I tried the $950 blood test Silicon Valley thinks can revolutionize cancer screening

I watch as my crimson-red blood quickly fills two small glass vials, then gets packed into an ice-cold sleeve to be shipped off to a cancer-hunting lab.

I've just completed my first cancer blood draw. There's about two teaspoons worth here, and I'm shelling out $824 to get it analyzed. (A seasonal discount brought it down from the listed price of $950.)

The process is simple and close to painless. A phlebotomist comes to your house, draws a little blood, you check the paperwork to ensure everything is correctly labeled, and off it goes to a lab. It will be tested for more than 50 different cancers. I'll have results in about a week, my phlebotomist says.

My stomach does a few flips, which surprises me. I don't think there's any reason to be nervous. I'm 40, still 10 years shy of the age when the test is recommended. I don't have any cancer symptoms, and I'm not considered high risk. I don't have a family history of cancer. But the thought of something lurking in my seemingly healthy body is unnerving. What if something is amiss and I just don't know it yet? The wait for results begins.

I've taken the Galleri test from Grail, a blood-testing startup for early cancer detection, one of a handful of tech companies that are Silicon Valley's latest answer to our collective cancer anxiety.

The sales pitch is basically this: If you can spot cancer in your bloodstream right now, you might be able to do something about it.

A third of us will get cancer in our lifetimes. Many of us won't catch it as quickly as we'd like, and there are deadly cancers, like liver and pancreatic cancer, that aren't routinely screened for. Cancer is meanwhile becoming less of an older person's predicament. More young people are getting diagnosed with late-stage colon cancer, a trend researchers and doctors don't fully understand. More young women are getting lung cancer and uterine cancer.

This and similar blood tests aim to get ahead of all that with good tech. They're designed to detect a cancer signal — traces of cancer DNA shed into the blood — and identify where in the body the signal is coming from.

That's the idea, at least. In a laboratory study Grail conducted in 2021, it caught 51.5% of cancers overall. That's a coin toss. Its accuracy fell to 16.8% for stage 1 cancers. Disappointing results from several real-world rollouts, including from a major, three-year trial in the UK, have dimmed the company's outlook. The test is catching cancer, but it's not clear it's really saving lives. At least not yet.

As I await my results, I worry about bad news. But I also worry I won't learn much of anything at all.

Grail has its origins in the human sequencing boom of the early 2000s.

When the Human Genome Project released the first draft of our genetic blueprint in 2000, a wave of biotech companies sprang up, focused on making sequencing more accessible across research, medicine, and eventually consumer testing. Early players like 23andMe were followed by companies that offered more advanced clinical applications, like prenatal tests for Down's Syndrome and other genetic abnormalities.

One of these companies, the genome-sequencing giant Illumina, was among the first and biggest providers of non-invasive prenatal blood tests (NIPTs). In the mid-2010s, Rick Klausner, Illumina's chief medical officer and the former director of the National Cancer Institute, noticed something interesting. When a phlebotomist showed him some unusual chromosomal patterns she'd spotted in a series of prenatal tests, he guessed, correctly, that they could be signs of cancer in the mothers' bodies.

"I immediately thought this could be the breakthrough we've long looked for: a screening test for cancer," Klausner says.

Getting that idea to market would require spending a lot of money, likely to the tune of $1 billion, on development and testing. With that in mind, Grail was spun off in 2016.

A few years later, Illumina reacquired Grail for $8 billion and then spun it off again. Grail went public in 2024 with a $500 million valuation.

"I called it Grail because this is one of the holy grails of cancer: Can you create a blood test for early detection?" Klausner says.

Jeff Huber, a former Google executive who helped develop Google Maps and had served on Illumina's board, was tapped to be Grail's first CEO. As it happened, Huber's interest in cancer detection had just become deeply personal. His wife, Laura, died at age 47 of a late-stage cancer that doctors were slow to diagnose. Huber says he was excited by the idea of harnessing data to increase a patient's odds.

"When cancer is diagnosed late, you are starting behind, and it's an uphill fight the whole way," he tells me. "There are times where you can make progress. But, ultimately, the dominant outcome is you lose because of the starting point."

