Why There Aren't Yet Nanobot Doctors

Scientists have long said that tiny robots would soon be able to conduct surgery and deliver drugs deep inside the body. Here’s why they’re still not a reality.

Jaehyun Park, a postdoctoral student at the University of California San Francisco, works at the Marvell Nanofabrication Laboratory in Berkeley. (Noah Berger / Reuters)

For years, the cutting edge of medicine has promised nanobots. Tiny little machines that could run around your body delivering drugs, checking up on arteries, and generally keeping people healthy. But so far, those machines haven’t quite come to dominate the way some people thought they might. The human body is vastly more complicated than any robot we’ve ever made. So creating a miniscule robot to go inside of it, to work with that vast infrastructure, and to do our bidding, is a huge challenge.

The first thing to know about nanorobots in medicine is that they’re not like the robots you’re probably imagining. “A robot means a machine that can handle things automatically,” says Aniket Margarkar, a researcher at the Centre for Drug Research at the University of Helsinki in Finland. Scientists who build nanorobots are building tiny packages that can complete tasks in an automated way. “Sensing, responding, detecting friend of foe, delivering a payload, delivering cargo, these are tasks that we build robots to do on the macro scale,” adds Shawn Douglas, a researcher at the University of California in San Francisco.

Douglas admits that calling this technology nanorobots is a bit of a PR move. He talks about his own work on a nanorobot paper. “We’re submitting this to Science and it’s going to get a lot more attention if we call this something that will resonate with people,” but that doesn’t really bother him. “I don’t begrudge anybody for overhyping their work,” he says.

Even if they’re not robots in the same sense that your or I might think of, the tiny molecules that can automate their tasks are quite useful. Both Douglas and Margarkar are working on nanorobots that can sense where they are, and deliver a drug at the right location. So if a bot is carrying a drug that is meant to kill liver-cancer cells, it would also be outfitted with a sensor to tell when it encounters those cells.

This targeted drug delivery has all sorts of upsides: You’re not dousing the body in poison hoping it reaches the thing the poison is meant to kill, for example. But it’s not easy to design. In fact, the field of nanorobotics has encountered a whole lot of challenges since its invention, ones that hold us back from the promise of those predictors and science-fiction writers.

One challenge when it comes to synthesizing nanorobots is that it takes a lot of work to actually make enough material to test your ideas on. “It’s a matter of all the practical annoying steps of actually doing it,” Douglas says. “The big thing that half of my lab is stuck on is making enough material that we can test. We could do a very small amount of material, we synthesize the DNA and we run a few experiments and show in a petri dish and say ‘Hey this looks promising. We made millions of little things and some of them do what we expect.’” But the next step in that process means picking those promising things up, and testing them on a lot more material.  “We need 100,000 times that much material, which would bankrupt the entire lab. So we have to invent a new way to make it, so that takes a couple of years. And then before you know it five years have gone by.”

Then, once you make enough material, and have a promising candidate, Margarkar says that sometimes there’s not a great system in place to test your idea. Pharmaceutical studies are designed to look for results in the end organ or illness. Take a drug for liver cancer: If the drug reduces the liver cancer, it is successful. But when it comes to nanoparticle delivery, researchers also want to be able to figure out if the process, the actual movement from injection through the body to the liver cells, works the way they designed it to. If it fails, did it fail to detect the cells? Or did it fail to release the drug when it did detect the cells? Or did the drug itself just not work to kill the cancer? The answers to these questions are key, and Margarkar says that for a long time there wasn’t a good way to figure them out. Now, researchers have started using computer models and scans to predict and then track the nanorobots on their whole trip.

Then, once you have a delivery system that works, and a drug that works, researchers have to tackle the FDA. Getting nanoparticle robots approved isn’t necessarily the issue, but getting the combination of drug and robot approved can be. Margarkar says that sometimes, even if both the nanoparticle and the drug have both been proven safe individually, the FDA may require the combination to be run through the entire approval process again. “So you’re testing it again in the lab, then on animals, then on people. From the moment you enter that process it can take eight to 10 years to finish,” he says.

All this is to say: It’s not easy to build and implement nanorobots. Margarkar makes that point another way: The first drug encapsulated in a nanoparticle was approved by the FDA in 1995. In the 30 years since, there have only been about 30 nanoparticle drugs put on the market. In that same time, the FDA approved about 700 new drugs in total.

In fact, the progress on these methods is so slow that sometimes Douglas gets discouraged with his own work. “This sounds crazy, but people are making so much progress on cancer therapies that I’m almost not as excited now to work on our projects, because it’s clear that people are already curing certain types of cancer with other technologies. People are already on the right track with immunotherapy. We’re 20 years behind the people who’ve been working on the stuff that’s just going to work.”

And yet, Douglas says, they know that it has to work. “We’re using the same materials that cells use to really great effect, we’re repurposing those for medicine. We’re in the stage where it’s totally obvious that it has to work, it already works in nature, it would be surprising if it didn’t work in the end.”

Margarkar is more optimistic about the enterprise. He imagines diagnostic nanorobots, ones that will float around in the body, fitted with sensors to detect possibly dangerous changes. If an artery starts to get a little clogged, or if a cancer node starts to grow, these nano-monitors would detect those changes, and alter their chemistry in response. An MRI scan could reveal the markers that have changed, and flag health problems early. But before any of that can happen, researchers have to overcome all those other hurdles.