The Atlantic: Cancer Supertests Are Here

[Colonoscopy]

Cancer Supertests Are Here​

But are they really such a good idea?
By Benjamin Mazer




It takes a certain amount of confidence to call your biotech company Grail. According to its website, the Menlo Park–based firm got its name because its “co-founders believed a simple blood test could be the ‘holy GRAIL’ of cancer detection.” Now the company claims that its “first-of-its-kind” screening tool, called Galleri, “redefines what’s possible.” At the cost of a needle stick and $949, the company can check your blood for more than 50 forms of cancer all at once.

The Galleri test, as well as many others of its type that are in development, is meant to sniff out malignant DNA floating in a person’s veins, including bits of tumors that otherwise might not be identified until they’ve spread. But the rapid introduction of this new technology, which is now available through major U.S. health systems, isn’t really guaranteed to help patients. Indeed, a contentious debate about its potential benefits has been playing out in the scientific literature for the past few years. Multi-cancer-screening tools—or “cancer-finding supertests,” as Galleri has been called—aren’t yet endorsed by the U.S. Preventive Services Task Force, or formally approved by the Food and Drug Administration. For the moment, health-care providers can offer Galleri only through a commonly used regulatory loophole that the government is desperately trying to close. Being able to distribute the company’s “prescription-only, well-validated test” in advance of full FDA approval is a good thing, Kristen Davis, a Grail spokesperson told me, because it gives patients “timely access to an important tool in the detection of unscreened cancers and allows for important real-world evidence collection.” That’s one way to look at it. Here’s another: The rush to get Galleri and related products into doctors’ offices skips right over the most important step in clinical development: proving that they really work.


“The status quo for cancer screening remains unacceptable,” Davis said. She’s right. Even traditional early-detection tests are controversial within the medical community. As a hospital pathologist who diagnoses cancer daily, I’ve seen firsthand how mammograms and Pap smears, among other traditional procedures, save some people’s lives—and also how they cause a lot of overtreatment. (They miss many lethal cancers, too.) Blood-based cancer screening, in particular, had an ignominious start. Most men middle-aged and older in the U.S. get PSA tests, which look for abnormal levels of a protein secreted from the prostate gland that may indicate malignancy. But many of the tumors those tests identify are slow-growing, harmless ones; their discovery leads to an epidemic of unnecessary surgery and radiation—and a subsequent epidemic of incontinence and impotence. Recognizing this harm, the scientist who first identified PSA more than half a century ago expressed his regret in 2010, calling widespread screening “a profit-driven public health disaster.”

Modern blood-based cancer tests (or “liquid biopsies”), which look for a tumor’s genetic material, have been more promising. The first was approved by the FDA in 2016. It allows patients who already know that they have lung cancer to avoid an invasive tissue-collection process while still receiving the right, targeted therapy for their particular disease. Today, liquid biopsies exist for other kinds of cancer, too, and are used to tailor treatment for people who are aware of being sick.

Unleashing the same technique on the general population, in an effort to find hidden cancers in healthy-seeming people, is in principle a reasonable idea. But in 2020, when Grail started trying its technology on thousands of adults without cancer symptoms, the company found that a majority of positive signals—the signs of potential tumors that it identified—weren’t real. Dozens of healthy participants were flagged as possibly having cancer; most suffered through unnecessary laboratory and imaging follow-up. One unlucky subject described in the published study even had his testicle removed in the hunt for a malignancy that didn’t exist. Another blood-based supertest called CancerSeek—which forms the basis of a multi-cancer test now under commercial development—had shown the same problem when an early iteration of it was studied in some 10,000 women: Registered blood “abnormalities” led to confirmed cancer diagnoses less than half of the time. False positives with CancerSeek caused some patients to have operations on their ovaries, colon, or appendix.

 

[Colonoscopy]

No form of cancer screening will be perfect, and Davis pointed out that “when used as recommended, in addition to current single-cancer screenings, the Galleri test can help screen for some of the deadliest cancers that often come with no warning today.” For cancers of the pancreas, ovaries, esophagus, and liver, she suggested, any form of screening will be better than what we currently have: nothing. Grail researchers have also noted that its technology “compares favourably” to other, more familiar single-cancer tests in the sense that a smaller proportion of patients end up with spurious results. (One in 200 people will experience a false positive with Galleri, while the same is true for about one in 10 women who get a mammogram.)