Huber, who left Grail in 2021, says he still does the Galleri test every year alongside his annual check-up as part of a broader bouquet of cancer screening options. Every other year, he also does a full-body MRI, which costs about $2,500 and can catch abnormal growths. "I do feel deeply offended by the current state of things where Grail is a rich people product," he tells me.

Grail filed for FDA approval of its Galleri test in late January. It scored a major win in February, with a new law that says that Medicare will cover FDA-approved "multi-cancer early detection tests" (MCEDs) starting in 2028, even though none of the tests have made it that far.

The one blood test for cancer screening that's FDA approved is Guardant Health's Shield test, but, so far, it only screens for colon cancer. A recent clinical trial found it detected around 83% of colon cancer cases in people over 45, and the National Cancer Institute is running a trial to determine if the test could eventually be used more broadly.

Other companies may not be far behind, though they're still gathering data. Cancerguard, the makers of Cologuard colon cancer stool testing tests, offers a $690 test that also screens for around 50 cancers and, thus far, performs about as well as the Galleri test. MCED tests are also offered through OneTest and Episeek.

In the meantime, Grail has partnered with lab-testing startups to enable easy, at-home tests at lower prices. Telehealth giant Hims and Hers began offering the Galleri test this year, as did Superpower, a San Francisco-based health coaching startup.

The question is whether Grail's technology, and the tech behind MCEDs in general, is as good as those promoting it say, or if it's more of a work in progress.

The American Cancer Society doesn't recommend MCED tests since "much still needs to be learned about the accuracy." The US Preventive Services Task Force, the independent expert panel that approves tests for insurance coverage, has yet to weigh in.

Several major rollouts of the Galleri test — one involving California firefighters; another done in collaboration with the UK's National Health Service — have yielded disappointing results.

In late 2022, the San Francisco Firefighters Cancer Prevention Foundation spent over $1 million to offer the Galleri test to 1,786 active and retired firefighters, hoping this new test might save some lives.

A cancer "signal" was detected in 11 tests, according to a foundation press release. Cancer was later confirmed in five of those cases. The diseases were late-stage in all five cases the test flagged and each of the patients have since died. In the other six cases where the Galleri test detected a "cancer signal," follow-up tests found no evidence of cancer.

Within six months, at least three other firefighters who had no cancer signal detected by the test were diagnosed with either melanoma, prostate cancer, or lymphoma.

"We were sort of disappointed," says Tony Stefani, president of the San Francisco Firefighters Cancer Prevention Foundation. "We were under the impression when we talked to them that this was a state-of-the-art type of new screening tool and it would definitely give us an opportunity to catch some of these cancers at their earliest stages."

The foundation now says it "cannot support the routine use" of the Galleri or similar tests due to "false reassurance, missed cancers, unnecessary follow-up procedures, or added stress."

"If they fine-tune this thing to where it does catch these cancers at early stages, like they said, then it's a phenomenal thing," Stefani says. "But right now, it's just not working."

When I asked Grail about this, a spokesperson pointed to the test's success in identifying cancers that aren't routinely screened for, as well as cancers that other screening has missed.

Klausner says it was never expected that the test would have "a positive predictive value of 100%," and that finding even one new cancer is a good thing.

"It's been very strange to me, the extraordinary skepticism of the cancer community," he says. "I'm cautiously optimistic that the performance is going to continue to hold up, and hopefully will demonstrate real meaningful clinical benefit."

Disappointing results have continued to pile up.

Preliminary results from a Grail trial, released in October and involving 23,000 people in the US and Canada found the test detected cancer in 133 people that was later confirmed. However, it missed 196 cases that were diagnosed by doctors within a year, and it produced 83 positive signals that could not be confirmed as cancer.

Then, in mid-February, just three weeks after Grail filed for FDA approval, Grail released preliminary results from a major trial of its test in the UK.

Grail had partnered with the UK's National Health Service on a controlled, three-year trial of 140,000 people between the ages of 50 and 77. The NHS had called the testing "potentially revolutionary" if it could reduce the number of cancers found at stage 3 and stage 4, when the chances of survival are lower, and the overall costs are higher.

