NEJM - NHS: Galleri - Early Detection Only Matters If Diagnosis Follows Fast
The next question for multi-cancer screening is how much of the time gained by early detection survives the diagnostic work-up
Early cancer detection has one central purpose: to gain time.
Finding a cancer before symptoms appear should create an opportunity to diagnose and treat it while disease remains localized and potentially curable.
But detecting a molecular cancer signal early is not the same as establishing an early cancer diagnosis.
Between those two events sits the diagnostic pathway: imaging, specialist referral, additional testing, biopsy and pathology.
If that pathway takes weeks or months, part of the time gained by screening can be lost before the patient reaches an actionable diagnosis.
For multi-cancer early detection tests such as Galleri, this distinction deserves greater attention.
The relevant question is not only:
How early can Galleri detect a cancer signal?
It is also:
How quickly can that signal become a definitive diagnosis?
NHS-Galleri leaves that question open
The NHS-Galleri randomized trial enrolled 142,250 participants between 50 and 77 years of age.
After three screening rounds, adding multi-cancer early detection testing to usual care did not significantly reduce stage III or IV cancer incidence across the 12 prespecified cancers. The incidence-rate ratio was 1.03 (95% CI, 0.92–1.14; P=0.63).
Stage IV cancer, a key secondary endpoint, showed a more favorable incidence-rate ratio of 0.86, but the authors concluded that further follow-up is needed.
The negative primary endpoint should not be minimized.
But it answers a different question from the one examined here.
NHS-Galleri asks whether population screening changes cancer stage at diagnosis.
It does not establish what the fastest or most resource-efficient pathway should be after Galleri produces a positive result.
What Galleri can do today
Galleri analyzes cancer-associated methylation patterns in cell-free DNA.
Its current architecture can provide:
1. Cancer Signal Detected / Not Detected
and, when positive:
2. Cancer Signal Origin — CSO
The CSO attempts to identify the anatomical origin most likely responsible for the signal.
FDA describes Galleri as a test intended both to detect a cancer-specific methylation signal and predict where that signal may have originated. A positive result must still be followed by diagnostic work-up.
That localization is clinically important.
A test saying:
Cancer signal detected somewhere in the body
creates a broad search problem.
A result saying:
Cancer signal detected — probable lung origin
creates a much narrower one.
Galleri can also provide a third layer of information in a subset of positive cases.
Its Supplemental Cancer Signal Origin, or CSO-S, can report additional biological context derived from methylation patterns. Current categories include squamous-cell signal, neuroendocrine signal, HPV-associated signal and several hematologic lineages.
CSO-S is returned in approximately 40% of eligible Cancer Signal Detected cases.
So Galleri already demonstrates three increasingly specific levels of information:
Cancer signal
↓
Probable organ
↓
Additional biological characterization in selected cases
But the third level remains limited. Galleri does not currently provide comprehensive histological classification for every positive result.
The opportunity: use the initial blood draw to reduce uncertainty before imaging
Consider a patient whose test reports:
Cancer Signal Detected
followed by:
Cancer Signal Origin: Lung
If CSO-S also identifies a squamous-cell signal, the clinician has substantially more information before performing the first diagnostic image than would be available from a nonspecific positive blood test.
The next logical development question is therefore:
How much more useful tumor information could be extracted from the initial blood draw before the patient enters the imaging pathway?
A future workflow could collect Galleri samples together with additional plasma aliquots during the same venipuncture.
If Galleri were negative, those additional samples would not need further analysis.
If Galleri were positive, they could potentially be used for a predefined reflex molecular assay designed to refine the suspected tumor type before imaging.
For example:
Cancer signal
↓
Lung origin
↓
Squamous / adenocarcinoma / neuroendocrine probability
↓
Targeted chest imaging
↓
Targeted biopsy
↓
Pathological confirmation
This is not a current validated Galleri workflow.
It is a development hypothesis.
Its purpose would not be to replace tissue diagnosis. It would be to reduce uncertainty before deciding which image, which specialist and which biopsy route are needed.
There is an important biological limit
This concept faces a major technical constraint.
Early cancers often release extremely small quantities of circulating tumor DNA.
In early-stage disease, the tumor-derived fraction of circulating cell-free DNA can fall below 0.1%.
That creates a paradox.
The cancers for which additional molecular characterization would be most useful — small, early, potentially curable tumors — may be precisely those producing the least ctDNA.
Obtaining increasingly detailed genomic information from such samples may therefore require greater analytical sensitivity, deeper sequencing, error suppression and sufficient plasma volume.
Each improvement has consequences for:
limit of detection;
sequencing depth;
turnaround time;
sample requirements;
laboratory cost.
