← All Conditions

A Cancer Blood Test Came Back Positive
But the Scan Is Normal

A result exists and a destination does not, and that is the hardest situation these tests create. In people with no symptoms, most positive results turn out to be false — and proving that takes longer than finding a cancer would. There is currently no agreed answer to how closely someone in this position should be followed, and anyone who offers one with confidence is going beyond the evidence. What can be done is to close the chest half of the question properly, with a defined interval and a defined end point. Dr Lawrence Okiror, Consultant Thoracic and Robotic Surgeon (GMC 6150382). Appointments at London Bridge Hospital and The Lister Hospital Chelsea within 2–3 days, including video consultations. Self-referrals welcome.

Last reviewed: August 2026 · Dr Lawrence Okiror FRCS(CTh) FRCSEd(CTh) · GMC 6150382

Most Positives Are False

In a well population, serious disease is uncommon — so even a very accurate test raises more false alarms than true findings. That is arithmetic, not a fault in the test.

Proving Nothing Takes Longest

In the largest published experience, resolving a positive took around two months where a cancer was found and around five where none was, with a quarter waiting eight months or longer.

A Defined End Point

The chest half can be closed properly: what is being watched, with what test, at what interval, and at what point the matter is considered finished. Not an open-ended arrangement to keep scanning.

A result exists. A destination does not.

Almost every other test in medicine hands you on to someone. An abnormal liver test goes to a hepatologist. A suspicious mammogram goes to a breast clinic. A positive multi-cancer blood test in a well person with normal imaging goes nowhere in particular, because it does not name an organ, a specialty or a pathway. You are left holding a result that has no obvious owner.

That is not a failure of the people you have seen. The situation is genuinely new, it falls between specialties by design, and the published literature is explicit that the right follow-up is not yet known. Being told there is no clear answer is unsatisfying, but it is more useful than being given a confident one that has nothing behind it.

What follows is what these tests measure, what a positive result can and cannot tell you, what the evidence actually shows about resolving one, and what can be settled about the chest. For the imaging side of the same problem — a scan that has found something rather than a blood test — see a scan or blood test has flagged something in your chest.

What is the blood test
actually measuring?

Tests sold under similar names work in quite different ways, and what a positive means differs between them. Two broad kinds account for almost everything on sale.

Tests that read fragments of DNA

What they look at
Every cell in the body sheds small fragments of DNA into the bloodstream, all the time and quite normally. These tests examine the chemical marks carried on those fragments, because the pattern of marks differs between cancer cells and healthy ones. Some also attempt to predict which organ a signal came from.
What a positive says
That a pattern consistent with cancer has been detected somewhere. Not that a cancer has been located, sized, or confirmed. The word detected in these reports carries much less than it appears to.

Tests that look for whole cells

What they look at
Whole tumour cells circulating in the bloodstream, isolated from a blood sample and examined. A different technique from the DNA-based tests, and one that generally targets solid-organ cancers rather than blood or lymphatic ones.
What a positive says
Again, that something consistent with cancer has been identified in a blood sample. Where a report also suggests a likely site of origin, that suggestion is a probability rather than a finding, and it is worth knowing that these tests are usually not designed to detect blood or lymphatic cancers at all.

One point of practical confusion is worth clearing up. Galleri is the test most people have heard of, because of the large NHS trial that reported in 2026. Until recently it was available in the UK only within that trial; it carries a UKCA mark and began to be offered privately here during 2026 through a small number of clinical providers, though it remains unavailable on the NHS. UK clinics also sell several other tests, and because the reporting language is similar, people frequently believe they have had Galleri when they have not. Bring the actual report rather than the brand name — the difference matters to how the result should be read.

Why are most positive results
false in people who feel well?

This is the part that most changes how a result should feel, and it has nothing to do with the quality of the test. It is about who the test was given to.

