FDA Panel Backs GRAIL's Galleri Multicancer Early Detection Test in Split Votes
核心洞察
The FDA's Molecular and Clinical Genetics Devices Panel voted 7-2 with one abstention that the benefits of GRAIL's Galleri (搜索) multicancer early detection test outweigh its risks.
Panel members voted 10-0 that Galleri (搜索) is safe and 6-4 that there is reasonable assurance the test is effective for its proposed indication in adults 50 and older.
NHS-Galleri (搜索) results published in the New England Journal of Medicine found no significant reduction in stage III or IV cancer (搜索) incidence after three screening rounds.
The FDA's Molecular and Clinical Genetics Devices Panel of the Medical Devices Advisory Committee voted on September 23, 2026 to endorse GRAIL's Galleri (搜索) multicancer early detection (MCED) test, concluding in three separate votes that the blood-based screening assay offers more benefit than risk for adults aged 50 years and older.
The 10 voting members voted 10 to 0 that there is reasonable assurance Galleri (搜索) is safe for use in patients meeting the proposed indication, 6 to 4 that there is reasonable assurance the test is effective, and 7 to 2 with one abstention that the benefits of Galleri outweigh its risks. The FDA is not bound by advisory committee recommendations but typically takes them into consideration. A final decision on the premarket approval (PMA) application, submitted on January 29, 2026, is expected in the coming months.
"Today's vote reinforces the strength of Galleri (搜索)'s clinical evidence," said Josh Ofman, MD, MSHS, CEO at GRAIL. "We believe the panel's recommendations reaffirm a high evidence bar that GRAIL has set for a multi-cancer (搜索) early detection test."
The endorsement was not unanimous in its reasoning. Panel member Dr. Victor van Berkel of UofL Physicians in Louisville, Kentucky, framed his support cautiously. "Perhaps my yes is a hopeful rather than an accurate one," he said, adding that more research is needed even if the test is rolled out. "If we can find more early-stage cancers in the population that we can't screen for currently, that will end up having benefit."
Proposed Indication and Test Design
GRAIL has proposed that Galleri (搜索) be indicated for screening for the early detection of multiple types of cancer (搜索) in adults 50 years or older, and for prediction of Cancer Signal Origin in individuals with a cancer signal detected result. Galleri is a next-generation sequencing-based test that identifies DNA methylation patterns on cell-free DNA (搜索) (cfDNA) in plasma. Whole blood is collected prior to plasma isolation, cfDNA extraction, and bisulfite conversion of the extracted cfDNA.
The test returns three classes of reportable results: a cancer (搜索) signal classifier that determines a result as cancer signal detected (positive) or no cancer signal detected (negative); a cancer signal origin classifier that predicts the likely anatomic location of a detected signal, returned only with a positive result; and a cancer signal origin supplemental classifier that predicts cancer biology such as histologic type and cellular lineage, also returned only with a positive result.
The FDA's executive summary, released before the meeting, listed precautions and limitations: Galleri (搜索) is not a replacement for existing recommended single-cancer (搜索) screening tests or diagnostic modalities; a no cancer signal detected result does not rule out cancer; and a cancer signal detected result still requires diagnostic testing with a medically established procedure to confirm the presence of cancer. The document also noted that the test may produce false positive or false negative results and that the benefits and risks of programmatic cancer screening with Galleri are still being studied.
Galleri (搜索) received breakthrough device designation in August 2018 and has been granted two investigational device exemptions, the first in September 2021 based on the PATHFINDER 2 study (NCT05155605) and the second in November 2022 based on the NHS-Galleri trial (NCT05611632).
NHS-Galleri: No Reduction in Late-Stage Incidence
Results from NHS-Galleri (搜索) were published in the New England Journal of Medicine on September 22, 2026, the day before the panel meeting. The trial evaluated the incidence rate of detecting cancer (搜索) with Galleri MCED testing plus usual care versus usual care alone. A lower incidence rate of stage III or IV cancers across 12 prespecified cancers was not observed following three screening rounds in the intervention group.
Among 142,250 patients who underwent randomization, 71,034 were clinically eligible in the intervention group and 71,033 in the control group. Of these, 3,637 patients in the intervention group and 3,400 in the control group received a diagnosis of a routinely staged cancer (搜索); 1,421 and 1,298 patients, respectively, were diagnosed with one of the 12 prespecified cancers.
After three rounds of screening, 706 and 688 patients, respectively, were diagnosed with stage III or IV cancer (搜索) among the 12 prespecified cancer types, corresponding to an incidence rate per 100,000 person-years of 300.5 (95% CI, 278.7-323.5) in the intervention group and 292.5 (95% CI, 271.0-315.1) in the control group. No significant difference was observed between the groups (incidence rate ratio, 1.03; 95% CI, 0.92-1.15; P = .63).
A post hoc analysis indicated the overall result was driven by a higher incidence rate of stage III or IV cancers in the intervention group during the first screening round (incidence rate ratio, 1.19; 95% CI, 0.98-1.43). In the second and third rounds, incidence rate ratios were 0.95 (95% CI, 0.77-1.17) and 0.88 (95% CI, 0.73-1.07), respectively.
For patients with the 12 prespecified cancers, the incidence rate of stage IV cancer (搜索) per 100,000 person-years was 145.2 (95% CI, 130.2-161.4) in the intervention group and 168.5 (95% CI, 152.3-185.9) in the control group, an incidence rate ratio of 0.86 (95% CI, 0.74-1.00).