But an imperfect screening tool is not always better than no screening tool at all. We already have reasonably accurate early-detection tests for pancreatic and ovarian cancer, for example, but experts recommend against their widespread use because—counterintuitively—screening healthy patients does little to extend their lives and comes with its own harms. And although it is true that Galleri’s false-positive rate is quite good in comparison to those of mammograms, PSA tests, and Pap smears, that’s only half the story. A glitchy answer from a cancer supertest like Grail’s may well be worse than the equivalent mistake in, say, a breast exam. The latter would only lead to further hunting for a tumor in the breast—perhaps with an ultrasound or MRI. In contrast, the follow-up for a suspect finding from a screen for 50 different cancers could be body-wide, producing yet more ambiguous results—such as the discovery of kidney cysts or lung nodules—that generate their own tests and surgeries.

When Galleri finds a potential tumor, it does provide doctors with some hints as to where that tumor might be located. In practice, though, doctors will likely err on the side of running lots of tests. Positive signals are often followed by a PET-CT scan, for example, which costs about $2,500 and exposes people to 62 times the radiation of a mammogram. In Grail’s own research, participants who received a false-positive result were generally subjected to multiple additional lab and imaging tests—sometimes as many as 16 laboratory studies and 10 clinic visits.


Read: When cancer screening stopped

More thorough and extensive testing takes longer, too. An errant mammogram might be resolved fairly quickly, with conclusive follow-up testing done a few weeks later. The equivalent delay after an abnormal Pap smear is less than two months, generally speaking. In the aftermath of multi-cancer blood-test screenings, though, worried patients may have to bide their time for almost half a year before a doctor reassures them that they do not, in fact, have cancer. Subjects in Grail’s study who received a false-positive result spent an average of 162 days in suspense before being cleared.

When I asked Grail about potential harms of the test, including this delay, the spokesperson told me that Galleri offers diagnostic guidance for doctors and patients who test positive through “a suite of services, including direct support from our medical science liaisons.” Grail has also presented data suggesting that the distress of patients who receive false positives tends to go away over time. Some people, however, may never feel completely at ease knowing that cancer-related genetic code is circulating in their veins. The medical system is very good at puncturing patients’ confidence in their own health.

Some anxiety may be worth experiencing for the opportunity to catch an actual cancer before it turns fatal. But that exchange would only work if curable cancers could be consistently picked up in our blood. Galleri is much better at detecting advanced malignancies—which shed more genetic material, and many of which are incurable—than small ones that are worth finding sooner. Galleri is billed as an early-detection test, but just one out of five cancers it finds are identified at Stage 1, which is the earliest stage. At this point, the same is true for other blood-based screening strategies, as well.

The only way to know for sure whether cancer-finding supertests truly save lives is to evaluate them in a large randomized, controlled trial. The U.K.’s National Health Service has enrolled 140,000 participants in such a study of Galleri; the main results, on whether the test can find cancers before they spread, are expected in a year or two. Then researchers will keep track of whether participants have their lives extended in the years that follow. In the meantime, U.S. efforts are running far behind. The National Cancer Institute is planning for a 24,000-person pilot study of multi-cancer screening, but any bigger and more useful randomized trial won’t begin for a long time.

The fact that all of this research is ongoing hasn’t stopped Grail from offering its wares to the public. The company recently sponsored a PGA Champions Tour event in California, where players and fans were offered cancer-screening blood tests on the golf course at a $100 discount; more than 100,000 Galleri tests have been performed in the U.S. since they first became commercially available. Meanwhile, hundreds of advocacy groups are lobbying the government to pay for multi-cancer-screening tests through Medicare. By one estimate, widespread adoption could cost Americans more than $100 billion annually—dwarfing the $7.8 billion spent on mammograms as of 2010, or the $6.6 billion spent on Pap smears.

It’s hard to miss the scientific challenge that still remains. In what might be a bit of corporate retconning, when Barron’s spoke with one of Grail’s co-founders about the story behind the company’s name in 2021, he wasn’t quoted saying that the company thought its blood test could be the holy grail of cancer screening. Rather, he said the name was chosen “out of humility,” because “the Holy Grail was never found.” That humility isn’t in the pitch to patients, though. Most people who use the product today will have no idea that they are generating “real-world evidence” for a technology that may yet be found unable to extend their lives. They’ll assume that if cancer-finding supertests are available in clinics, then we must already know that they’re worth using. We don’t.