Multicancer Early Detection Blood Tests Current Evidence and Limitations

Recommended screening in the United States covers five cancer types: breast, lung, colorectal, cervical, and prostate. Pancreatic cancer, ovarian cancer, liver cancer, and bile duct cancer have no routine screening pathway. These cancers are typically found only after a patient develops symptoms, and by that point the tumor has often advanced past the stage where surgery or other curative treatment is still possible.

The biology behind multi-cancer early detection testing offers a plausible way around that gap. In principle, a single blood draw can screen for signals from many cancer types at once, rather than testing for one organ system at a time the way a mammogram or colonoscopy does. Klein and colleagues, in a 2022 summary of the field, described the appeal in direct terms: a test that can agnostically detect a cancer signal with good sensitivity and a low rate of false alarms, predict where in the body the signal originated with a high degree of accuracy, and catch some of the most lethal cancers that have no screening test to call their own.

That is the pitch, and it is a serious one, aimed at the exact place where current medicine has the least to offer a patient before symptoms arrive. Whether the tests built on this biology actually deliver on that promise, at a level of evidence strong enough to change how cancer screening is practiced, is the question the rest of this piece works through.

Reading a cancer signal from a blood sample

What an MCED test can and cannot find starts with a basic constraint of the biology: cell-free DNA circulating in blood arrives in short, fragmented pieces, not intact chromosomes. The more regions sampled, and the more independent the information each one carries, the smaller the tumor signal the test can plausibly catch.

Galleri, the MCED test developed by GRAIL, takes a methylation-based approach. Klein and colleagues describe a classifier built around more than one hundred thousand independently informative regions, each covering multiple CpG sites, the specific spots on DNA where methylation marks accumulate. The classifier was trained and validated using data from the Circulating Cell-Free Genome Atlas study (NCT02889978), the large analytical validity study that underpins Galleri's development.

Abbott's Cancerguard test takes the biology a step further by combining more than one class of biomarker. In early-stage disease specifically, many of the positive signals come from just one biomarker class at a time, with methylation and protein contributing independently rather than always agreeing with each other. That pattern is the rationale for stacking biomarker types: it offers two separate chances to catch a faint signal rather than one.

Both Galleri and Cancerguard share a design choice that shapes everything about how they perform in practice. A test built to avoid flagging healthy people will, by the same mechanism, miss some early cancers. That trade-off is the central design decision any screening test has to make, and it is the lens through which the sensitivity numbers and false-positive rates in the clinical studies that follow need to be read.

What the largest interventional trials found

Three major interventional studies now provide the evidence base for MCED testing, and together they show real cancers caught that standard screening would have missed, alongside a performance picture that varies by stage and by cancer type. None of the three has yet shown that an MCED test reduces cancer mortality.

PATHFINDER 2, the Galleri trial, published results on September 22, 2026, in Nature Medicine. Adding Galleri to standard screening substantially increased the number of cancers caught through screening. Among the cancers the blood test found, nearly half were diagnosed at stage I or II, and nearly three-quarters at stage I through III, stages at which treatment options are considerably broader than at stage IV. Seven in ten of the cancers Galleri newly detected were types with no national screening recommendation, the exact population the test was built to reach. One case from the trial illustrates what that combination can mean in practice: Franklin Weingarten, a retired physician in Oregon enrolled at OHSU at age 84, had cholangiocarcinoma, a bile duct cancer with no routine screening test and one typically found only after symptoms appear and the disease has become inoperable, caught by Galleri while he was still asymptomatic. Surgeons removed part of his liver, and at age 88 he remains cancer-free.

DETECT-A, the forerunner to Abbott's Cancerguard and built on the CancerSEEK technology, was the first large, prospective, interventional study of a blood-based MCED test. It enrolled a large cohort of women with no history of cancer and identified nine cancer types, several of which have no routine screening option. After a median follow-up of roughly four years, every patient treated for a stage I or stage II cancer found through the study remained alive and cancer-free. The multiyear outcomes from that study, published in 2024 in Cancer Prevention Research, earned the authors the AACR Cancer Prevention Research Award for Outstanding Journal Article.

The NHS-Galleri trial tells a more complicated story, and it is the first large randomized controlled trial of an MCED test to report results. The market's reaction to the primary result was swift. GRAIL lost roughly half its valuation in February 2026 after the findings were announced.

Read together, the three trials show a technology that reliably finds real cancer, including cancer in organs with no screening infrastructure today, while also showing that the relationship between finding more cancer and improving outcomes at the population level is not yet settled. The NHS-Galleri result, taken alongside the stronger PATHFINDER 2 and DETECT-A data, sets up the question the next section has to answer.

Stage shift versus proven mortality benefit

Catching more cancers, or catching them earlier, does not by itself prove that a screening program saves lives. That distinction is the pivot on which the entire MCED evidence debate turns, and the evidence base for these tests has not yet crossed it.

The gold-standard endpoint for judging any cancer screening program is a reduction in cancer-specific and all-cause mortality, measured in a randomized controlled trial. No completed RCT has shown that an MCED blood test reduces cancer mortality in a general population. But the surrogate itself is contested. A 2024 analysis of randomized cancer-screening trials found that the relationship between a reduction in late-stage cancer incidence and a reduction in cancer-specific mortality varies by cancer type, which makes late-stage incidence an unreliable stand-in for mortality across the board rather than a dependable proxy for every cancer an MCED test might find.