In a post hoc analysis of early-stage diagnoses across the 12 prespecified cancers after three screening rounds, stage I or II cancers were identified in 647 patients in the intervention group and 559 in the control group (relative risk, 1.16; 95% CI, 1.03-1.30); stage I, II, and III cancers were reported in 1,007 and 846 patients, respectively (relative risk, 1.19; 95% CI, 1.09-1.30). Results were similar for patients with all routinely staged cancer (搜索) types.
Trial-related, device-related, or combined adverse events occurred in 371 patients (0.52%) in the intervention group and 321 (0.45%) in the control group; none were considered serious. Device-related adverse events occurred in 27 patients (0.04%) in the intervention group. The most common trial-related adverse events across both groups were vessel bruises at the puncture site, syncope, dizziness, vessel hematomas at the puncture site, and anxiety.
Eligible patients were 50 to 77 years old, had no cancer (搜索) diagnosis or treatment within three years prior to enrollment, and were not being assessed for possible cancer. The 12 prespecified cancer types were lung, head and neck, colon or rectum, pancreas, multiple myeloma or plasma cell neoplasm, liver or bile duct, stomach, esophagus, anus, lymphoma, ovary, and bladder, which account for more than 60% of cancer deaths.
PATHFINDER 2: Sensitivity and Predictive Value
PATHFINDER 2 results were published on September 22, 2026 in Nature Medicine. The study evaluated Galleri (搜索) in patients 50 years or older without clinical suspicion of cancer (搜索). A total of 32,007 enrolled patients completed their 12-month cancer assessment. Of them, 173 had true-positive results, corresponding to a detection rate of 0.54%, a positive predictive value of 60.3%, and a negative predictive value of 99.2%. Specificity was 99.64%, and the 12-month episode sensitivity was 39.3% for all cancers and 69.8% for a prespecified subgroup of 12 cancers.
A total of 267 patients had false-negative results, of whom 163 had United States Preventive Services Task Force grade A/B/C-recommended screening. Among all 35,335 patients in the safety analysis set, the proportion who underwent an invasive procedure after a positive test result was 0.6% (213/35,335), or 0.01 invasive procedures per patient. Nearly twice as many patients with true-positive results underwent invasive procedures during targeted evaluation compared with those with false-positive results (88.6% versus 46.3%). A total of 114 patients received a false-positive result, 59 fewer than the 173 true positives. Four patients with false-positive results underwent a surgical procedure to clear suspicion of cancer (搜索); benign neoplasms were found in all cases.
Patient-reported anxiety increased temporarily among patients with true-positive results and returned to baseline by 12 months. Anxiety did not increase among patients with false-positive or false-negative results. Eligible patients were 50 years or older, capable of giving signed and legally effective informed consent, and not undergoing or referred for diagnostic evaluation due to clinical suspicion of cancer (搜索).
Panel Debate Centers on Missed Cancers
FDA staffers and outside experts spent hours debating the technology's limitations. Risks cited included that the test would miss a significant number of cancers while also flagging potential cancer (搜索) in healthy patients, and that it was less accurate at detecting early-stage cancers. GRAIL did not provide data showing that screening with its test led to fewer cancer deaths, a metric that would have required years of follow-up.
FDA reviewers at the meeting reframed the company's data to emphasize potential downsides. The agency's chief reviewer noted that the test missed two of three cancers diagnosed over the coming year, giving false reassurance to many patients. GRAIL's primary study in 140,000 older adults showed the test detected about one third of cancers diagnosed over a one-year period, and about two-thirds of patients with a cancer (搜索) signal detected result were later confirmed to have cancer.
The panel declined to describe the test as providing early detection, saying it failed to show effectiveness for that term. Members also specified precautions that physicians and potential patients should be told: the test must be used in addition to routine cancer (搜索) screenings rather than as a replacement, a negative test does not guarantee the absence of cancer, and more study is needed to prove whether the test is as beneficial as mainstay screenings such as mammograms.
Consumer advocates raised concerns about behavioral consequences. "When they start having symptoms there will be tendency to delay going to the doctor, just a little bit, and that could be deadly," said Diana Zuckerman of the nonprofit National Center for Health Research, who spoke during the public comment period. "They will be told to continue other screening, but will they?"
Representatives for GRAIL argued the test can pick up signs of more than 50 cancers, including dozens for which no screening options exist, such as liver and ovarian cancer (搜索). "Patients want opportunity, they want options," said Dr. Mylynda Massart of the University of Pittsburgh, who presented on behalf of GRAIL. "They want the chance to fight their disease while the chance to survive is possible."
GRAIL said more than 200 physicians, nurses, professional societies, patients, and patient advocates shared support for Galleri (搜索) through the written public docket and open comment period during the panel. The company describes Galleri as the only MCED test supported by large interventional and randomized controlled studies in intended-use populations, including the PATHFINDER 2 and NHS-Galleri IDE studies. In clinical studies, GRAIL reports that Galleri has detected approximately four to seven times more cancers compared with standard of care alone, with most Galleri-detected cancers found at earlier stages while maintaining a low false-positive rate and high Cancer (搜索) Signal Origin prediction accuracy.
Galleri (搜索) is currently available in the United States, typically at a cash price of $700 or more. FDA approval would likely widen availability by opening the door to coverage from Medicare and private insurers.