Imagine a very accurate test applied to a large group of well people. Serious cancer is uncommon in such a group. The small number of people who do have a cancer will mostly be picked up. But the very large number who do not have one will produce a small proportion of incorrect alarms — and because that group is so much bigger, that small proportion can easily outnumber the true findings. The result is that in people without symptoms, a positive is more often wrong than right.

Why this is not an argument against the tests

The same arithmetic applies to every screening test ever devised, including the national programmes, which is why those programmes are offered to defined groups at defined ages rather than to everyone. It is the reason screening is targeted rather than universal, and it is a feature of screening rather than a flaw in any particular product.

What it does mean is that a positive result in a well person is a reason to investigate carefully, not a reason to assume the worst. And it means the sensible response to a clear scan is considerable reassurance rather than a search for what the scan must have missed.

There are other explanations for a positive result besides cancer. Some benign conditions of the blood and bone marrow produce the same signals; in one reported series a substantial share of false positives were traced to a benign condition of the white blood cells. Where a positive result is unexplained after imaging, a haematology opinion is sometimes the more useful next step than another scan.

How long does this
usually take to settle?

Longer when nothing is found than when something is, which is the opposite of what most people expect and the single most useful thing to know in advance.

In the largest published experience of a multi-cancer blood test used in this way, the time taken to reach a resolution was around two months where a cancer was eventually diagnosed, and around five months where none was. A quarter of people waited eight months or longer. A later and considerably larger study reported shorter times overall, so the picture is improving — but the shape of it does not change. Finding a cancer ends the question. Proving there is no cancer has no equivalent moment.

That asymmetry is the real cost of these tests, and it is not measured in money. It is months of a well person's life spent inside an unresolved question. Knowing that in advance changes how the waiting feels, and it is a strong argument for a plan with a stated end point rather than an open arrangement to review things periodically.

The reassuring counterweight from the same body of work: very few people in these studies ended up having an invasive procedure as a consequence, and of those who did, the great majority were not operations.

Is there an agreed protocol
for what happens next?

No. And that deserves to be said in one word rather than dressed up.

Published analysis of exactly this situation states that individuals with risk factors whose first round of investigation comes back negative or equivocal remain at some risk of a cancer diagnosis later, and that uncertainty remains as to how closely such people should be followed. It is unusual for a scientific paper to concede that so directly, and it should be taken at face value. Anyone offering you a confident surveillance schedule for this situation is going beyond what is known.

A useful comparison from a different setting

Blood tests for cancer DNA are also used in people who already have a diagnosed cancer, to look for traces remaining after treatment. That is a much easier problem — the cancer is known and the target is defined. Even there, the European Society for Medical Oncology states that current assays give information about outlook but not about which treatment will work, that sensitivity is suboptimal with a large proportion of false negatives, and recommends against using them to guide treatment decisions outside clinical trials.

That is not a statement about screening tests in well people, and it should not be presented as one. But if the technology is not yet considered ready to guide decisions in the situation where the answer is known, it is worth holding that in mind when a result arrives in someone entirely well.

What can be done, in the absence of a protocol, is to make the plan explicit. What is being watched. With what test. At what interval. And at what point the matter is considered closed. A plan with those four elements is worth more than a confident-sounding schedule with nothing behind it.

Should I just keep scanning
until something turns up?

It is the natural response and it is usually the wrong one. Repeat whole-body imaging of a well person mostly produces further incidental findings rather than an explanation for the original result. Each of those findings then needs its own assessment. Some lead to procedures. And the original question stays exactly where it was.

The more useful approach runs in the other direction: go narrow rather than wide. Identify anything in the imaging you already have that is genuinely equivocal. Characterise that one thing properly with a targeted test aimed at the specific question. Then set a defined interval with a defined end point for anything that remains uncertain, and stop.

One further point about the chest specifically. If lung cancer is the concern behind all this, the scan that answers it is a CT. A whole-body MRI is progressively unreliable for small lung nodules and a chest reported as clear on one has not had its lungs examined to CT standard. That is set out in detail on the chest finding page, and what a lung nodule is and how it is assessed on the lung nodule page.