Two well-documented biases explain why a surrogate can mislead even when the underlying test is working as intended. Overdiagnosis bias cuts the other way: an MCED panel covers many cancers with very different natural histories, and a pooled stage-shift number averaged across all of them can hide a genuine benefit in one cancer type while masking harm, or no benefit at all, in another.

The NHS-Galleri trial's missed primary endpoint is the clearest demonstration of this problem in practice. Galleri found real cancers in that trial. It did not produce a statistically significant reduction in late-stage diagnoses during the first round of screening, and mortality data from the trial are not yet mature enough to settle the question either way.

None of this means the RCT framework is free of costs of its own. Waiting fifteen years for mature mortality data before making any decision about access is also a policy choice, and it has a price: patients who might benefit from early detection today go without it while the trial runs its course. The evidentiary gap between stage shift and mortality benefit is real and well documented, and recognizing the cost of closing that gap slowly is part of weighing it honestly.

The false-positive problem and downstream workup

A low false-positive rate is not the same as zero false positives, and what happens to a patient after a positive result is where the ledger of harm and benefit actually gets written. Even a well-designed MCED test tuned for high specificity will generate false positives once it is used across a screening-scale population, and the burden that follow-up testing places on the people who receive a false alarm is a real cost that has not yet been fully measured.

PATHFINDER 2 reported a very low false-positive rate. At the scale of a national screening program, though, even a very low rate translates into a large absolute number of people referred for follow-up workup who turn out not to have cancer. A systematic review of false-positive MCED results published in 2026 found that the studies it examined often lacked detailed, standardized accounts of what diagnostic steps were used to resolve a positive signal. The magnitude of harm from the biopsies, imaging studies, repeat blood draws, and the anxiety that a positive signal produces while a diagnosis is pending has not been established. That remains an open question in the published literature rather than a settled one.

Overdiagnosis sits alongside false positives as a related but separate risk. A cancer an MCED test detects may grow so slowly that the patient was never going to be harmed by it, meaning earlier detection adds treatment and anxiety without adding benefit. Cancerguard's case-control data offer one encouraging data point on the false-positive side specifically: none of the false-positive results in that dataset were positive on both the methylation and protein biomarkers at once, which suggests a multi-biomarker design may cut down on false alarms. The downstream consequences for the false positives that do occur still need direct study.

Context matters here. PATHFINDER 2's positive predictive value means a positive Galleri result carries a meaningfully higher likelihood of being a true cancer than many traditional screening tests manage. That confidence level is a real strength of the technology, and it belongs in the same frame as the false-positive discussion rather than apart from it, because it is the reason a positive result justifies prompt follow-up even while the harms of that follow-up remain only partly understood.

The regulatory and coverage landscape as of late 2026

Commercial marketing, FDA review, and Medicare coverage legislation are all moving forward on their own timelines, even though no MCED test currently holds FDA clearance or approval and no major clinical guideline body recommends routine use of one.

As of August 2026, the FDA has not cleared or approved any MCED test. GRAIL submitted the final module of its premarket approval application for Galleri in January 2026, and an FDA advisory committee reviewed the supporting evidence, including the PATHFINDER 2 results, in late September 2026. A submission under advisory committee review is not the same thing as an approved product, and Galleri remains in that review stage rather than on the market with FDA clearance. Cancerguard, meanwhile, is marketed today as a laboratory-developed test, a regulatory category that has not been cleared or approved by the FDA or by any other regulatory authority.

Federal legislation passed in February 2026 established a pathway for Medicare to eventually cover MCED tests, but the law specifies that the federal agency overseeing that coverage cannot begin it before 2028, timed to line up with the FDA's anticipated approval schedule. No clinical practice guideline developer, including the American Cancer Society and the US Preventive Services Task Force, currently recommends MCED testing for cancer screening in any population. Guidance for primary care doctors emphasizes shared decision-making with patients who are interested in testing, rather than a standing recommendation, and is explicit that an MCED test should not replace established screening like mammography or colonoscopy.

What a positive or negative MCED result can tell

A positive MCED result is a strong signal that follow-up is warranted. It is not a diagnosis. A negative result does not substitute for established screening like mammography, colonoscopy, or a low-dose CT scan for lung cancer. Both limits come from the current state of the evidence itself, a limitation shared by Galleri and Cancerguard alike.

A positive result carries real weight. PATHFINDER 2 reported a positive predictive value of 60.3% for Galleri. When the test comes back positive, a confirmed cancer is more likely than not to be present. That is a meaningfully higher level of confidence than many traditional screening tests offer, and it is reason enough to move promptly to follow-up imaging rather than wait. What a positive result does not do is establish the cancer type, its stage, or a treatment plan. It opens a diagnostic process. It does not close one.

A negative result carries its own limit, and it is just as important as the positive one. Given the sensitivity constraints built into these tests, especially for early-stage disease, a negative MCED result cannot be read as clearance from cancer, and it provides no basis for skipping a mammogram, a colonoscopy, or any other screening test with an established mortality benefit behind it. The absence of a signal on a blood draw reflects the limits of what the biology allows the test to detect at a given tumor burden.

The evidence assembled so far is strong on analytical validity and on finding real cancers that current screening cannot reach, but still incomplete on the question of whether finding those cancers earlier translates into longer lives at the population level, and that gap sets how much weight a clinician or patient can responsibly place on a single result today. That is the standard this technology has met, and it is the standard still ahead of it.