Next-generation sequencing, or NGS, can analyse large numbers of DNA sequences simultaneously.
Instead of looking for one predetermined mutation, an NGS panel can examine dozens or hundreds of genes or genomic regions.
This makes the method particularly useful when the objective is to establish a broader molecular profile of a cancer.
For a patient with an established cancer diagnosis, doctors may use molecular testing to identify alterations associated with:
eligibility for a targeted treatment;
sensitivity to a particular drug;
resistance to a previous treatment;
potential clinical trial options.
The technical challenge is considerable. When an abnormal DNA molecule represents only a very small fraction of all DNA in the sample, the laboratory must distinguish a true cancer-associated variant from sequencing errors and unrelated biological signals.
Modern assays can therefore use very deep sequencing, specialised bioinformatics and techniques such as unique molecular identifiers to reduce error rates and improve confidence in low-frequency findings.
Pre-analytical conditions also matter. Blood collection, sample transport, processing time, plasma preparation and DNA extraction can all influence the final result.
Digital PCR: highly sensitive detection of a known mutation
Digital PCR, including techniques such as dPCR and droplet digital PCR (ddPCR), takes a different approach.
It is particularly useful when the laboratory already knows exactly which genetic alteration it wants to detect.
The sample is divided into a very large number of separate microscopic reactions. The laboratory can then determine how many contain the target mutation and estimate its abundance.
This can make digital PCR extremely sensitive for detecting or monitoring a predefined alteration present at a low concentration.
A simple way to understand the distinction is:
NGS asks: “Which relevant alterations can we find?”
Digital PCR asks: “Is this specific alteration present, and how much of it is there?”
The two technologies therefore address different laboratory and clinical questions.
Tumour-informed vs tumour-agnostic ctDNA testing
Not every ctDNA assay starts with the same information.
One important distinction is between tumour-informed and tumour-agnostic testing.
Tumour-informed testing
With a tumour-informed approach, the laboratory already knows the molecular profile of the patient's tumour.
For example, tissue removed during surgery or collected during a biopsy may reveal several mutations characteristic of that particular cancer.
A personalised blood test can then be designed to search for those same alterations in plasma.
Because the laboratory knows exactly what it is looking for, this strategy can improve the ability to distinguish tumour DNA from unrelated circulating DNA.
Tumour-informed assays are particularly important in research into molecular residual disease, where the aim is to identify very small quantities of cancer remaining after potentially curative treatment.
Tumour-agnostic testing
A tumour-agnostic assay does not require a previously analysed tumour sample.
Instead, it searches the blood directly for molecular patterns that may indicate cancer.
This is essential when no tumour has yet been characterised, as would be the case when trying to screen apparently healthy people.
Tumour-agnostic approaches may examine mutations, methylation patterns, fragmentation characteristics or combinations of different biological signals.
These strategies form an important part of research into multi-cancer early detection, often abbreviated to MCED.
Multi-cancer blood tests do not rely only on mutations
Looking for a rare mutation becomes particularly difficult when a tumour is very small and releases almost no DNA into the circulation.
Researchers have therefore developed methods that examine other characteristics of cell-free DNA.
One of the most important is DNA methylation.
What is DNA methylation?
Methylation involves the addition of small chemical groups to particular regions of DNA.
These chemical markers help regulate which genes are active or inactive. Cancer cells often develop methylation patterns that differ from those of healthy cells.
Instead of searching for one specific mutation, a test can analyse large numbers of methylation sites and look for a broader pattern associated with cancer.
Machine-learning algorithms may then attempt to answer two questions:
Is there evidence of a cancer-related signal?
If so, which tissue or organ is the signal most likely to come from?
This principle is used by several multi-cancer early detection tests.
What is liquid biopsy actually used for today?
A crucial distinction needs to be made between two very different applications:
using ctDNA in a patient who already has, or is strongly suspected of having, cancer;
screening an apparently healthy population for multiple cancers.
The same broad technology may be involved, but the level of clinical validation is very different.
Finding genetic alterations that can guide treatment
One of the most established uses of ctDNA is molecular profiling in patients with cancer.