When Grail announced it hadn't met that goal, its share price fell 50%. The full results won't be released until later this spring, and an NHS spokesperson told the BBC that "the NHS will carefully study the full results from this major trial in the coming months to help determine how blood tests like this could be used in the future."

Grail has pointed out that the test reduced the number of stage 4 cancer diagnoses by more than 20%, at least for a particular subset of 12 especially deadly cancers, when used a second and third time. Grail also said its testing had led to a fourfold improvement in the detection of breast, colon, lung, and cervical cancers.

Professor Richard Sullivan, the director of the Institute of Cancer Policy at King's College in London, calls Grail's response "incredibly disingenuous." He points out that the test simply didn't achieve its goal of preventing more stage 3 and stage 4 cancer diagnoses.

"The bar was so low, and they couldn't even show that," he says. "The answer is the test does not work. End of story."

Many cancer researchers still see exciting potential in blood testing, even if the results thus far haven't been as good as some had hoped.

"I actually still think that these tests are our best hope for bending the cancer curve," Dr. Scott Ramsey, a cancer researcher at Fred Hutch Cancer Center, told The Cancer Letter when Grail's NHS results were announced. Ramsey is a co-principal investigator for the National Cancer Institute's ongoing MCED trial, which is trialing products from Guardant and ClearNote, but not Grail. "We're looking at the results for a first-generation test, and perhaps a trial design that wasn't ideal," he told the outlet.

As someone who writes about health for a living, I spend a lot of time monitoring cancer research and talking to cancer patients. And the truth is, we don't always have great options.

Doctors can be dismissive of symptoms, and no screening test is perfect. Mammograms miss about one in every eight cases of breast cancer. Colonoscopies, which are now recommended for adults over 45, can miss about 5% of colorectal cancers.

The demand for more accessible and reliable screening options, from MRIs to blood tests to wearables, is enormous. Most of us want more detailed information about our health so we can take proactive steps to live longer, healthier lives. And it would be great if we didn't have to smash our boobs or stick a probe inside our butts, only to test for one cancer at a time, with imperfect results.

It's no wonder that people who can afford it will shell out for any promising new technology, even if the data is mixed or the potential benefit is marginal. And while we typically set a high bar for screening tests — they should save lives and have real preventative value — are we willing to bend that standard in the name of supporting innovation?

A week passes after the phlebotomist's visit, and there's still no word from Grail. For all my growing ambivalence about the test, I can't kick the feeling of uneasiness.

Finally, two weeks after my blood draw, my results arrive in an email from Grail. I wait to open them alongside my husband, just in case the news is bad.

Relief passes over me as I read the result: "no cancer signal detected."

Still, there's no detailed analysis to look at, no numbers to track, no data to keep on file or compare to future tests. I think about other blood tests and genetic workups I've done. They almost always included a range of "normal" values so you can see where you rank, or detailed a long list of genetic mutations you do or don't have. There's none of that detail here.

When I tell people I took the Galleri test, the first thing they usually ask is whether they should give it a try.

In weighing my response, I'm haunted by the story of Danielle Hoeg, a mom to three young kids who took the Galleri test at 43 after some routine blood work had come back with confusing results.

Hoeg had received the same email that I got: "no cancer signal detected." Still convinced that something was wrong, she kept digging and, a few months later, went in for a Prenuvo full-body MRI. The scan found stage one lung cancer. Doctors were able to quickly remove her tumor in a single procedure without any chemotherapy or radiation.

It's just one story. But I wonder about others who don't follow their gut as Hoeg did.

In the weeks since I got my results, here's what I've landed on. If it's peace of mind about cancer you're after, comprehensive full-body MRIs will tell you the most about what's going on in your body right now, for about two and a half times the price of a Galleri draw.

The number one thing you can do — and this is something I've heard from nearly every cancer patient I've interviewed — is to talk to your doctor as soon as something feels amiss. Don't wait months or years to go in. And if they dismiss your concerns, go get a second opinion.

I certainly wish we could do more to both catch cancers earlier and treat them more easily. In the absence of a perfect solution, there are at least some imperfect ones out there.

Hilary Brueck is a health correspondent at Business Insider.