Google promised a 'tricorder' that would instantly diagnose cancer - but things haven't gone as planned

Science fiction is great for inspiring scientists to tackle big problems with innovative, out-of-the-box solutions.

Take one of Google's signature life sciences projects: a device that offered the promise of diagnosing cancer long before someone showed any symptoms. It's inspired by Star Trek's medical tricorder, a diagnostic super-computer.

But turning science fiction into reality, it turns out, is a tricky nut to crack.

In fact, the tricorder is giving Verily, Google's life sciences arm, so much trouble that, according to an investigation by STAT, it's now being "seen internally more as a way to generate buzz than as a viable project."

Here's the idea they've been chasing: One problem with cancer is that we often don't catch it soon enough, and screening tests as they are right now don't necessarily solve that problem.

There are a few initiatives trying to figure out how to decode cancer diagnostics. But so far, they're focusing on testing blood samples. Verily's tricorder project is trying to go a step further - to be looking out for cancer cells all the time from within your body.

Say you decided to use the tricorder. You would start by swallowing a pill - but instead of medicine, it would contain tiny magnetic nanoparticles primed to look for cancer cells. When they found one, they would light up and alert your Fitbit-esque wristband.

If that all sounds perhaps a little incredible, that's because it is.

STAT highlighted a few problems with this idea. First, researchers would need to figure out how to keep the nanoparticles from being absorbed by your liver or even the very cancer cells they're meant to flag. Second, they'd need to make sure the particles don't disrupt blood flow. Third, they'd need to find a balance between missing cancers they should have caught (false negatives) and seeing cancers that aren't really there (false positives).

And that's by no means a comprehensive list of the potential road blocks.

When contacted by STAT, Verily noted that its projects are "inherently difficult," saying that while they think they can succeed, "as with all true innovation, some projects can and will fail."

It isn't clear how much progress Verily has made so far on the tricorder, since they haven't published any papers about the process. But when the program was announced, Verily CEO Andrew Conrad said they'd have a prototype in six months. It's been three years.

You can read Charles Piller's whole deep dive at STAT.