If a cancer carries a particular genetic alteration, that information may determine whether a targeted medicine is appropriate.
Traditionally, these alterations are identified from tumour tissue. However, obtaining enough tissue is not always straightforward.
A tumour may be:
difficult to access;
unsafe to biopsy;
too small to provide sufficient material;
previously sampled but without enough usable DNA;
biologically different after months or years of treatment.
In these circumstances, a blood-based genomic test may provide useful additional information.
The ESMO recommendations on the clinical use of circulating tumour DNA recognise plasma ctDNA genotyping as clinically useful in selected patients with advanced cancer, while also stressing that negative results may require follow-up with tumour tissue because some cancers do not release enough detectable DNA.
In England, ctDNA testing is now used in defined NHS pathways
The situation in England has changed significantly in recent years.
Following national evaluation and pilot programmes, NHS England moved liquid biopsy into routine genomic testing for selected patients.
By 2026, ctDNA testing commissioned through the NHS Genomic Medicine Service includes defined indications such as:
genomic testing in patients with suspected non-small cell lung cancer;
ESR1 testing in certain patients with breast cancer where the result can influence targeted treatment;
testing of alterations involving PIK3CA, AKT and PTEN in particular breast cancer treatment pathways.
This is an important development because it shows that liquid biopsy is no longer purely experimental.
But it must be interpreted correctly.
The NHS is not offering a general ctDNA blood test to everyone to determine whether they have cancer. These analyses are used in specific clinical situations where the result can contribute to diagnosis or treatment selection.
In suspected lung cancer, England has also introduced a “blood-test-first” molecular pathway for eligible patients, allowing actionable genomic information to become available earlier in some cases.
This does not make tissue pathology obsolete. A tissue biopsy can reveal the structure, cell type and histological characteristics of a tumour — information that circulating DNA alone cannot provide.
For readers wanting to place medical genomic testing within the much broader field of genetic analysis, our overview of the different types of DNA tests explains how relationship, ancestry and genetic profile tests differ from clinical cancer testing.
Can ctDNA show whether cancer treatment is working?
One particularly useful characteristic of ctDNA is that its concentration can change over relatively short periods.
When treatment successfully reduces the number of cancer cells, the amount of tumour-derived DNA in the bloodstream may fall.
Conversely, rising ctDNA or the appearance of new molecular alterations can sometimes indicate tumour progression or the development of treatment resistance.
This creates the possibility of using blood samples to follow how a cancer evolves over time.
Researchers are investigating ctDNA dynamics in many tumour types to determine whether molecular changes can reveal treatment response before conventional imaging shows a clear difference.
However, this does not mean CT scans, MRI, PET scans or other clinical investigations can simply be replaced by blood tests.
The most useful approach is generally to combine information from:
symptoms and clinical examination;
imaging;
pathology;
genomic analysis;
treatment history.
Liquid biopsy adds another layer of information rather than replacing all the others.
This broader movement towards more personalised medicine is also discussed in our article on recent advances in DNA, genetic testing and cancer screening.
Molecular residual disease: can ctDNA detect cancer that imaging cannot see?
After surgery or other treatment intended to cure cancer, a very small number of malignant cells may remain in the body.
There may be too few cells to form a lesion visible on a scan.
If these remaining cells release tumour DNA into the bloodstream, an extremely sensitive ctDNA test may nevertheless detect a molecular signal.
This is known as molecular residual disease, or MRD.
Why MRD attracts so much attention
Across several cancer types, detectable ctDNA following curative-intent treatment has been associated with a substantially higher risk of recurrence.
This makes ctDNA potentially valuable for identifying patients whose cancer is more likely to return.
However, there is an important distinction between predicting risk and improving treatment.
A test may be able to show that one group of patients has a higher probability of recurrence without proving what doctors should do differently as a result.
For example, researchers still need to determine whether using ctDNA to:
intensify chemotherapy;
extend treatment;
change treatment;
reduce therapy in lower-risk patients;
actually improves survival or quality of life.
For many solid tumours, using ctDNA routinely to make these post-treatment decisions remains under clinical investigation.
Can a blood test screen for cancer in someone with no symptoms?
This is where the subject becomes more complicated.
