On September 23, 2026, the field of multi-cancer early detection reached a significant milestone. The U.S. FDA's Molecular and Clinical Genetics Panel reviewed GRAIL's Galleri multi-cancer early detection test and delivered a clear verdict through three key votes.

The panel voted 10-0 that Galleri has a reasonable assurance of safety, 6-4 that its effectiveness has reasonable assurance, and 7-2 (with one abstention) that the clinical benefits of Galleri outweigh its risks. While the FDA's final decision on nationwide sales approval is still expected in the coming months, the advisory panel's endorsement has already significantly boosted market confidence. As Guggenheim analyst Subbu Nambi noted, this has essentially "eliminated the risk".

The concept of detecting more than 50 cancer types from a single blood draw sounds like something out of science fiction. However, those with industry insight recognize that the most critical aspect of this shift is the pivotal role AI and data are beginning to play in cancer screening.

Abandoning the Traditional Path: Selling an AI-Powered Cancer Identification System

When discussing liquid biopsy, many think of searching the blood for specific genetic mutations released by tumors. Galleri, however, takes a distinctly different approach. When tumor cells die, they release cell-free DNA (cfDNA) into the bloodstream, and this DNA retains tissue and tumor-specific methylation patterns. Galleri's core technology involves targeted methylation sequencing to read these signals, which are then processed by a machine learning model. This model not only determines if a cancer signal is present in the sample but also predicts where in the body the signal is most likely originating from.

From a commercial and technical standpoint, GRAIL is no longer a traditional "diagnostic reagent company". What it truly sells is an AI-based cancer identification system composed of molecular testing, algorithmic classification, and clinical interpretation.

35% Sensitivity and 99.85% Specificity: A Philosophy of Restraint

If the expectation is "one blood draw to detect all cancers," the Galleri data might surprise you. In the U.S. PATHFINDER 2 study, Galleri's overall 12-month sensitivity was approximately 35.0%, which means a significant portion of cancers were not detected during the test. So why did the FDA panel still vote in favor? The answer lies in another remarkable statistic: 99.85% specificity.

In screening populations of tens of millions, or even hundreds of millions, of healthy individuals, the greatest danger is not a "missed diagnosis" but a "false positive". Even a small percentage of false positives, when scaled to a population level, could generate immense anxiety and lead to countless unnecessary CT scans, PET scans, endoscopies, and even biopsies, potentially overwhelming the healthcare system. Galleri's 99.85% specificity ensures it rarely creates a "false alarm". Additionally, its positive predictive value reaches approximately 77.0%, meaning that when it does signal an alert, the probability of a confirmed cancer diagnosis is already quite high.

"Prioritizing the control of false positives while precisely targeting deadly blind spots" is the product philosophy of Galleri. Data indicates that for 12 types of highly lethal cancers that currently lack established screening methods—including pancreatic, ovarian, and biliary tract cancers—and which account for about two-thirds of cancer deaths in the U.S., Galleri's sensitivity reaches approximately 61.3%.

The FDA panel explicitly emphasized that Galleri's role is intended to complement, not replace, standard screenings like mammograms and colonoscopies. Its target market is to fill the most evident gaps in the current cancer screening system.

The 140,000-Person Trial: Missing the Finish Line Yet Validating Real Value

The most significant controversy surrounding Galleri stems from the NHS-Galleri randomized controlled trial in the U.K., which involved over 140,000 participants. While the large-scale trial did not meet its pre-specified primary endpoint of reducing late-stage cancers overall, it did show a 22% and 26% reduction in the most fatal late-stage cancers during the second and third rounds of screening. The trial successfully demonstrated the ability to shift detection of deadly cancers from advanced stages to earlier stages.

This trial aimed to answer a fundamental public health question: does adding Galleri to the healthcare system lead to a decrease in late-stage (III/IV) cancers? From a strict statistical perspective, it did not reach its primary endpoint.

In the field of early cancer detection, "finding more" does not equate to "screening success" due to the risk of overdiagnosis. The true measure of success is "moving deadly cancers, which would otherwise only be exposed at stage IV, to stages I or II". This is the core reason why the FDA vote showed a 6-4 split on effectiveness, yet resulted in a 7-2 decision that the benefits outweigh the risks. The experts concluded that while it is not perfect, its potential to reduce late-stage cancers has crossed the threshold for clinical use.

Disrupting Old Rules: A New Screening Model for the AI Era and the Data Flywheel

For decades, cancer screening has been organized by organ: low-dose CT for the lungs, colonoscopy for the colorectal area. Galleri introduces a paradigm shift, changing the fundamental unit from "organ" to "person". Future physical exams may begin with a single blood draw for a full-body signal scan. If the AI flags an abnormality, the algorithm would then guide a targeted diagnostic workup for the indicated organ.

In this transformation, the opportunities for AI are just beginning: AI tracking over a longitudinal timeline can monitor dynamic changes. Cancer is dynamic, and if an individual provides blood samples for five consecutive years, a single signal might never reach the "positive" threshold. However, to an AI model, the subtle trajectory of change across those five years could already reveal a high-risk trend. This includes closing the decision loop after a positive result. Once a blood test indicates an issue, should a patient undergo a CT or an MRI first? When is it appropriate to stop testing? In the future, large language models and multimodal AI, integrated with medical records and family history, could become the most powerful external support for guiding physician decisions. There is also the un-copyable data flywheel. GRAIL currently holds data from the CCGA and PATHFINDER series, as well as real-world data from the over 140,000-person NHS-Galleri trial. When this methylation data is fully integrated with clinical outcomes such as patient survival and treatment modalities, GRAIL will possess an ever-evolving "real-world AI training platform" that presents a formidable moat difficult for any follower to cross.

Conclusion: An AI That Is Not Infallible, But a Future That Has Arrived

Of course, AI cannot create something from nothing. If an extremely early-stage, tiny tumor has not released enough cfDNA into the bloodstream, no algorithm, no matter how powerful, can detect it. To ensure safety at a population level, the Galleri test under FDA review currently uses "locked machine-learning classifiers", meaning it is not yet able to automatically update its learning on a daily basis in clinical applications.

Nevertheless, this expert vote on September 23, 2026, is far more than just a passage ticket for a blood-based diagnostic product. It demonstrated that simultaneously capturing signals from dozens of cancer types within the vast sea of blood is already technically feasible. If, in the coming years, Galleri can further demonstrate in the real world that it can reduce cancer mortality and succeed in making the cost economics of commercial deployment work, then today we stand at the dawn of a new era in cancer screening, hearing the first clarion call of the AI age.