The Test Sold While the Trial Was Still Running
Last year, 185,000 Americans paid $949 each for a blood test called Galleri that promised to detect more than 50 types of cancer early [1]. At the same time, 142,000 NHS patients in the UK were enrolled in a clinical trial designed to determine whether the test actually works [1]. The trial results, presented at the American Society of Clinical Oncology's annual meeting in Chicago, show it failed its primary goal: no statistically significant reduction in advanced stage three and four cancer diagnoses [1][2]. Yet the test remains on the market, sold under a regulatory loophole that allows laboratory-developed tests to bypass FDA approval entirely [4].
This isn't an anomaly. It's how the system works. Medical innovation can reach hundreds of thousands of paying customers before the evidence proving it helps them is collected, analyzed, or published. The gap between "available for purchase" and "proven to work" isn't measured in weeks, it's measured in years and hundreds of millions of dollars in revenue.
What the Trial Actually Showed
The NHS-Galleri trial was the world's first randomized controlled trial of a multi-cancer early detection test [1]. Between 2021 and 2023, researchers enrolled 142,000 participants aged 50 to 77 with no cancer symptoms [1]. Half received annual blood draws analyzed with the Galleri test, which looks for minuscule shards of cancer DNA in the blood [4]. The other half served as controls. The primary endpoint was straightforward: reduce late-stage diagnoses, stages three and four, in a pre-specified group of 12 cancers [1].
The test didn't meet that goal [1]. But buried in the secondary analyses was a different signal: stage four cancers alone dropped 14% in the Galleri group [1]. These are the deadliest cancers, the ones that have already spread beyond their original site. Grail, the California-based company that developed Galleri, said it was "encouraged" by this finding [5]. Prof Richard Houlston, head of the division of genetics and epidemiology at the Institute of Cancer Research in London, reached a different conclusion: "There is no evidence base to justify implementation of Galleri at a population scale" [1]. A senior cancer figure was blunter: "The trial flopped" [1].
Both statements are true. The pre-specified primary endpoint failed. A secondary observation showed a reduction in the most lethal cancer stage. How you interpret this depends largely on what you're selling and who bears the risk of being wrong.
The Revenue Model Built on Uncertainty
Grail has sold nearly half a million Galleri tests since launching in the United States in 2021 [4]. At $949 per test, that represents roughly $474 million in revenue from an intervention whose clinical benefit remains unproven. Very few insurers cover the test, so most buyers pay out of pocket [4]. The company marketed Galleri aggressively, including a Super Bowl advertisement [4], while the trial that would validate its core claims was still enrolling patients.
This was possible because Galleri is classified as a laboratory-developed test, a category that exists in a regulatory gray zone [4]. These tests can be marketed without FDA approval or review [4]. The logic originally made sense: hospital labs needed flexibility to develop diagnostic tests for rare conditions without navigating a years-long approval process. But that loophole now allows companies to sell complex screening tests directly to consumers at scale, collecting revenue and real-world data simultaneously while the controlled trial meant to establish efficacy runs in parallel.
The incentive structure is clear. Every month the test remains on the market generates tens of millions in revenue. Every month the trial continues is another month of uncertainty about whether the test actually saves lives. Grail's chief scientific officer, Harpal Kumar, is the former chief executive of Cancer Research UK [1]. Prof Peter Johnson serves as national clinical director for cancer at NHS England [1]. These are not fringe actors or fly-by-night operators. They are established figures navigating a system that allows promising technology to reach patients before the evidence is conclusive.
What Patients Are Actually Buying
The 185,000 people who bought Galleri in the year before the trial results were announced [4] were not purchasing a proven intervention. They were purchasing a hypothesis: that detecting cancer DNA in blood before symptoms appear would lead to earlier treatment and better outcomes. That hypothesis is biologically plausible. It's also unproven.
The psychological stakes are high. A positive result triggers a diagnostic cascade, imaging, biopsies, specialist consultations, that may or may not find cancer. A negative result offers reassurance that may or may not be justified, since the test's sensitivity varies widely by cancer type. The financial stakes are equally concrete: $949 is more than many Americans have in emergency savings, spent on a test their insurance won't cover for a benefit that a 142,000-person trial just failed to demonstrate.
Meanwhile, the mortality data, the ultimate measure of whether early detection actually saves lives, won't be available for another couple of years, according to Prof Houlston [1]. By then, Grail will likely have sold the test to hundreds of thousands more people, each one paying nearly $1,000 to participate in what amounts to an uncontrolled, uncompensated extension of the clinical trial.
The System That Prioritizes Access Over Evidence
Other cancer screening tests followed a different path. Mammography underwent decades of randomized trials before becoming standard care. Colonoscopy was studied extensively in controlled settings before insurers agreed to cover it. The FDA approval process for diagnostic tests is slow, expensive, and risk-averse, but it ensures that by the time a test reaches widespread use, there's a substantial evidence base supporting the claim that it helps more than it harms.
The laboratory-developed test loophole creates a parallel system where market access comes first and validation comes later, if at all. Companies argue this accelerates innovation and gives patients access to cutting-edge technology. Critics counter that it shifts risk from companies to consumers, who pay out of pocket for tests that may offer no benefit and may cause harm through false positives and unnecessary procedures.
Grail now has real-world data from half a million customers and controlled trial data from 142,000 NHS patients. The controlled trial failed its primary endpoint but showed a signal in a secondary analysis. The company remains "encouraged." Independent researchers say there's no evidence base for population-scale implementation. And the test is still for sale, still marketed, still generating revenue while the definitive mortality data accumulates slowly in a dataset that won't be analyzed for years.
The question isn't whether Galleri might eventually prove useful, the 14% reduction in stage four cancers suggests it might [1]. The question is who should bear the cost and risk of that uncertainty while we wait to find out. Right now, the answer is patients with their wallets and their anxiety, while the company collects data and revenue simultaneously. That's not a bug in the system. That's the system working exactly as designed.
