Opinion: Liquid biopsy: misplaced faith in early cancer detection?

The trial of Theranos CEO Elizabeth Holmes for defrauding investors about the company’s blood tests highlights a more general reality: huge sums of money are being devoted to develop and promote diagnostic technologies with little evidence they work as advertised.

In Holmes’s case, the technology was a sham, albeit one that lured in high-profile investors, like Henry Kissinger and Rupert Murdoch, and corporate partners, like Walgreens and Safeway. Companies today are making real testing technologies, but that doesn’t mean more testing will help individuals live longer or live better.

In particular, companies engaged in using a single blood test to detect multiple cancers — so-called liquid biopsies — are attracting a lot of investors. Both the companies and those funding them are banking on the public’s faith in the claim that early cancer detection always saves lives.


PetDx, for one, promises on its website to deliver “breakthrough liquid biopsy technology for the benefit of dogs and their families” with its OncoK9 test. That’s followed by the requisite disturbing statistics: 6 million dogs are diagnosed with cancer every year in the United States and 1 in 3 dogs will develop cancer in their lifetimes.

That’s an interesting choice of statistics, because more cancer testing could only increase those numbers.


PetDx makes the case that, because canine cancers are typically found at a late stage, many dogs could enjoy more healthy years if their cancers were found at earlier stages. Perhaps. Or perhaps many more dogs will be found to have cancer, many more dogs will be subjected to cancer treatment, and many more dog owners will be shuttling back and forth to the vet worried about their pet — and their monstrous vet bills.

Don’t have a dog? Fear not. Multiple companies are mounting arguments to market liquid biopsy tests for human use. These include products like CancerSEEK from Thrive, Galleri from Grail, and PanSeer from Singlera Genomics.

Liquid biopsies work by identifying fragments of cancer DNA circulating in the bloodstream. Everyone has some DNA in their blood, shed from their bodies’ cells. People with cancer may also have some DNA from cancer cells. The challenge is finding the tiny amount of cancer DNA circulating with a much larger amount of DNA from healthy cells.

The FDA has approved liquid biopsies for people with certain types of advanced cancer, who are likely to have a lot of cancer DNA in their blood. These tests may well prove to be a good way to detect the recurrence of cancer and to guide cancer therapy. But the real money isn’t in testing the relatively few people with advanced cancer, it’s in testing everyone else: those not known to have cancer. Testing asymptomatic people for cancer is known as screening.

For the past quarter-century, my research has focused on cancer screening. It’s no secret that I believe the benefits of cancer screening have been systematically overstated and its harms largely ignored. Because aggressive, rapidly growing cancers would be expected to shed cancer DNA into the blood, I have little doubt that liquid biopsy can find some deadly cancers earlier. And because innocuous, non-progressive cancers wouldn’t be expected to shed DNA, I worry less about cancer overdiagnosis with liquid biopsy than with other screening tests.

My issue with liquid biopsies is more fundamental: I want to know if multicancer liquid biopsy screening of the general population will help people live longer.

Effective cancer screening requires more than early detection. It also requires that starting therapy earlier helps people live to older ages than they would if they started treatment later. If that doesn’t happen, liquid biopsies will only lead to people living longer with the knowledge they have a potentially incurable disease without extending their lives. These people would be subjected to cancer therapies and their toxicities earlier, but at a time when they would otherwise be experiencing no cancer-related signs or symptoms.

Ironically, screening will look like it helps — even when it doesn’t. Survival times and five-year survival rates always increase with screening even if the age of death is unchanged. That’s because both are measured from the time of diagnosis. The problem is so ubiquitous it even has a name: lead time bias.

Another reason why multicancer liquid biopsy screening might not help people live longer is the aging soma hypothesis: Many advanced cancers are believed to reflect an aging body — some combination of accumulating cellular damage and diminished immune function.

The risk of dying from cancer is strongly related to chronological age and would almost certainly be even more strongly related to biological age (if we could measure it). But that means individuals at high risk for dying from cancer may also be at high risk for dying from other causes. If that’s the case, screening could reduce the number of deaths due to cancer only to have them replaced by deaths from heart disease, infection, and other causes.

The only way to sort all this out is to do a randomized trial: Randomly assign participants to one of two groups — liquid biopsy screening or no screening — follow both for a set number of years, and compare the death rates in the two groups. To its credit, Grail has committed to doing a randomized trial of Galleri in collaboration with the United Kingdom’s National Health Service. Screening trials like this typically require tens of thousands of people followed for a decade or so.

Or tens of thousands of dogs followed for 10 dog-years.

The fact that so many participants are needed is a tip-off of what’s expected: the trial is designed to detect a very small benefit. For context, the life-saving benefit of dexamethasone for Covid-19 patients on ventilators was demonstrated in a trial of 1,000 patients — half of whom got the steroid and half of whom didn’t — followed for one month. Big benefits can be demonstrated in small trials, while small benefits require big trials.

It’s easy to imagine how companies will get liquid biopsies for screening enmeshed in American medicine: Publish a few studies showing the test finds cancer. Market the findings to patients, physicians, and health systems. Hire physician-researchers at salaries that far exceed those in academic medicine. Have them publish a few more studies showing increased 5-year survival. Invest early profits into campaign contributions.

By that time, there will be demands that liquid biopsies be covered by private health insurance, Medicare, and Medicaid. Screening liquid biopsies could become the standard of care before we ever get a clear answer to the question: Do they really save lives?

The only thing that is clear to me is that this will cost a lot of money. That’s another thing the companies and their investors are banking on. The Galleri test costs $949 a pop and is recommended every year in those 50 and older. With 100 million Americans in this age group, that’s about $100 billion a year — 15 times the budget of the Centers for Disease Control and Prevention. And that’s not even counting the cost of all the subsequent testing and treatment that will invariably follow.

Just as Aduhelm, the newly approved treatment for Alzheimer’s, is the drug that could break American health care, liquid biopsy is the screening test that could do the same thing.

If saving lives is the goal, it would be far better to plow that $100 billion into genuine determinants of human longevity. Don’t raise health insurance premiums and out-of-pocket health care costs, forcing people — particularly the near poor (those not poor enough to be on Medicaid) — to have less to spend on real food, housing, and education. And don’t bankrupt public budgets, diverting funds from food stamps, schools, and safe, walkable public spaces.

Liquid biopsies may prove useful in the relatively few people for whom it is FDA approved. But the financial incentive — both for the test manufacturers and our volume-driven medical care system — is to screen the rest of us.

Don’t get duped into thinking we can biopsy our way to better health.

H. Gilbert Welch is a senior investigator in the Center for Surgery and Public Health at Brigham and Women’s Hospital and is the author of several books, including “Less Medicine, More Health: 7 Assumptions That Drive Too Much Medical Care” (Beacon Press, 2015).

Source: STAT