What can a thoracic surgeon
actually do here?

Take responsibility for the chest half, and close it properly. That is a narrower offer than you might want, and it is the one I can make honestly.

What that means in practice

Reviewing the imaging itself rather than the report, because reports are written to be defensible and often sound more equivocal than the pictures are. Deciding whether anything in the chest genuinely needs characterising, and arranging the one targeted test if it does. And setting a defined interval and a defined end point, so that this stops being an open question.

Where the honest answer is that nothing in the chest requires anything at all, that is what you will be told — plainly, and with the reasoning, so that it holds up when the worry comes back at three in the morning.

What I will not do is take on the findings outside the chest. These investigations routinely produce results in the liver, kidneys, adrenal glands, prostate, spine and brain, and working outside my specialty would not serve you. I will tell you which of them look as though they warrant a specialist opinion and which are the common harmless ones, rather than leaving you to work that out from a report.

And I will not offer a view on whether you should have had the test. People choose these deliberately, usually after reading a great deal, and being told afterwards that it was a mistake helps nobody. My involvement starts at the point the result exists. Where a case is genuinely uncertain it is discussed at the chest multidisciplinary team meeting, and if you have already been given an opinion you are unsure about, a second opinion is a reasonable thing to want.

Where the figures above
come from

The statements on this page are written without technical vocabulary. The underlying figures are set out here in full.

  • Time to diagnostic resolution. In the PATHFINDER study, median time to diagnostic resolution following a cancer signal detected result was 79 days overall — 57 days (IQR 33–143) in true positives and 162 days (IQR 44–248) in false positives. Most participants had both laboratory tests and imaging. Schrag D et al, The Lancet, 2023.
  • Positive predictive value. In the same study, positive predictive value was 43.1% with the refined test and 38.0% with the earlier version, with specificity of 99.5% and 99.1% respectively. In a single-institution subset of 1,735 participants, 27 returned a signal-positive result (1.6%) with 12 confirmed cancers, giving a positive predictive value of 44%; six of the 15 false positives correlated with monoclonal B-cell lymphocytosis, a benign precursor condition.
  • The larger, more recent cohort. PATHFINDER 2 (35,878 participants, presented at ASCO 2026) reported cancer signal origin accuracy of 91.3%, median diagnostic resolution of 48 days, invasive procedures in 0.6% of safety-analysable participants (213 of 35,335), of which 90.5% were non-surgical, and 71% of newly detected cancers at stages I–III.
  • Uncertainty about follow-up. Published analysis of cancer signal origin prediction and diagnostic resolution in PATHFINDER states that participants with known risk factors and negative or equivocal initial evaluations remain at risk of receiving a cancer diagnosis, and that uncertainty remains around how closely such patients should be followed to rule out malignancy at a future date. Cancer Prevention Research 2025;18(8):475.
  • The NHS-Galleri trial. Reported at ASCO in May 2026, the trial did not meet its primary endpoint of reducing the combined number of cancers diagnosed at stage III or IV. It reduced stage IV diagnoses, increased overall cancer detection when added to standard screening, raised stage I–II diagnoses, and reduced diagnoses made through emergency presentation. Confidence intervals were wide and mortality follow-up is not yet available.
  • Availability of Galleri in the UK. The Galleri test carries a UKCA mark. Until 2026 it was available in the UK only within the NHS-Galleri trial; private provision began during 2026 through a small number of UK clinical providers and private GP clinics. It remains unavailable on the NHS, and the NHS-Galleri trial website continues to state that the test cannot be bought in the UK, reflecting the position before commercial launch.
  • Blood tests in known cancer. The ESMO Clinical Practice Guideline for early and locally advanced non-small-cell lung cancer states that circulating tumour DNA detection or clearance after curative-intent treatment is not recommended to guide treatment escalation or de-escalation outside of clinical trials, noting that current assays are prognostic but not predictive, with suboptimal sensitivity and a large proportion of false negatives. Zer A et al, Annals of Oncology 2025;36:1245–1262. This guidance concerns residual disease testing in patients with a known cancer, not multi-cancer screening in well people.
  • MRI and lung nodules. Pooled per-lesion sensitivity of MRI for nodules ≥4 mm was 87.7%, rising to 98.5% at 8–10 mm and falling to 80.5% below 8 mm; in a direct comparison of 113 nodules, sensitivity was 57.1% at ≤4 mm. CT was statistically superior to MRI on sensitivity, specificity and AUC in meta-analysis. Radiology: Cardiothoracic Imaging 2024; Cieszanowski A et al, PLOS One 2016;11:e0156272.
  • Screening guidance. The UK National Screening Committee does not recommend whole-body screening in healthy adults. No guideline body currently recommends a blood test for cancer screening in the general population.