Biologically, the idea is attractive.
If a tumour starts releasing abnormal DNA before a person develops symptoms, detecting that DNA might create an opportunity to diagnose the disease earlier.
But a screening test used in large numbers of healthy people faces much stricter requirements than a test used in someone already known to have cancer.
Screening requires very high specificity
Cancer is relatively uncommon in an apparently healthy screening population at any single point in time.
Even a small false-positive rate can therefore send large numbers of people without cancer for additional investigations.
Those investigations may involve:
repeat blood tests;
imaging;
endoscopy;
biopsies;
specialist appointments;
anxiety while waiting for results.
A population screening test therefore needs extremely high specificity.
It also needs sufficient sensitivity
At the same time, the test needs adequate sensitivity.
A highly specific test that rarely gives false alarms is not necessarily a good screening test if it misses many genuine cancers.
This is particularly difficult with ctDNA because early tumours may release almost no detectable DNA.
The challenge is therefore to achieve both:
very few false positives;
very few missed cancers.
That balance is difficult.
Galleri: what does the best-known multi-cancer blood test do?
Galleri, developed by GRAIL, is one of the most widely discussed MCED tests.
Rather than looking only for individual mutations, Galleri analyses methylation patterns in cell-free DNA.
Sequencing data and computational models are then used to identify whether a cancer-associated signal appears to be present and to predict its probable tissue of origin.
The test is designed to identify signals associated with more than 50 cancer types.
Importantly, it is a screening test rather than a standalone diagnostic test.
A positive result does not establish that a person has cancer. Further medical investigations are required.
Likewise, a negative result cannot prove that cancer is absent.
PATHFINDER 2: high specificity does not mean 99% of cancers are detected
PATHFINDER 2 provided important prospective data on Galleri in people aged 50 and over without a known cancer diagnosis.
Results presented in 2025 included 23,161 participants with 12 months of follow-up.
Among this population:
specificity was 99.6%;
positive predictive value was 61.6%;
sensitivity across all cancer episodes diagnosed during follow-up was 40.4%;
sensitivity reached 73.7% for a predefined group of 12 cancers responsible for a large proportion of cancer deaths in the United States;
the first prediction of cancer signal origin was correct in approximately 91.7% of true-positive cases.
These figures illustrate why statements such as “99% accurate” can be misleading when discussing cancer screening.
Specificity and sensitivity measure different things
A specificity of 99.6% means the test correctly returns a negative result for the vast majority of people without cancer.
It does not mean that the test detects 99.6% of cancers.
In the PATHFINDER 2 analysis, overall sensitivity was around 40%.
In other words, a substantial proportion of cancers diagnosed during follow-up had not generated a positive MCED result.
The study also largely reflected a US population, and the Galleri programme is commercially developed by GRAIL. These factors need to be considered when interpreting the findings.
NHS-Galleri: the major England trial did not meet its primary endpoint
The most important test of Galleri in an NHS population has been the NHS-Galleri trial.
More than 142,000 people aged 50 to 77 across England took part.
Participants had no symptoms of cancer when they entered the trial. The study examined whether adding annual Galleri testing to usual NHS screening and care could shift cancer diagnosis towards earlier stages.
The central question was not simply whether Galleri could detect cancer.
It was whether using the test would reduce the number of cancers diagnosed at stages III and IV in a predefined group of 12 cancers.
The primary endpoint was not met
Full results presented in 2026 showed that the trial did not achieve its primary endpoint.
There was no statistically significant reduction in the combined number of stage III and stage IV cancers in the main predefined analysis.
This result matters.
A screening programme should not be judged simply by how many cancer signals it detects. Researchers need evidence that screening changes clinically important outcomes.
There were encouraging secondary findings
The trial was not completely negative.
A pre-specified secondary analysis found fewer stage IV cancers among participants undergoing annual MCED testing, with reductions becoming more apparent during later rounds of screening.
Researchers also reported fewer cancers diagnosed through emergency presentations among people offered Galleri testing.
These findings are scientifically interesting and justify continued investigation.
However, secondary outcomes cannot be used to pretend that the original primary objective was achieved.
The question of mortality is even more important.
