“We’re Hoping to Build the Tricorder”

Google X’s Andrew Conrad is developing an early disease detection system that could reinvent health care

Photo by Stephen Lam.
Sometimes moon shots come in small packages — small enough to float in the bloodstream and send out alerts when danger is afoot. That was Andrew Conrad’s vision when he came to Google X in March 2013 after a storied medical research career. Now he is revealing details of that vision — and reporting that experiments are well under way in realizing it.

Basically, Google X is creating a system for early detection of disease that involves ingesting specially “painted” nanoparticles that target various molecular harbingers of disorder. If the nanoparticles find these microscopic malefactors, they send out signals that will be picked up by wristbands. The early alerts means that potentially deadly ailments might be apprehended soon enough to be dispatched by minimal treatments. Conrad thinks that everybody, healthy or not, will be popping those pills and wearing those devices (or something like it) in the not-distant future.

Conrad’s track record demands that we take this seriously. In the 1990s the molecular biologist’s work led to a test that dramatically reduced the time and cost it took to test blood and plasma donations for HIV and other viruses. He continued his ground-breaking research as chief scientist at LabCorp, but after spending time with Sergey Brin and others at Google, he decided that he could reach for most audacious goals at that company’s well-funded advanced lab, which has been the source of projects like self-driving cars, Google Glass, and Project Loon.

Google X Life Sciences, which Conrad heads, now consists of over 100 researchers including some of the leading chemists, scientists and biologists as well as experts in machine learning and data mining. Earlier, Google announced that the group is conducting what it calls a Baseline Study of 175 healthy people (later scaling to a much larger number), using genetics and molecular analysis to really understand what “healthy” means — the better to much earlier detect variations that portend danger. The reach of the program is extended considerably with close partnerships with labs at Stanford, MIT and others. Another project puts tiny sensors and microelectronics into a contact lens, so it can measure glucose from tear ducts to monitor diabetes.

Conrad, 50, is a surfing aficionado who sports a blonde goatee and a wry sense of humor. He gave BACKCHANNEL a rare interview about his new project as well as an overview of what his team is doing at Google X.

[Steven Levy] You are announcing a wildly ambitious project involving nanoparticles, wearables, and big data. Is this the kind of thing that led you to Google X?

[Andrew Conrad] It’s pretty much what they expected me to do. Healthcare is a huge problem, and it’s Google X’s mission to take on huge problems Sergey [Brin] said come to Google X, try something crazy, do something that’s a 10x change.

You’ve had a storied career. I imagine you’d have a zillion options.

I was not unemployed or worried about things at the time, I was having a pretty good time. But this was the chance to totally reinvent something I had seen and felt from the inside, and that was ready for some change. We knew it we would take partners, but I thought Google X was a big enough soapbox that we could try do it. It was this notion of trying to do something ten times different. That was both hugely intriguing and hugely challenging—how do you make a ten times difference? That’s sort of a dream aspiration.

It’s also ten times the chance you’ll fail, right?

Yeah, but at this point failure is not terribly scary because I would just fail back to being a happy, retired guy. I would be living in our Hawaii house and surfing more. So the risk didn’t scare me as much as the opportunity to really make a huge difference.

So tell me how this project began.

For the last two thousand years, healthcare has been this transactional, reactive system. You go to the doctor when you’re sick and the doctor prods you, taps you, and gives you a prescription or does some procedure, then sends you on your way. But with serious disease, we frequently encounter the physician when we already are very sick. In fact, a majority of cancers are diagnosed in the later stages because they’ve become clinically apparent. Some cancers have ninety percent success rate if you diagnose them in early stage one. But most cancers have a five or ten percent survival rate if you diagnose them in stage four. We’re diagnosing cancer at the wrong time. It’s analogous to only changing the oil on your car when it breaks down. If you think of airplanes or cars or any complex entity, preventative maintenance has been proven without a doubt to be the better model. Yet for some reason we don’t concentrate on that in Western medicine. So our central thesis was that there’s clearly something amiss. So we needed to see if, with partners, we could change the system in healthcare from being reactive to proactive.

How do you start on that?

The first thing you realize is the triggers of diseases usually start way before they’re clinically apparent. They are usually subtle and rare. Most of the time people are not sick. That means monitoring would have to be done continuously. You have to measure all the time because if you only measure once a year when people visit he doctor—or in men’s cases, once a decade—you’re going to miss huge swaths of the possibility of detecting disease early. So we have to make a continuous monitoring and measuring device. Since it’s continuous, it has to be something people wear, right? Can you imagine if you had to carry a sixty-pound thing around with a radar dish on your head and poke yourself with needles every hour? People just wouldn’t do it.