Questions About
A Positive Cancer Blood Test

Questions people ask after a multi-cancer or circulating tumour cell blood test has returned a positive result and the imaging has found nothing.

Book a Consultation →

Or call Jo Mitchelson:
020 7952 2882

My cancer blood test was positive but every scan has been normal. What does that mean?
Most often it means the test raised a signal that was not caused by a cancer. In people with no symptoms, most positive results on these tests turn out to be false — not because the tests are badly made, but because serious disease is uncommon in a well population, and even a very accurate test produces more false alarms than true findings when it is applied to people who are unlikely to be ill. The other possibilities are that a cancer is present but too small or in a place the imaging could not resolve, or that the result reflects something else in the blood entirely. What can honestly be said is that a clear scan makes a serious, treatable cancer considerably less likely, and does not reduce that to zero.
What are these blood tests actually measuring?
There are two broad kinds and they work differently. The first looks at cell-free DNA — fragments of DNA that all cells shed into the bloodstream — and examines the chemical marks on it, since the pattern of those marks differs between cancer cells and healthy cells. Some of these tests also try to predict which organ a signal came from. The second kind looks for whole tumour cells circulating in the blood. Neither locates a cancer; both report that a pattern consistent with cancer has been detected. That distinction is the source of most of the confusion, because a report saying a cancer signal has been found reads as though something has been discovered when what has been found is a statistical pattern.
How long does it usually take to resolve a positive result?
Longer when nothing is found than when something is, which is the opposite of what most people expect. In the largest published experience of a multi-cancer test used this way, the median time to reach a resolution was around two months where a cancer was eventually diagnosed and around five months where none was, with a quarter of people waiting eight months or longer. A later and much larger study reported considerably shorter times overall, so this is improving. But the pattern holds: proving that nothing is there takes longer than finding something, because there is no single test that ends the question.
Should I just keep having scans until something is found?
No, and this is where real harm happens. Repeat whole-body imaging of a well person mostly generates further incidental findings rather than an explanation — each of which then needs its own assessment, and some of which lead to procedures. The useful approach is the opposite: identify anything in the existing imaging that is genuinely equivocal, characterise that one thing properly with a targeted test, and otherwise set a defined interval with a defined end point. An arrangement to keep scanning indefinitely is not surveillance; it is an unresolved question being deferred.
Is there an agreed protocol for what happens next?
No, and anyone who offers you one with confidence is going beyond the evidence. Published analysis in this field states plainly that individuals with risk factors whose first round of investigation is negative or equivocal remain at some risk of a later cancer diagnosis, and that uncertainty remains as to how closely such people should be followed. That is an unusual thing for a scientific paper to say and it is worth taking at face value. What can be done is to make the plan explicit rather than open-ended: what is being watched, with what test, at what interval, and at what point the matter is considered closed.
Is the Galleri test available in the UK?
That has changed recently, and you may encounter contradictory information. For several years Galleri was available in the UK only within the NHS-Galleri trial, whose main results were reported in 2026 — it did not meet its primary endpoint, although it did reduce the number of cancers found at the most advanced stage and increased the number found early. The test carries a UKCA mark, and during 2026 it began to be offered privately in the UK through a small number of clinical providers and private GP clinics. It is still not available on the NHS, and the trial's own website continues to state that the test cannot be bought here, which reflects the position before that commercial launch rather than an error. UK clinics also sell several other cancer blood tests, and what a positive result means differs considerably between them — so bring the actual report rather than the brand name.
If these tests are unproven, why do doctors use blood tests for cancer at all?
Because there is a real difference between using a blood test in someone who is known to have cancer and using it to screen a well population. Even in the first situation, where the cancer is already diagnosed and the target is known, major guidelines are cautious. The European Society for Medical Oncology states that current assays for detecting residual cancer DNA after treatment are prognostic but not predictive, that sensitivity is suboptimal with a large proportion of false negatives, and recommends against using them to guide treatment decisions outside clinical trials. If the technology is not yet ready to guide decisions where the cancer is already known, that is worth holding in mind when interpreting a result in someone who is entirely well.
What can a thoracic surgeon actually do for me here?
Take responsibility for the chest half of the question, which is the part I can answer properly. That means reviewing the imaging itself rather than reading the report, deciding whether anything in the chest genuinely needs characterising and arranging the one targeted test if it does, and giving you a defined interval and a defined end point rather than an open-ended arrangement. Where the honest answer is that nothing in the chest requires anything at all, I will say that plainly. What I will not do is take on findings outside the chest, which these investigations produce in quantity — though I will tell you which of them look as though they warrant a specialist opinion.
Will you tell me I should not have had the test?
No. People choose these tests deliberately and usually after reading a good deal about them, and being lectured about it afterwards helps nobody and is not what you came for. My view is that these tests detect real signals, that the gap between detecting a signal and explaining it is where the difficulty lies, and that the job in front of us is closing that gap for the part of the body I can speak to. My involvement starts at the point the result exists.
Do I need a GP referral?
Self-referrals welcome — no GP letter is needed before booking. Appointments at London Bridge Hospital and The Lister Hospital Chelsea within 2–3 days, including video consultations, with outpatient clinics also at Canary Wharf and the City of London. Send the blood test report itself, the radiology report, and the images if you have them on disc or as a download link. Most major insurers accepted, and self-pay estimates are provided before any commitment is made.

