Beyond the All-Clear: What a Negative Multi-Cancer Blood Test Really Rules Out

A blood test that scans for signals across dozens of tumour types is an instinctively attractive idea, and it is no longer hypothetical. Multi-cancer early detection (MCED) assays are available through some private providers in the UK, continue to be examined within NHS-linked research, and are discussed openly by patients who have read about them in the press. The technology is not the problem. The problem is the word “clear”.

A negative result lands differently from a positive one

A positive MCED result activates a pathway: referral, imaging, investigation, a named clinician taking ownership. A negative result activates a feeling. Patients who have paid several hundred pounds, or waited weeks for a kit to be processed, tend to treat the absence of a signal as the conclusion of an investigation rather than a single observation within one. Primary care then inherits a consultation in which reassurance has already been issued, often in writing, by somebody who will never see the patient again.

That asymmetry is worth naming explicitly. Diagnostics do not simply produce information; they produce confidence, and confidence is distributed unevenly across results.

Sensitivity: what the assay can actually see

Sensitivity describes the proportion of people who genuinely have cancer and who test positive. For MCED assays, published evaluations have generally reported overall sensitivity in the region of 50%, with substantial variation by tumour type and, more importantly, by stage. Detection in advanced disease is markedly better than in stage I disease, where circulating tumour DNA may be present at vanishingly low concentrations.

Specificity — the proportion of people without cancer who correctly test negative — is typically reported above 99%, a figure that sounds decisive until it meets a large denominator.

Two practical consequences follow. A negative result cannot exclude a cancer that was present on the day of sampling, particularly an early-stage one. And even a very specific test will generate a meaningful absolute number of false positives when applied to tens of thousands of people, which is precisely why confirmatory pathways matter more than the headline accuracy figure.

Positive predictive value, or why rarity changes everything

Positive predictive value (PPV) is the probability that a positive result is a true positive, and it depends as much on how common the condition is as on how good the test is.

Consider an illustrative cohort of 100,000 people aged 50 to 77 in whom 0.5% have an undiagnosed cancer — 500 individuals. At 50% sensitivity, roughly 250 cancers are flagged. At 99.5% specificity, roughly 497 people without cancer also receive a positive result. The PPV sits near one in three: two of every three positive results would be false alarms. Meanwhile 250 cancers pass through undetected.

These numbers are illustrative and shift with prevalence, age and tumour mix, but the arithmetic is not optional. It is the reason MCED results must be interpreted alongside symptoms, history, examination and conventional investigation rather than in place of them.

Interval cancers and the cancers these tests do not seek

A negative test is a snapshot taken on a particular morning. Cancers too small to shed detectable signal, or tumours that shed little circulating DNA at any size — early renal, prostatic, thyroid and some central nervous system tumours among them — may remain invisible. Fast-growing tumours can become clinically evident within months of a normal result. These are interval cancers, a predictable property of any test delivered at intervals, not necessarily a defect in the assay itself.

It is also worth restating the obvious, because it is frequently lost: an MCED result does not substitute for the NHS bowel, breast or cervical screening programmes, which have been validated in large populations over decades. Attendance at those programmes should continue regardless of what a private blood test said.

What the NHS evaluation is and is not telling us

The UK has invested substantially in this question through large-scale NHS-linked research, including the Galleri evaluation study, which recruited well over 100,000 participants. That work is intended to answer questions a commercial launch cannot: whether adding an MCED test to standard care shifts the stage at which cancers are diagnosed, whether it causes harm through over-investigation, and whether it is affordable at population scale. Interim signals are not the same as outcome data, and the UK National Screening Committee has so far judged the evidence insufficient to support a national programme.

GPs should be comfortable saying precisely that, without either dismissing the technology or endorsing it ahead of the evidence.

Questions clinicians should be ready to answer

• Which test was used, and is it CE/UKCA marked for this intended use?
• What sensitivity and specificity does the manufacturer report, for which cancer types, and at which stages?
• Was the result returned as “no signal detected” or as “no cancer detected”? The two statements are not equivalent.
• Does the patient have symptoms that were present before the test? Symptoms override screening logic entirely and warrant investigation on their own terms.
• When is the next safety net, and what symptoms would trigger an earlier review?

Questions patients should ask

Patients should be encouraged to ask what a negative result excludes, what it does not, and what happens next. A written result that contains no statement of limitation is a result that deserves a second conversation.

The bottom line

An MCED test can be a useful addition to a diagnostic pathway. It is not a certificate. The clinically honest position in 2026 is that a negative result lowers, but does not eliminate, the probability of undiagnosed cancer — and that a patient who develops symptoms after a clear result needs assessment, not reassurance by recall.