Blood test detects multiple cancers early, showing promise in landmark study

A blood test that finds cancer before symptoms appear
GRAIL's MCED test detected malignancy in asymptomatic participants, including early-stage cancers with no existing screening methods.
Mark

So this test found cancer in people who had no idea they were sick. How does a blood test even know to look for cancer?

Mimi

The test searches for DNA fragments and other signals in the blood that cancer cells shed. Researchers trained it to recognize patterns that point to specific cancer types—and to ignore the benign conditions that earlier versions kept flagging.

Luke

But we should be careful here. They detected a signal in 92 people and confirmed cancer in 35. That's a 38% confirmation rate. What happened to the other 57?

Mimi

That's a fair question. The study doesn't fully explain those cases—whether they were false positives, whether follow-up testing is still pending, or whether some people declined further investigation.

Mark

The test identified where the cancer came from with 97% accuracy. That seems almost too good to be true.

Mimi

It's genuinely impressive, but it's worth noting that accuracy in origin prediction is different from accuracy in detection itself. The test can tell you it's pancreatic cancer, but it still has to find the cancer in the first place.

Luke

And 70% of the cancers detected were types with no standard screening. That's either a huge opportunity or a sign that the test is picking up things we don't yet know how to treat.

Mimi

Both, probably. It opens a door to catching these cancers early, but it also means doctors will need new protocols for what to do once they find them.

Mark

What's the next step?

Mimi

Larger studies in more diverse populations. The current study was limited to people 50 and older. We need to know if it works as well in younger people, in different ethnic groups, and in people with different health backgrounds.

Luke

And we need to know the false positive rate in real-world use. A 1% detection rate sounds low, but if millions of people take this test, that's a lot of people getting told they might have cancer.

  • Cancers of the pancreas, liver, and small intestine are among the deadliest precisely because they hide until it is too late — this test found them while patients still felt fine.
  • Of 6,662 asymptomatic adults over 50, roughly 1% harbored a cancer signal, and 70% of confirmed diagnoses were for cancer types that have no existing screening protocol.
  • The test detected twice as many cancers as conventional screening methods, including early-stage tumors that would otherwise have gone unnoticed for months or years.
  • With 97% accuracy in predicting the tissue of origin, the tool moves beyond mere detection — it points doctors toward where to look, compressing the diagnostic timeline.
  • The study's sample is promising but limited, and broader validation across more diverse populations remains the critical next step before clinical deployment.

For generations, cancer has revealed itself on its own terms — through pain, weight loss, or a shadow on a scan — often arriving too late for easy remedy. A study presented in Paris in September 2022 suggests that a single blood draw may begin to change that calculus: GRAIL's Multi-Cancer Early Detection test identified cancers in people who felt entirely well, finding malignancies that standard screening would have missed, including early-stage tumors of the pancreas and liver that medicine has long struggled to catch in time. The test is not yet in clinical use, but its precision — 97% accuracy in identifying where a cancer originated — signals a possible turning point in how humanity confronts one of its oldest adversaries.

A blood test has found cancer in people who had no idea anything was wrong. In a study of more than 6,600 adults aged 50 and older — all asymptomatic — GRAIL's Multi-Cancer Early Detection test identified malignancy in roughly 1% of participants. The findings were presented at the European Society for Medical Oncology Congress in Paris in September 2022.

The test had been refined from earlier versions that struggled to distinguish true cancer signals from benign blood conditions. The updated tool proved far more precise: of the 92 people who received a positive signal, 35 were ultimately confirmed to have cancer, and the test identified the tissue of origin with 97% accuracy. Among those diagnosed were patients with Stage I liver, small intestine, and uterine cancers, as well as Stage II pancreatic and bone cancers — diseases that typically surface only after symptoms appear and treatment options narrow.

What set the results apart was the nature of the cancers found. Seventy percent were types for which no standard screening currently exists. There is no routine test for pancreatic or liver cancer in an otherwise healthy person; medicine waits for the patient to report symptoms, by which point the disease has often advanced. GRAIL's chief medical officer Jeffrey Venstrom noted that the test caught twice as many cancers as conventional screening would have.

The study's scale is real but modest, and the test has not yet entered clinical practice. Larger, more diverse trials are needed before it can reshape medical care. But the early signal is striking enough to command serious attention: if validated, a single blood draw could one day replace the fragmented, disease-by-disease approach to cancer screening — offering earlier detection, more treatment options, and a fighting chance against cancers that today so often win by arriving unannounced.

A simple blood test has caught cancer in people who felt perfectly fine. The discovery comes from a study of more than 6,600 people, all aged 50 and older, who had no symptoms of disease. Researchers at GRAIL, a healthcare company focused on cancer screening, ran what they call the Multi-Cancer Early Detection test—or MCED—on these participants and found malignancy in roughly 1% of them. The results, presented at the European Society for Medical Oncology Congress in Paris in September 2022, suggest the test may fundamentally change how doctors hunt for cancer before it announces itself.

The MCED test is not entirely new. Researchers refined earlier versions that had a problem: they picked up pre-malignant blood conditions that were common but not necessarily dangerous. The updated test narrowed its focus, learning to distinguish true cancer signals from benign noise. What emerged was a tool that appears to work with striking precision. Of the 92 people in whom the test detected a cancer signal, 35 were ultimately diagnosed with 36 confirmed cancers. The test identified the tissue of origin—where the cancer started—with 97% accuracy.

What makes this result remarkable is not just the detection itself but what was detected. Jeffrey Venstrom, the chief medical officer at GRAIL, noted that the test found twice as many cancers as standard screening methods would have caught. Among them were Stage I cancers of the liver, small intestine, and uterus, along with Stage II pancreatic, bone, and oropharyngeal cancers. These are cancers that typically announce themselves only after they have grown large enough to cause symptoms—by which time treatment becomes harder and outcomes worse.

The most striking finding concerns the types of cancer the test caught. Seventy percent of the diagnosed cancers were varieties for which no standard screening protocol currently exists. A person with no symptoms has no routine way to know if they are developing cancer of the pancreas, liver, or small intestine. Doctors wait for patients to report pain, weight loss, or other signs. By then, the disease has often progressed. The MCED test, by contrast, can find these cancers while they are still early and potentially more treatable.

The study's scope was limited—6,662 participants is substantial but not enormous, and all were 50 or older. The test has not yet been deployed in clinical practice or validated in larger, more diverse populations. But the early signal is clear enough that it has drawn serious attention from the medical establishment. If the test proves reliable in further studies, it could reshape cancer screening from a disease-specific approach—mammograms for breast cancer, colonoscopies for colon cancer—into a single blood draw that searches for multiple malignancies at once. The implications are profound: earlier detection, more treatment options, and potentially better survival rates across a range of cancers that today often go undetected until they are advanced.

MCED testing detected double the number of cancers than standard screening, including Stage I cancers of the liver, small intestine, and uterus, and Stage II pancreatic, bone, and oropharyngeal cancers.
— Jeffrey Venstrom, chief medical officer at GRAIL
Contact Us FAQ