The result has no owner.
The chest half can have one.

Self-referrals welcome. Appointments at London Bridge Hospital and The Lister Hospital Chelsea within 2–3 days, including video consultations, with outpatient clinics at Canary Wharf and the City of London. Send the blood test report, the imaging report and the images themselves.

Book a Consultation → Second Opinion

Jo Mitchelson, PA  · 020 7952 2882 · pa@lungsurgeon.co.uk

St Thomas’ Hospital #1 UK · Guy’s Hospital #2 UK · London Bridge Hospital #10 UK · Newsweek World’s Best Hospitals 2026

Continue Reading

A Scan Has Flagged Something in Your ChestIf the imaging did find something — what these scans see well, what they miss, and which second test settles it. A Finding in the Front of the ChestPossible thymoma on a health-check scan — and why most of these are normal thymus or a harmless cyst. What Is a Lung Nodule?Size, appearance and behaviour over time — how a nodule is assessed and when it needs action. Shadow Found on a ScanA shadow on a chest X-ray or CT — what it can be, and what happens next. Chest Imaging ExplainedWhat CT, MRI and PET each contribute at each stage, and what the phrases in a radiology report mean. The Lung Screening GapWho qualifies for lung cancer screening, who does not, and what the criteria miss. Lung Nodule Precision PathwayHow a nodule that does need answering is sampled and diagnosed without an open operation. Specialist Second OpinionIndependent review of your imaging and your diagnosis, within 2–3 days. Book an AppointmentLondon Bridge Hospital, The Lister Chelsea, Canary Wharf and the City of London.
📅Book 📞020 7952 2882