New blood test detects cancers 3 years before typical diagnosis, study hints

Blood plasma can harbor DNA changes that could flag cancer years before existing diagnostic tests, an early study hints.

The recent study, published May 22 in the journal Cancer Discovery, found traces of free-floating DNA from dead precancerous or cancerous cells in plasma that had been donated three years before a diagnosis.

"It's an important step toward preclinical cancer detection, which could potentially revolutionize cancer screening," said Catherine Alix-Panabières, a cancer researcher at the University of Montpellier in France who was not involved with the work. "Earlier detection typically correlates with better outcomes across many cancer types due to earlier intervention," she told Live Science in an email.

The prognosis for cancer patients generally grows worse the later their disease is caught, especially once it has grown and spread to other tissues. Yet the gene changes, or mutations, that give rise to tumors tend to appear decades beforehand. Consultant oncologist Dr. Yuxuan Wang at Johns Hopkins University and her colleagues wanted to see if they could detect tumor DNA in plasma long before cancer manifests.

They examined plasma — the liquid that blood cells are suspended within — that was collected from patients roughly 40 years ago for an unrelated study. They focused on 26 participants who had developed cancer within six months of donating blood, as well as 26 controls who did not develop cancer for at least 17 years post-donation.

Related: Simple blood tests could be the future of cancer diagnosis

Wang's group found between one and three common cancerous mutations in seven of the plasma samples, all of which were taken from participants that developed cancer within four months of donating blood.

Six of these patients had also donated blood between 3.1 and 3.5 years beforehand, so Wang's team turned back the clock further and assessed those earlier samples for the same mutations. Two of the early samples contained the same DNA errors, confirming that these warning signs were detectable years before the tumors appeared, at least in some people.

Since they found only a few common mutations in two of the six plasma samples taken three years before diagnosis, they then sequenced the plasma DNA to find additional mutations that were unique to each patient. Using the genomes of their white blood cells — a type of immune cell — as a reference, they found between four and 90 unique mutations in the plasma DNA from three patients. All told, they found hints of cancer in three of the five early samples they examined.

The patients in this study had a variety of cancers, including breast, colon, liver, lung, pancreas, and rectal cancer. However, it's not clear if the testing method works equally well for all tumor types. "Some organs will shed tumor DNA more than others," Wang told Live Science, noting that the blood-brain barrier, a protective membrane, may prevent brain cancer DNA crossing out of the organ and into the bloodstream.

In addition, the new research didn't find any cancer DNA in 18 of the 26 participants who developed tumors in the months after their samples were collected. That's not ideal for a clinical test, Wang said. But she suggested that detection could potentially improve if doctors took larger volumes of plasma from each patient.

Since the test could potentially detect cancer years before symptoms first appear, it could one day be useful for screening patients preemptively. However, further experiments are needed to ensure this diagnostic doesn't lead to false positive results, which could unnecessarily alarm patients and possibly lead to unnecessary treatments or invasive diagnostic procedures, like biopsies.

"Ethically, implementing such tests in routine screening would require clear guidelines on how to handle incidental findings," Alix-Panabières said.

And because the study only included plasma samples from 52 people, larger investigations involving hundreds or thousands of participants would be needed to validate the test before doctors could use it with confidence. "Realistically, widespread clinical adoption may take another 5–10 years," Alix-Panabières predicted.

Finding personalized mutations requires sequencing the patient's DNA, which can cost several hundred or thousands of dollars, Wang said. So even if such a test can be validated in larger trials, it's "probably not going to be something we can provide for everyone who we want to screen," and the test may need to be reserved for at-risk groups whose families have known histories of cancer, for instance.

The recent study consisted mostly of Black and white men and women between the ages 45 to 64 from four U.S. states. Future investigations could explore the efficiency of the test in people from other, genetically diverse backgrounds.