Many cancer doctors hope that tests like Galleri, which are sometimes called “liquid biopsies,” will transform cancer care by detecting the earliest signs of malignancy, enabling treatment before the disease becomes deadly. They are part of an accelerating effort to find cancer earlier and earlier, through blood tests as well as new techniques and artificial intelligence.

The goal is to take the same energy and skill that has been deployed to develop treatments for cancer, and put it to work at finding cancer early enough to stop it in its tracks, said Dr. Betsy O’Donnell, director of the Multi-Cancer Early Detection Clinic at the Dana-Farber Cancer Institute.

“How do we move from a reactive model where we are only picking up cancers when they declare themselves?” she said. O’Donnell is leading a study of the Galleri test, recruiting 1,000 people at high risk for cancer.

Today, 70 percent of cancers have no screening test. People can be screened for cancers of the colon, breast, cervix, prostate, and, for heavy smokers, lung. Other tumors often grow without symptoms and are discovered when it’s too late to effectively treat them.

The path to catching more cancers early is fraught with challenges. To be most effective, screening has to do more than just detect cancer: It has to find the right cancer at the right moment. When a test turns up cancer so indolent it was never going to harm the patient, it only sows fear and can lead to needless procedures. When it finds a cancer that is so aggressive there’s no stopping it, its value can be limited.

“You really want a screening test that finds the cancer and leads to the patient not dying,” said Dr. Otis Brawley, a noted expert in cancer prevention and control at Johns Hopkins University, who has consulted for GRAIL.

When Burns White, 66, of Randolph, returned from the jazz festival, she was not immediately concerned. She’d been told that more often than not, a positive result from Galleri proves to be a false alarm. Her primary care doctor told her not to worry.

But the result was pretty specific: It said cancer might be lurking in her breast, esophagus, and stomach.

So Burns White consulted with doctors from the Dana-Farber Cancer Institute, where she works in community outreach and engagement. Two years ago, the hospital launched its Centers for Early Detection and Interception, which works with patients at high risk of cancer, those who receive troubling results from liquid biopsies, and others who learn of hard-to-interpret abnormalities from tests done for another purpose.

Burns White endured an endoscopy, an uncomfortable procedure involving threading a tube with a camera down her throat. The result was good news: no cancer of the esophagus or stomach. It would be a different story for breast cancer.

A mammogram she had taken a few months before had detected masses — but they looked benign. The results of the GRAIL test “pushed us to look closer,” said O’Donnell, of Dana-Farber.

Burns White underwent another set of mammograms, followed by MRIs and biopsies of both breasts. Doctors told her she had triple-negative breast cancer, a form that can be difficult to treat but that was still at an early stage.

Multicancer blood tests eventually will “change the face of early detection for cancer,” said Timothy R. Rebbeck, professor of cancer prevention at the Harvard T.H. Chan School of Public Health.

But they still need more work because, while they catch some cancers at an early stage, others are missed, he said. And even a screening tool that saves lives can have costs: follow-up testing and sometimes needless treatment, as well as “the psychosocial costs of having to go through all of this.”

Recent preliminary results from the largest study of GRAIL’s Galleri test showed its promise and limitations. Galleri found many types of cancer for which there is no screening test, and 92 percent of the time it correctly identified where in the body the cancer originated.

The test was administered to 36,000 people aged 50 and older. Among the first 25,000 analyzed, Galleri found a “cancer signal” in 216 cases — and cancer was confirmed in 133, about 60 percent. That compares favorably with mammography, said Megan P. Hall, vice president of medical and corporate affairs at GRAIL; she cited data showing that when a woman is called back for a follow-up mammogram, only 4 to 5 percent turn out to have cancer.

But there were more false negatives than true positives — 196 people were told the test found no cancer but went on to develop it within the next year.

And, even though its purpose is early detection, nearly half the cancer cases the test found were already advanced. That’s because advanced cancers shed more DNA and are easier to pick up.

The endgame in early detection may come down to better defining who is most at risk and thus most likely to benefit from screening.

For example, pancreatic cancer is almost always detected at an advanced stage. But it’s so rare that it would make no sense to screen everybody. Researchers are looking for ways to identify a subset of high-risk people who would benefit from screening,

Dr. Limor Appelbaum, a staff scientist at Beth Israel Deaconess Medical Center, is deploying AI to comb through the medical records of millions of patients, looking for patterns common among those who went on to develop pancreatic cancers, such as changes in blood chemistry or more frequent doctor’s visits.

At Dana-Farber, researchers have identified one possible warning sign of pancreatic cancer: muscle loss. By studying blood samples from people who developed pancreatic cancer going back years before their diagnosis, and also looking at CT scans, researchers led by Dr. Brian Wolpin and Dr. Michael H. Rosenthal found evidence that the pancreas starts to falter, performing its key functions less well, long before any tumor forms. This weakens the body’s ability to absorb proteins in the diet, causing it to break down muscle. But many factors can cause muscle loss, so further study is needed before these measures can be put to use.

At Massachusetts General Hospital, doctors are working on a warning system for lung cancer. An AI program called Sybil scanned thousands of chest X-rays of people who were later diagnosed with the disease and learned the patterns seen before someone develops cancer in the lung, said Dr. Lecia Sequist, program director of the Early Detection and Diagnostics Clinic at the Mass General Brigham Cancer Institute.

“My vision for 25 years from now,” said Dana-Farber’s O’Donnell, “is that when you go to your primary care physician’s appointment, they say, ‘This is your risk score, and here are the screening tests you need.’”

Dr. H. Gilbert Welch of Brigham and Women’s Hospital, who calls himself a “healthy skeptic” of cancer screening, would like to see the opposite: “If you want to lower the burden on patients,” he said, “develop the test that tells people they’re at low risk for cancer and [can] stop worrying about it.”

As for Burns White, she’s happy with how the Galleri test played out for her. Once her breast cancer was diagnosed, she underwent numerous rounds of chemotherapy, a lumpectomy, radiation therapy, and immunotherapy. Today, a little over a year since her diagnosis, she continues to take a medication aimed at eradicating the remaining cancer cells.

Through it all, Burns White continued to work, taking off only a few days for the surgery. She expects her treatment to end in mid-December.

“I remain strong and optimistic,” she said, “and grateful to have had access to treatment.”

Felice J. Freyer can be reached at felice.freyer@globe.com. Follow her @felicejfreyer.