The Blood Test That Finds Cancer but Can't Prove It Saves Lives
Grail's Galleri blood test found cancer four times more often than standard screening in a massive NHS trial. There's just one problem: it completely failed to prove that catching cancer earlier actually reduces late-stage disease.
A $3.5 Billion Bet Just Hit a Wall
Imagine building a smoke detector so sensitive it catches every whiff of trouble in your house. Sounds great, right? Now imagine the fire department shows up and finds that, despite all those early alarms, roughly the same number of houses still burn down.
That's essentially what just happened with Grail's Galleri blood test, the most ambitious cancer screening tool ever built. The NHS-Galleri trial enrolled roughly 140,000 people across England, gave half of them annual blood draws on top of their usual screening, and asked one big question: would catching cancer earlier actually reduce the number of people diagnosed at dangerous late stages?
The answer, presented at ASCO 2026, was a deflating "not really."
What the Trial Was Actually Designed to Prove
Galleri works by scanning cell-free DNA in your blood for methylation patterns (chemical tags on DNA that look different in cancer cells). It's designed to detect signals from dozens of cancer types with a single blood draw. Think of it as a cancer metal detector for your entire body.
The NHS-Galleri trial was built around a concept called "stage shift." The idea is straightforward: if you catch cancers at stage I or II (early, more treatable) instead of stage III or IV (late, much deadlier), fewer people should die. The trial's primary goal was to show a statistically significant drop in stage III and IV cancers combined across 12 of the deadliest cancer types, including pancreatic, ovarian, lung, and liver cancers.
It didn't.
The incidence rate ratio came in at 1.03 with a p-value of 0.63. For non-statisticians: that's about as far from meaningful as you can get. The number of late-stage cancers in the Galleri group was virtually identical to the control group.
The Good News Buried Inside a Bad Headline
Before you write Galleri's obituary, the story gets more complicated. Several secondary findings were genuinely encouraging.
Stage IV cancers did drop. The reduction grew with each screening round: 9% in the first year, 22% in the second, and . Overall, stage IV diagnoses fell by about 14% across those 12 deadly cancers. That's not nothing.
Get tomorrow's biotech intelligence before your competitors.
Join thousands of biotech professionals who start their day with our free, daily briefing.
26% by the third round
The test also caught cancer four times more often than standard NHS screening alone, with a detection rate of roughly 0.48% across all three rounds. And cancers diagnosed through emergency room visits (often a sign of very late detection) fell by 25% in the Galleri group.
So why did the primary endpoint fail if stage IV went down? Because stage III went up. Galleri essentially pulled some cancers out of the stage IV bucket and dropped them into stage III. The net effect on combined late-stage disease: zero.
It's like rearranging deck chairs on the Titanic, except the ship might actually be slightly more seaworthy. We just can't prove it yet.
The Spin Versus the Science
This is where things got contentious. Grail's communications leaned heavily on the stage IV reductions and the four-fold detection improvement, painting the results as a glass-half-full story.
Experts weren't buying it. Prof Richard Houlston of the Institute of Cancer Research said the findings were "presented far more positively than the overall results justify." One senior UK cancer leader, quoted by The Guardian, called the trial result a "flop, clear and simple."
The UK Science Media Centre issued a pointed critique, warning that Grail's messaging over-emphasized secondary findings while glossing over the failed primary endpoint. Their concern: in screening research, a "favorable trend" can sound almost as convincing as a statistically robust result, especially in a press release. But it should only be treated as hypothesis-generating, not proof.
ASCO's chief medical officer took a more diplomatic route, calling the trends toward tumor downstaging "encouraging" while stressing that the primary endpoint failure is what matters most right now.
The Deeper Problem: Finding Cancer ≠ Saving Lives
This trial exposes a fundamental tension at the heart of the liquid biopsy screening revolution. Detecting more cancers earlier sounds like an unambiguous win, but biology is messier than that.
For some cancers, earlier detection genuinely saves lives. Catching colon cancer at stage I instead of stage IV is transformative. But for others, like pancreatic cancer, even early-stage patients face limited survival with current treatments. Finding it earlier doesn't help much if the treatments haven't caught up.
There's also the overdiagnosis problem. Some tumors grow so slowly they'd never cause symptoms or death. Detecting them creates patients out of people who were fine, triggering anxiety, biopsies, scans, and sometimes unnecessary treatment. With a false-positive rate of about 0.4%, Galleri performs well on that front; better, in fact, than many existing single-organ screens. But at population scale, even small rates add up to real human impact.
Perhaps most sobering: one methodological study found that patients whose cancers were detected by MCED tests actually had worse overall survival than those whose cancers were MCED-negative, even at the same stage. The likely explanation is that these tests preferentially catch biologically aggressive tumors, the very ones that are hardest to treat regardless of when you find them.
What This Means for the $50 Billion MCED Race
Grail has poured roughly $2 billion into Galleri since 2016. It submitted a Premarket Approval (PMA) application to the FDA in January 2026. The NHS-Galleri data will be central to that review, and the primary endpoint failure is going to be a very awkward conversation.
NHS England has already signaled it won't rush into a national rollout. The health service had a conditional commitment to purchase up to one million tests, but only if strict thresholds were met (including a 30% reduction in stage IV cancers). Current data don't clear that bar.
Meanwhile, the competition isn't standing still. Guardant Health already has FDA approval for its Shield blood test for colorectal cancer screening and is expanding toward multi-cancer detection. Exact Sciences is running a 20,000-person trial for its Cancerguard MCED test, with an FDA filing expected around 2026 to 2027. Both companies are watching Galleri's stumble closely.
Grail does have real strategic advantages: its April 2026 partnership to embed Galleri ordering into Epic's electronic health records across ~450 health systems is a powerful distribution play. But distribution doesn't matter if regulators and payers aren't convinced the test actually helps people live longer.
The Verdict (For Now)
The NHS-Galleri trial proved that a blood test can find cancer. Lots of cancer, in fact. What it hasn't proven is the thing that ultimately matters: that finding it this way saves lives.
Mortality data from this trial won't arrive for years (follow-up extends to eight years post-randomization). Until then, most experts are united on one point: Galleri and similar MCED tests should stay inside clinical trials and research settings, not your annual physical.
The smoke detector works. The question is whether the fire department can actually do anything different when it arrives earlier. That's the billion-dollar answer we're still waiting for.
5 min read
The Lung Cancer Drug Where Half the Patients Are Still Winning After 7 Years
Seven years into a clinical trial, more than half of metastatic lung cancer patients on Pfizer's lorlatinib still haven't progressed. Oncologists are starting to whisper the word "cure," and the data might actually justify it.