So the radical solution was to move away from the episodic, “Wait ‘til you feel a big lump in your chest before you go into the doctor” approach, and do a continuous measurement of key biological markers through non-invasive devices. And we would do that by miniaturizing electronics. We can make a little computer chip which has three hundred and sixty thousand transistors on it, yet it’s the size of a piece of glitter. One of the other ways is to functionalize nanoparticles. Nanoparticles are the smallest engineered particles, the smallest engineered machines or things that you can make. Nature does its business on the molecular level or the cellular level. But for two thousand years we’ve looked at medicine at the organ or the organism level. That’s not the right way to do it. Imagine that you’re trying to describe the Parisian culture by flying over Paris in an airplane. You can describe the way the city looks and there’s a big tower and a river down the middle. But it’s really, really hard to opine or understand the culture from doing that. Its the same thing when we look at systems—you can see that there’s a complex system, but unless you’re down at the level where the transactions occur, it’s very hard for you to imagine how it works.

That’s when we began to realize that this idea of nanotechnology, miniature electronics, and continuous measurements of biological parameters were all possible. So instead going to the doctor who says, “Let me draw blood and in three days I’ll call you if there’s anything wrong,” the doctor can can look and say, “Oh, I just checked all your blood over this last year, and it looks like your kidneys are good, your liver is good, I don’t see any indication of oncologic cells, pretty good, thanks.” We use Star Trek as our guiding force around Google because there used to be a computer called Tricorder —you’d talk to it and it would answer any question. That’s what we’re really looking for at Google X. We want to have a Tricorder where Dr. McCoy will wave this thing and say “Oh, you’re suffering from Valerian death fever.” And he’d then give some shot in a person’s neck and they’d immediately get better. We won’t do the shots—our partners will do the shots. But we’re hoping to build the Tricorder.

Can you describe your system?

So this nanoparticle platform we’re talking about essentially does the following: You take a capsule chock full of the nanoparticles, and they absorb into your body and into your bloodstream. These nanoparticles are two thousand times smaller than a single red blood cell. They’re tiny. They’re so little that they can pass through parts of your body, they go through the blood, they go through your lymph system, they just walk around. They’re essentially very benign particles—-there’s already lots of FDA approved nanoparticles for imaging and stuff like that, because they’re simply made out of an iron oxide core, like you take in a One-A-Day Plus Iron pill. And they’re decorated with proteins and amino acids and DNA to make them bind to certain things.

Detecting Disease early with nanoparticles

So there’s no issues of unexpected consequences of having these in your body?

No. Actual particles like these have been tested for quite a while to be safe. The trick is decorate them with smart molecules on their surface so they do smart things.

You say you decorate the nanoparticles. I’ve also heard the word painting. I’m having trouble wrapping my mind around how you paint a nanoparticle.

It’s done with chemistry. The core of the nanoparticle is iron oxide. So you take all the little particles, you can’t see individual ones, but you take a spoonful of particles, and you throw it into a mix of almost a polymer, like paint, that coats the outside. And coating the outside of it makes it permissible to attach other things to the surface.

So you have a generic nanoparticle and paint different things on its surface to target different kinds of markers for disease.

Yes. You can use these nanoparticles to detect rare things like a cancer cell or you can use them to measure common molecules. For example, in one case we put a coating on the nanoparticle that finds sodium — it’s a super common molecule but very important in renal disease. When a sodium molecule comes into the nanoparticle, it causes the nanoparticle to fluoresce light at a different color. So by collecting those nanoparticles at your wrist, where you have a device that detects these changes, we can see what color they’re glowing, and that way you can tell the concentration of sodium. In another case, by having a magnet at your wrist you can tell whether the nanoparticles are bound to cancer cells. This allows us to let these messengers walk around Paris, bring them all back to a central location, and ask them what they saw, what they did, what they encountered. And imagine that is the way in which we’re trying to understand the culture of the French.

There’s not going to be a traffic jam at my wrist as you call all these particles home?

