Catch cancer early for just $699? That is the pitch from Pathway Genomics, a diagnostics company in San Diego whose new blood test, which it began marketing to consumers today, is supposed to detect DNA mutations harbored by unseen cancer cells in your body.

The test sounds like great idea. Who wouldn’t want to know if they have cancer lurking somewhere? But it’s not quite so simple. The Verge has a detailed explanation of the test and its grave problems.

If true, the [test] would be a big deal for medicine because it could be used to catch cancer before a person shows symptoms. But Pathway Genomics doesn’t have data that shows that the test actually works.

That’s right. They don’t know how accurate it is or, more importantly, if it’s good medicine. Finding hints of cancer could save lives, but it might also send consumers jumping into MRI machines searching for cancers that don’t exist or that won’t ever hurt them. Wonder what kind of medical mayhem could result? Check out this scene from Woody Allen.

Jim Plante, CEO of Pathway Genomics, told the Verge that his company’s test had been validated through tests on “hundreds of samples.” Not too impressive. To find out if the test actually helps or harms people would require thousands of samples, many years, and millions of dollars that Pathway hasn’t spent.

What’s more, Pathway is marketing the test, called CancerIntercept, directly to consumers, having them consult online with a “telemedicine” doctor to order it. I went ahead and filled out Pathway’s online purchase form, and a salesperson called me back minutes later to explain that the company would send a “mobile phlebotomist” to my home to collect blood. 

They also offer a subscription plan, the salesperson said, so that I can get my blood tested four times a year. That’s just $1,196, billed to my credit card.

A promotional video from Pathway Genomics urges consumers to get their blood tested for signs of cancer.

The company is pitching its test to people who have reason to think cancer is likely. That includes people with inherited cancer risks and longtime smokers, or those exposed to radiation. But the subscription plan is one way Pathway could be misleading people. I am not aware of any evidence whatsoever that frequent, repeated testing for cancer mutations has any benefit. Privately owned Pathway has pushed the edge of diagnostic testing before. According to Reuters:

The company has already run into trouble on a similar issue. In 2010, Walgreens pulled Pathway’s genetic testing kits from its pharmacies after U.S. health regulators warned patients might take action without understanding the results.

Even if it’s premature, what’s important about Pathway’s test is that it’s a clear sign of things to come. These “liquid biopsies,” as the blood tests are called, really do have the potential to catch cancer early and possibly revolutionize treatment by letting surgeons cut tumors away before they cause problems.

In fact, careful scientists are now accumulating evidence from thousands of patients at risk for specific cancers to determine if early detection is possible, better than existing tests, and if it actually saves lives. In February, for instance, we wrote about a large study in Hong Kong led by researcher Dennis Lo involving thousands of Chinese at risk for a type of nasal cancer.

The study involves 20,000 healthy middle-aged men recruited in Hong Kong, and it’s halfway done. Among the first 10,000 men screened, the researchers picked up 17 cases of cancer—13 of those at stage I, the earliest kind. Nearly all these men have now beaten the cancer with radiation treatment. “They would normally be just walking on the street not knowing that there was a time bomb waiting to go off, and now we have alarmed them,” says Lo.

The tests are able to detect the presence of cancer because dying cancer cells release fragments of DNA into a person’s blood, although often in small amounts. The first scientific reports demonstrating the phenomenon were back-to-back Nature Medicine papers in 1996 (here and here). Only more recently has DNA sequencing technology become powerful and inexpensive enough to offer the chance of a routine screening test cheap enough for a mass market.

Much of the foundational science and technology development around liquid biopsies has been done at Johns Hopkins University. Last year, we profiled the efforts of famed cancer biologist Bert Vogelstein, who told us he was convinced that early detection, not drugs, is the best way to defeat cancer.

To Vogelstein, the blood tests mean it may be possible to catch more than half of cancers early on, and potentially cure them with surgery. “If there were a drug that cured half of cancer you’d have a ticker-tape parade in New York City,” he says.

The Hopkins scientists demonstrated that early detection is possible, but only sometimes, in some cancers. And they weren’t so bold as to launch an early detection test. When I spoke to Vogelstein’s colleague Luis Diaz about screening tests earlier this year, he warned me that some companies could take a “reckless” approach. “Without the proper science and clinical evaluation it can be dangerous to deploy,” he said.

The problem, Diaz explained, is that there are several ways the test can lead patients and doctors astray. “There are three levels of errors: technical, biological background, and clinical context,” he said. The first problem is false positives. The earlier you are trying to catch a cancer, the more likely a positive result will be a “technical artifact” caused by the test. You might scare people into thinking they have cancer when they don’t.

A different problem is “biological artifacts,” or mutations that accumulate in people as they age but don’t mean anything. The last source of error is medical. What if you find and treat a cancer that will never kill you? That could be the most serious drawback of all. The costs to the medical system and to patients can be extraordinary. As we noted last year:

Medicine has a precedent of handling predictive cancer tests poorly. Consider the PSA test. Millions of men have ended up getting treated for prostate cancers that ultimately wouldn’t have affected them. Studies by researchers at Dartmouth College suggest that mammography also leads to overdiagnosis and overtreatment. About 25 percent of breast cancers discovered, and treated, would not have caused any symptoms.

That’s why Pathway’s test is such a big deal, and possibly such a big problem. Technically, it’s definitely possible to detect cancer DNA from a blood draw. But the hard work of knowing whether it’s good medicine hasn’t been done.