Methylation-based classification may have advantages because it interrogates broader biological patterns rather than relying on a single mutation, but it does not remove the fundamental problem of low tumor-derived signal.
The relevant development question is therefore not:
Can blood replace biopsy?
It is:
How much clinically useful uncertainty can blood remove before imaging, given the very low tumor fraction expected in early-stage disease?
Why this matters: diagnostic resolution already takes time
PATHFINDER provides a useful first look at what happens after a positive MCED result.
Among participants with a Cancer Signal Detected result, median time to diagnostic resolution was:
79 days overall
57 days for true positives
and
162 days for false positives.
These figures should not be treated as current universal Galleri waiting times.
PATHFINDER was conducted during the height of the COVID-19 pandemic, and the investigators explicitly acknowledged that disruption of diagnostic services may have prolonged diagnostic resolution. A later analysis also noted a median of approximately 43 days in GRAIL's subsequent real-world commercial experience.
Nevertheless, PATHFINDER establishes an important point:
the interval between molecular detection and diagnostic resolution is clinically meaningful and measurable.
A cancer signal can be generated quickly.
A definitive answer may take much longer.
False positives create a different time problem
For true positives, delay potentially consumes part of the clinical advantage created by early detection.
For false positives, the problem is different.
The patient spends time moving through a cancer diagnostic pathway despite ultimately not being diagnosed with cancer.
In PATHFINDER, 57 participants had a positive MCED result without subsequent cancer confirmation.
Of those:
53 of 57 — 93% — underwent imaging
and
17 of 57 — 30% — underwent an invasive diagnostic procedure.
The median time before those patients reached diagnostic resolution was 162 days.
Radiology analyses of PATHFINDER also show substantial use of PET/CT and broad CT imaging in patients whose results ultimately proved false positive, with some patients undergoing multiple studies.
That creates three separate burdens:
resource utilization
Imaging, specialist consultations and invasive procedures are consumed without ultimately finding cancer.
financial exposure
The burden falls somewhere — on the payer, health system or patient.
iatrogenic uncertainty
For months, a person may know that a blood test has detected a cancer signal without knowing whether cancer is actually present.
The false-positive pathway should therefore not be treated merely as a specificity statistic.
It is a diagnostic pathway with its own clinical and economic consequences.
Every unnecessary image can also mean another queue
Advanced diagnostic imaging is not an unlimited resource.
MRI, CT and PET/CT require equipment, specialized personnel and appointment capacity.
In July 2026, more than one fifth of patients waiting for MRI within NHS England had already been waiting at least six weeks.
Canadian data provide an even more pronounced example, with national MRI waiting times around 59 days at the median and much longer waits for patients toward the upper end of the distribution.
These figures do not mean that every patient with suspected cancer would experience those delays. Urgent oncology pathways are commonly prioritized.
But they illustrate the capacity constraint.
The important operational point is simple:
An unnecessary imaging study is not only another cost. It can also be another queue.
If the patient enters several diagnostic queues sequentially, the delays accumulate.
The objective should be one high-yield route to the lesion
Cancer Signal Origin creates an opportunity to avoid broad anatomical searching.
Instead of:
positive signal
↓
CT
↓
MRI
↓
PET/CT
↓
specialist referral
↓
another image
↓
biopsy
a sufficiently informative molecular pathway could move toward:
positive signal
↓
probable organ
↓
additional biological information when technically feasible
↓
one high-yield organ-specific investigation
↓
targeted biopsy
↓
pathology
This distinction matters particularly because different suspected cancers require different first-line investigations.
A pulmonary signal may appropriately lead to diagnostic chest CT.
A pancreatic signal may call for pancreas-protocol CT or MRI/MRCP.
A colorectal signal may lead toward colonoscopy.
The objective is not to reduce imaging indiscriminately.
It is to avoid images that do not materially reduce diagnostic uncertainty.
Early PATHFINDER data suggest localization can help
A later analysis of PATHFINDER examined how Cancer Signal Origin influenced diagnostic evaluation.
Among the selected participants analyzed, 82% achieved diagnostic resolution after the initial evaluation. Of those, 78% reached resolution through a CSO-directed work-up.
Whole-body imaging contributed to diagnostic resolution in only 49% of Cancer Signal Detected cases.
The analysis was small and retrospective, and diagnostic pathways were not standardized.
It therefore does not prove that CSO-guided investigation reduces costs or waiting times.
But it supports a practical hypothesis:
If molecular localization is sufficiently accurate, imaging can be used to locate a suspected lesion rather than search broadly for an unknown cancer.
The endpoint we should measure: Detection-to-Diagnosis Interval
Sensitivity, specificity and positive predictive value remain fundamental.