No. Two thousand of them are the size of one red blood cell. You have millions and millions of red blood cells running through your wrist at any time. So if we able to get all the nanoparticles that you take in that pill to collect in your wrist, maybe it would have some tiny effect. Also, there are super para-magnetic nanoparticles, they’re iron oxide. When you take the magnet away, they don’t retain their magnetism, they just disperse back into the wind. You bring them to your wrist for, let’s say, only an hour a day. Or you can do it for a minute every hour. It depends on the algorithm that you want to use.

What about false positives?

This goes back to our Baseline Study. We are looking at thousands of normal, healthy people and we’re going to measure everything we can think to measure in an effort to answer questions like how many cancer cells should a normal healthy person have, Zero? I don’t know. One? I don’t know. Ten? I don’t know. Because we might have cancer going around all the time but the immune system stifles it. So if you really want to be proactive, you need a ground truth. And the baseline is enrolling thousands of super healthy people, measuring all these things on them, then putting these devices on them to make sure we know what do when we’re looking for someone who’s transitioning from health to disease.

How much have you done?

We’ve done a lot, to be quite humble about it. Enough to give us great confidence that this is all likely to work. At our Google facilities, we’ve been able to build the nanoparticles, decorate them, prove that they bind to the things that we want them to bind to, in really clever artificial systems. We’ve made these molded arms where we pump fake blood through them and then try devices to detect the nanoparticles. We’re pretty good at concentrating and detecting nanoparticles. We’re pretty good at making sure that those particles bind only to cancer cells and not to other cells. You know that sodium experiment I just spoke about? That’s the actual data that I was talking about. We’ve built particles that can measure small molecules.

So you’re saying that of the four components of this system — delivery, targeting, detecting, and counting— you’ve sort of got a proof of concept of each?

Yes, we have really compelling experiments that demonstrate each of the four sub-components of this.

And how about working together?

We have many of them working together.

In the last couple days I’ve spoken to a couple people who are working with you— Dr. Sam Gambhir of Stanford, and Dr. Robert Langer at MIT. Both of them are excited about the project but emphasized remaining challenges in delivery, detection, and other areas. They’re bullish in the long run but haven’t expressed the certainty that you’re displaying now.

I don’t want to say I have any certainty whatsoever. There’s a big journey in between the in vitro [lab testing] demonstration principle and in vivo [testing in living organisms]. Both Gambhir and Langer are working with us on the in vivo part of it. We’ve done as much as we can do in vitro. We know that much of this works: we’ve become very good at nanoparticle decorating, we’ve become very good at concentrating them and understanding how they behave in magnetic fields. There’s still a million crazy things that happen with people, and there’s a long journey to put medicines into people, and it has to be done in the open because we’re going to do experiments— people will be wearing these devices at our Baseline Study. But I think we have years to go, not decades.

You’re not going test on animals first?

You don’t have to in this case, the medicines are well known. I think we have pretty demonstrable evidence that this concept is plausible, maybe even probable. But we still have this long journey and that’s why we’re partnering with MIT and Stanford. We’re having hospitals and doctors and folks beginning to think about how this works. And eventually we’ll find big partners who will take this to the next step.

Have you patented this stuff?

Yeah. That’s part of the reason we had to talk about this. There’s a pretty substantial body of patents that describe what we’re talking about in great detail that will be publicly noticeable in the next month or so.

That seems to indicate how you might monetize these advances, something that even Google X keeps in mind for its long-range projects. You would license it?

Yeah, and that’s what we did with our contact lens. We licensed to Novartis. It was extremely beneficial to both parties. They have expertise at doing exactly this, taking contact lenses, taking medicines and taking diagnostics and bringing them forward to the world.

Do you anticipate any pushback to this procedure? I can imagine someone saying, “I’m not going to let Google put stuff in my blood.”

Remember, most rational people let pharma companies put things in their blood all the time. What do you think happens when you swallow any pill?

But we do it reactively. I’ve got a headache, I take a pill. I have cancer, I get chemo. It’s different to ask a healthy person to take a pill and say, “This is going to watch you.”

That’s a good point. Imagine the first group of people who would wear this…

Yeah, Sergey and Peter Thiel.

That’s funny. But the first people using this might have had breast cancer, and they’re worried about reoccurrence. Half of those women reoccur within five years. Wouldn’t you want to know if that was going to happen earlier? So instead of eight rounds of chemo you only get one round of chemo? Boy, I bet you lots of women with breast cancer would be glad to swallow that pill. Now imagine the papers that would come out reporting that when women use these devices, we were able to detect their reoccurrence eight months sooner and therefore their treatments were 47% more effective than the women who were diagnosed conventionally. I’m pretty sure most people would do it then. Next would be women at high risk for breast cancer, they’d hear, “Even if you haven’t had it yet, you should do this.” This is why we take partners, because there are lots of very good healthcare companies that would promote this notion. We’re going to be inventors that work on the technology— disruptive, innovative technology—and then we’re going to look for partners who will bring it forward .