Longer follow-up is needed to establish whether adding MCED testing ultimately reduces deaths from cancer. Detecting a cancer earlier only benefits a population screening programme if that earlier detection translates into better outcomes without creating disproportionate harms from overdiagnosis, false positives and unnecessary procedures.
Galleri is not routine NHS population screening
As of 2026, the evidence from NHS-Galleri does not support Galleri as a routine NHS screening test for the general population in England.
People eligible for established NHS cancer screening should therefore continue to participate in the programmes recommended for them.
These include established screening for breast, bowel and cervical cancer, together with the ongoing national rollout of lung cancer screening for eligible people at higher risk.
An MCED blood test should not be treated as a substitute for those programmes.
Why can a ctDNA test be negative when cancer is present?
For ctDNA to be detected, the tumour must first release enough measurable material into the bloodstream.
That does not always happen.
A small or early cancer may:
release very little DNA;
release tumour DNA intermittently;
have biological characteristics that make its DNA more difficult to detect;
be located in a tissue from which relatively little ctDNA reaches the bloodstream;
contain alterations that are not covered by the assay.
The amount present may therefore fall below the laboratory's detection threshold.
A negative result means:
the test did not detect the cancer-associated signal it was designed to find in that particular sample at that particular time.
It does not mean:
the person definitely does not have cancer.
This distinction is essential both for clinical ctDNA testing and for multi-cancer screening.
Why can liquid biopsy produce false-positive results?
The opposite problem can also occur.
Not every genetic alteration detected in blood comes from a solid tumour.
One important source of interference is clonal haematopoiesis.
What is clonal haematopoiesis?
As people age, some blood-forming stem cells in the bone marrow acquire mutations.
A cell carrying one of these mutations may then multiply and produce a larger population of blood cells carrying the same genetic change.
This phenomenon is often referred to as clonal haematopoiesis of indeterminate potential, or CHIP, when specific criteria are met.
DNA released from these blood cells can enter the circulating cfDNA pool.
If an assay simply detects a mutation without determining where it came from, it could mistakenly attribute a blood-cell mutation to a solid tumour.
Modern liquid biopsy workflows therefore need to control for biological as well as technical sources of error.
These include:
clonal haematopoiesis;
sequencing errors;
sample degradation;
delayed blood processing;
contamination;
DNA extraction variability;
bioinformatics artefacts.
High-quality ctDNA testing therefore depends on the entire laboratory process, not only on the sequencing machine itself.
Liquid biopsy vs tissue biopsy: do we really have to choose?
Liquid biopsy and tissue biopsy provide different types of information.
A tissue biopsy allows a pathologist to examine the actual cells and architecture of a lesion under a microscope.
That information may establish:
whether a lesion is malignant;
the type of cancer;
its histological subtype;
cellular characteristics relevant to prognosis or treatment;
additional molecular and protein markers.
For many cancers, tissue remains essential for diagnosis.
A liquid biopsy, by contrast, provides relatively easy access to selected molecular information circulating in the blood.
Advantages of liquid biopsy
Potential advantages include:
blood sampling is less invasive than many tissue biopsies;
samples can be collected repeatedly;
it may provide genomic information when a tumour is difficult to biopsy;
it can help monitor molecular changes over time;
ctDNA may reflect DNA released by several tumour sites rather than one small tissue sample;
results can sometimes become available rapidly enough to accelerate treatment selection.
Limitations of liquid biopsy
The disadvantages are equally important:
some cancers release little or no detectable ctDNA;
a negative result can be falsely reassuring if interpreted incorrectly;
blood-derived mutations may be mistaken for tumour mutations;
the test cannot provide all the histological information obtained from tissue;
test performance varies between assays, tumour types and disease stages.
For most clinical applications, the more accurate view is therefore that liquid and tissue biopsies are complementary rather than interchangeable.
Circulating tumour DNA: what is established and what remains experimental?
ctDNA technology has progressed far enough to have a genuine role in modern oncology.
Some applications are now part of routine clinical care.
Others remain promising but unproven.