But they do not tell us how quickly a screening result becomes clinically actionable.
BBIU proposes adding a simple operational endpoint:
Detection-to-Diagnosis Interval
Defined as the time between:
initial screening blood collection
and
definitive diagnostic resolution.
For a true positive:
blood draw → confirmed cancer diagnosis
For a false positive:
blood draw → confident exclusion or resolution of the cancer signal
Both matter.
For true positives, the objective is to preserve the time advantage created by screening.
For false positives, the objective is to reach safe resolution quickly while minimizing unnecessary procedures.
The next study should test the pathway, not only the test
A useful prospective study could compare current physician-directed work-up against a predefined diagnostic pathway using:
Cancer Signal Origin
available CSO-S information
future reflex molecular characterization when analytically feasible
a predefined high-yield organ-specific diagnostic procedure.
The most useful endpoints would be concrete:
days from positive result to definitive resolution
days from positive result to biopsy in confirmed cancers
number of imaging studies per confirmed cancer
PET/CT utilization
whole-body imaging utilization
number of invasive procedures among false positives
number of specialist visits
cost per diagnostic resolution
and
percentage of the time advantage created by screening that remains when definitive diagnosis is reached.
The last measure may be particularly important.
A screening program does not create clinical value merely by moving the date of molecular detection earlier.
It creates value if that earlier detection allows action earlier.
BBIU Decision Frame
NHS-Galleri leaves the central population-level question unresolved:
Will MCED screening ultimately reduce clinically meaningful late-stage disease and improve outcomes?
That evidence still needs to mature.
But a second question can already be addressed:
Can Galleri shorten the route between a molecular cancer signal and definitive diagnostic resolution?
Current Galleri can detect a cancer-associated signal.
It can predict a probable anatomical origin.
And in a subset of positive cases, it can already provide additional biological classification.
The next development opportunity is to determine how much more uncertainty can safely and reliably be removed from the initial blood draw before the patient enters imaging.
That concept has a biological limit: very low ctDNA fractions in early-stage disease may restrict how much molecular detail can be extracted.
It also has a clinical limit: pathology will remain necessary in many cancers.
But neither limitation eliminates the opportunity.
The goal is not to replace imaging or biopsy.
It is to reach the right image and the right biopsy with fewer unnecessary steps.
Because early cancer detection does not create an unlimited time advantage.
Every additional appointment, inconclusive investigation and unnecessary diagnostic queue consumes part of it.
And for a false positive, those same steps consume healthcare resources while extending the period during which a healthy person remains under suspicion of cancer.
The relevant question for MCED therefore extends beyond:
How early was cancer detected?
It should also include:
How much of that time advantage remained when the diagnosis became definitive and actionable?
That is where early detection becomes early diagnosis.
And only then can it become earlier treatment.
Challenge Your Assumptions. Before You Decide.
References
Sasieni P, Johnson P, Round T, et al. Effect of Screening with Multicancer Early-Detection Test on Late-Stage Cancer Diagnosis. N Engl J Med. 2026. doi:10.1056/NEJMoa2505723.
U.S. Food and Drug Administration. FDA Executive Summary: GRAIL Inc., Galleri. Molecular and Clinical Genetics Panel of the Medical Devices Advisory Committee. September 23, 2026.
GRAIL, Inc. Galleri Multi-Cancer Early Detection Test: Executive Summary. Molecular and Clinical Genetics Panel of the Medical Devices Advisory Committee. September 23, 2026.
Schrag D, Beer TM, McDonnell CH III, et al. PATHFINDER: A Prospective Cohort Study of Blood-Based Multi-cancer Early Detection. Lancet. 2023;402(10409):1251-1260. doi:10.1016/S0140-6736(23)01700-2.
Marinac CR, McDonnell CH III, Nadauld LD, et al. Clinical Evaluation of Cancer Signal Origin Prediction and Diagnostic Resolution Following Multicancer Early Detection Testing in the PATHFINDER Study. Cancer Prev Res (Phila). 2025;18(8):475-483. doi:10.1158/1940-6207.CAPR-24-0468.
Hanna TP, King WD, Thibodeau S, et al. Mortality due to cancer treatment delay: systematic review and meta-analysis. BMJ. 2020;371:m4087. doi:10.1136/bmj.m4087.
Canadian Institute for Health Information. Wait Times for MRI Scan. Updated June 2026.
NHS England. Diagnostic Waiting Times and Activity: Monthly Diagnostics Data 2026-27. NHS England; 2026.
The Role of ctDNA for Diagnosis and Histological Prediction in Early Stage Non-Small-Cell Lung Cancer: A Narrative Review. Diagnostics. 2025;15(7):904.