Besides targeting and detecting known diseases, would this system lead to other benefits?

If we and our partners could bind nanoparticles to all the tumors maybe we could do something to get rid of them, too, right?

Is that on the road map?

Yeah. The mission of Google X Life Sciences is to change healthcare from reactive to proactive. Ultimately it’s to prevent disease and extend the average lifespan through the prevention of disease, make people live longer, healthier lives.

It sounds like that mission overlaps a little with another Google health enterprise, Calico. Are you working with them?

Let me give you the subtle difference. Calico’s mission is to improve the maximum lifespan, to make people live longer through developing new ways to prevent aging. Our mission is to make most people live longer, getting rid of the diseases that kill you earlier.

Basically you’re helping me live long enough for Calico’s stuff to kick in.

Exactly. We’re helping you live long enough so Calico can make you live longer. And I think what’s beautiful about Google is when Google attacks a problem, like healthcare, they really put some force behind it in a magical way.

Is all this data you collect going to be aggregated where analysis can be used to come up with new insights?

Of course. Imagine that every patient at Stanford University gets to use this device. The ability to understand the new description of the way people are, the molecular profiles of patients, would allow those doctors to make decisions that would be completely different than under other contexts. Physicians will now be empowered to say, “If you have a seventeen percent increase in such and such, does that have material, clinical effect? Let me see everyone else who had seventeen percent last year. Oh yeah, look, nobody had any clinical sequelae, that’s probably just a noise in the system.” It would be amazing to me if questions like that could be answered.

You’ve been at Google almost two years now. Have you found this to be a dramatically different kind of place to do work like this than in an institution or medical facility?

I will have been here two years this March. In nineteen months we have been able to hire more than hundred scientists to work on this. We’ve been able to build customized labs and get the equipment to make nanoparticles and decorate them and functionalize them. We’ve been able to strike up collaborations with MIT and Stanford and Duke. We’ve been able to initiate protocols and partnerships with companies like Novartis. We’ve been able to initiate trials like the baseline trial. This would be good decade somewhere else.

And you don’t have to have your staff spend endless time on funding requests.

No, you don’t. We’re super-fiscally responsible, though—we’re probably as cautious and thoughtful about spending money as anyone. But we’re unafraid. As long as you’re trying nobly, failure is not frowned upon. People frown upon a stupid attempt, but when it’s a brave and wise one, failure is actually a blessing because we often learn more from failures than success. If told you ten years ago I was going to make a computer that could do complex calculations, have a built-in radio, many different sensors, and put it in the size of a piece of glitter, you’d laugh at me, right? And even if I were to tell you I could do that in the next ten years, you would’ve laughed at me.

So in ten years am I going to have this thing on my wrist?

Oh yeah.

Really?

One would hope so. I would hope so.

How about five years?

I think we’ll find partners who will begin to do that in the next few years, yeah. Yeah, I think we’re a few years away. This is a giant program and the mission is a noble one — to prevent disease instead of just trying to find ways to treat it. It’s like we want to build your house out of fireproof materials instead of providing lots of fire extinguishers. By miniaturizing electronics, by creating an understanding of how to use nanoparticles, by having a ground truth that baseline provides, the opportunity for us to create many, many, many more devices, many, many more new innovations in healthcare become plausible.

There’s one more thing I want to say— this nanoparticle detection is a project, the contact lens is a project, the Baseline Study is a project. But those projects are manifestations of a program. Each of these things are fascinating and seem pretty science fiction-y, but the message is that we’re actually pretty methodical. We’re instituting a program that includes really, really powerful partnerships with universities, with healthcare providers, with pharma companies. And by having a philosophy that says partner wisely, we’re punching way above our weight, and we may have a chance to turn this battleship of healthcare around. Because each time we pick beautiful partners like Stanford and Harvard and Duke and Novartis we’re using part of the system in a sort jiu-jitsu-esque way, where we take the inertia of the system and sort of flip it around. Those are serious players. We know that we’re just upstarts but we’re trying to dream big and trying to work hard to have an effect on the overall system.

Follow Backchannel: Twitter | Facebook