Uses with an established or growing clinical role
Depending on the cancer and clinical setting, ctDNA can already be useful for:
detecting actionable genomic alterations in known or suspected cancers;
providing molecular information when tissue is limited or difficult to obtain;
supporting targeted-treatment selection;
identifying some mechanisms of acquired drug resistance.
England's NHS rollout of ctDNA testing in selected lung and breast cancer pathways is a clear example of this transition from research into clinical use.
Uses still undergoing evaluation
Research is continuing into:
monitoring treatment response;
molecular residual disease;
predicting recurrence;
changing treatment on the basis of post-treatment ctDNA;
screening apparently healthy people for multiple cancers.
The evidence is not equally mature across these applications.
A biomarker may be strongly associated with prognosis without yet being proven to improve outcomes when doctors act on it.
Likewise, a blood test may detect several cancer types without having demonstrated that population-wide use reduces cancer mortality.
Why a multi-cancer blood test cannot yet replace existing screening
The technical achievement behind MCED testing is substantial.
It is now possible to examine a single blood sample and identify molecular patterns associated with many different cancers.
But population screening requires more than technical detection.
A successful programme must demonstrate that it:
finds clinically important cancers sufficiently early;
misses acceptably few dangerous cancers;
produces few enough false positives;
directs patients reliably to the correct follow-up investigations;
avoids excessive overdiagnosis;
does not expose large numbers of healthy people to unnecessary invasive procedures;
is feasible within the health service;
provides benefits that outweigh its harms and costs;
ultimately improves meaningful health outcomes.
This is why the results of NHS-Galleri are so important.
The trial showed that multi-cancer testing can identify cancer signals at scale in an NHS setting. It also showed why detecting those signals is not, by itself, enough to prove that the test should become a national screening programme.
Conclusion: a blood test can reveal tumour DNA, but it cannot yet screen reliably for every cancer
Circulating tumour DNA is one of the most important developments in molecular oncology.
Cancer cells can release tiny DNA fragments into the bloodstream. With technologies such as NGS, digital PCR and methylation analysis, laboratories can extract useful molecular information from those fragments without directly sampling the tumour.
Liquid biopsy already has practical clinical applications. In England, ctDNA testing is now commissioned within selected NHS cancer pathways, particularly where molecular information can help guide treatment.
Research into treatment monitoring and molecular residual disease is also advancing rapidly.
Multi-cancer screening is a different matter.
Tests such as Galleri demonstrate that a single blood sample can identify molecular signals associated with many cancers with very high specificity. But sensitivity remains substantially lower, particularly when cancers release little DNA.
The NHS-Galleri trial reinforced this distinction. Although some secondary results were encouraging, the study did not meet its primary objective of significantly reducing combined stage III and IV cancer diagnoses.
For patients and the general public, the key message is straightforward:
Detecting tumour DNA in the blood is already possible. Turning that technology into a reliable universal blood test that can rule out or detect every cancer is not.
Can a blood test detect cancer?
A blood test can detect certain molecular signals associated with cancer, including circulating tumour DNA. However, performance varies considerably according to the cancer type, stage and technology used. No ctDNA blood test can currently diagnose or exclude every type of cancer on its own.
What is the difference between cfDNA and ctDNA?
Cell-free DNA (cfDNA) refers to all DNA fragments circulating freely in the bloodstream. Most come from normal cells. Circulating tumour DNA (ctDNA) is the smaller fraction of cfDNA that originates from cancer cells.
What is a liquid biopsy?
A liquid biopsy analyses tumour-derived biological material found in blood or another body fluid. ctDNA is one of the most widely studied liquid-biopsy biomarkers. Unlike a conventional tissue biopsy, it does not require removing a piece of the tumour itself.
Does a negative ctDNA result mean I do not have cancer?
No. Some cancers release very little tumour DNA, particularly when they are small or localised. A negative result means that the assay did not detect its target signal in the sample analysed. It cannot completely rule out cancer.
Is Galleri available as routine NHS cancer screening in England?
No. Galleri was evaluated in more than 142,000 people in the NHS-Galleri trial, but the study did not meet its primary endpoint. As of 2026, it is not part of routine NHS population cancer screening in England. Eligible people should continue to use established NHS screening programmes.