Invent the Medical Tricorder, Win $10,000,000

GeneralSecretary writes "If you've ever watched Star Trek and said, 'Hey, I could build that,' now's your chance. Qualcomm and the X PRIZE Foundation have teamed together to offer ten million US dollars to whomever can invent 'a mobile solution that can diagnose patients better than or equal to a panel of board certified physicians.' They call it the Tricorder X PRIZE. Hopefully the Tricorder will join the cell phone, MRI, and tablet computer in the list of Star Trek devices that are now part of our lives."

Comments

better than a group of doctors?!?!

[modmans2ndcoming]

Can't we start out with just one doctor?

Re:better than a group of doctors?!?!

[MoldySpore]

There is already a device that can diagnose you as well as 1 doctor. It's called a Magic 8-Ball.

Re:better than a group of doctors?!?!

[sortius_nod]

This has already been done by an Australian. It's a micro-lab basically. It was on a TV show called The New Inventors where inventions are showcased.

Maybe if they learnt to use a decent search engine they'd have found this:

http://www.lifescientist.com.au/article/309857/handheld_lab_receives_innovic_gong/

and this:

http://www.abc.net.au/tv/newinventors/txt/s2669552.htm

Looks like an Aussie gets the $10m.

Re:better than a group of doctors?!?!

[Anonymous Coward]

Well doc, give it to me straight, what have I got.

Signs point to yes

What?

yes

So i've got yes?

My sources say no

Is that a good thing?

no

Re:better than a group of doctors?!?!

[skylerweaver]

Actually, I find it interesting you used the Magic-8-Ball. While it is just a toy, the 20Q toy (which is somewhat similar in my mind) is very interesting because it tries to guess what you are thinking of by asking you yes/no/sometimes/don't-know questions. The neural net was then built by people playing the game and providing 'better questions' for when the AI got the answer wrong.

Could you not do the same for medical diagnosis?

[Do you have a headache?]
"No."
[Does your stomach hurt?]
"Yes."
[You have an ulcer?]
"No."
[What's wrong, and what would have been a better question?]
"Food poisoning, and 'Did you eat uncooked meat recently?'"
[Noted. Now I am smarter.]

It seems that you could make a diagnosis engine that as you rule things out it could come to as good a conclusion as a typical doctor.

What a deal

[suso]

So for only $10 millions dollars you can buy a device that is worth billions. Yeah right.

According to this thing...

[davidbrit2]

"...you talk like a fag, and your shit's all retarded!"

Re:According to this thing...

[CharlyFoxtrot]

"Okay. This one goes in your mouth. This one's for your ear. And... This one goes in your butt."
[scrambles plugs]
"No wait, THIS one goes in your mouth, ..."

Done

[TheSync]

Have an iPad point to http://easydiagnosis.com/

Medical expert systems already diagnose better than human doctors. What they can't do is figure out the best way to bill your insurance. That requires real intelligence.

Re:Done

[alcourt]

Just for fun, I tried out that tool. I plugged in the symptoms I had when I cracked a rib. When I went to the doctor, the nurse diagnosed me before I even got to the exam room based just on watching me walk and hearing where the pain was. The software didn't even begin to ask the right questions, and assigned a 94% probability of something completely unrelated.

I also tried plugging in the symptoms I had a number of years back before I realized I had asthma. It's diagnosis was for several possibilities, none better than 24%.
However, a trained doctor, hearing me cough just once immediately recognized it as an asthma specific cough pattern.

So no, I wouldn't trust that tool over common sense, not even close.

Difficult

[lymond01]

You'd need...

Broken bones: something that bounces off bone and can detect the time to travel which will determine fractures and breaks. If you're using a flat scanning device, everything needs to bounce off something inside the body, rather than pass through and imprint itself on x-ray paper, etc.

Diseases: Lasers can tell blood type now (I think)...might be you could fine tune it to detect anything from genes to bacteria.

Muscle and ligament tears: same deal as bone I suppose -- would need to reflect off of a certain type of material.

Internal bleeding: scan for pools of blood versus the normal trails of blood (veins, arteries, capillaries)

My only question is why we need 4 different devices (MRI, pad, phone, tricorder)...I'd fully expect this to have solar-rechargeable batteries and a form factor that can fit in my back pocket (which would require a wide-angle "lens" for the probes so it doesn't take you 20 minutes to scan someone). And I darn well better hear the "wee-ooo, wee-ooo" sound without having to put on headphones!

Re:Difficult

[Anonymous Coward]

No, the trick is understanding that different imaging modalities give different, and incomplete, pictures. The current level of technology in imaging currently precludes any sort of device like this. Decent MRI scanners still fill an entire room, ultrasound scanners with good resolution still take a trolley that has to be wheeled, not carried. X-rays don't bounce, they penetrate or are absorbed, so would need a two-part system (and a lead apron for the user). No one of these modalities will cover every possible diagnosis (for instance, ultrasound is good for abdomens, but useless for brains except in babies). I'm not saying that a 'medical tricorder' will never be possible, but at the moment there is no way to miniaturise and combine all those different technologies into one device. Something like the GE Vscan is amazing, but that sort of device would be for crude imaging, like am I going to hit the artery with this introducer, is there fluid around the heart, is there a fetal heartbeat present. You'd be crazy to use a handheld scanner for detailed imaging and diagnosis.

Re:Difficult

[ColdWetDog]

No, the trick is understanding that different imaging modalities give different, and incomplete, pictures. The current level of technology in imaging currently precludes any sort of device like this. Decent MRI scanners still fill an entire room, ultrasound scanners with good resolution still take a trolley that has to be wheeled, not carried. X-rays don't bounce, they penetrate or are absorbed, so would need a two-part system (and a lead apron for the user). No one of these modalities will cover every possible diagnosis (for instance, ultrasound is good for abdomens, but useless for brains except in babies). I'm not saying that a 'medical tricorder' will never be possible, but at the moment there is no way to miniaturise and combine all those different technologies into one device. Something like the GE Vscan is amazing, but that sort of device would be for crude imaging, like am I going to hit the artery with this introducer, is there fluid around the heart, is there a fetal heartbeat present. You'd be crazy to use a handheld scanner for detailed imaging and diagnosis.

That's correct - with the OTHER caveat is that interpreting the data is hard. Which is why we make radiologists sit in darkened rooms all day and night and it takes about 10 years to grow one.

There are limited areas where computer analysis has helped humans interpret radiological data (ie mammograms) but they aren't all good and typically just help out with the part of the analysis that humans are terrible at (looking at large quantities of boring data) rather than figuring out things on their own.

Re:Difficult

[Darinbob]

At a previous company we worked on hand held ultrasound before GE (and the thing in their commercials looks suspiciously like a prototype/patents we had so I may send off some emails to see what's up with that). Went away from that design since it was a bit bulky and inconvenient since there were a lot of electronics added to the transducer. Instead a portable hand carry device instead that has enough room to get top of the line image quality, but it can also be docked to a full size cart. People really want the hand held stuff for emergency rooms or disasters but not for day to day use. I think the biggest selling points were image quality and low cost and the portability was a "just in case we need it" afterthought.

But then this is just ultrasound. That doesn't give you a full range of stuff you need to know. It isn't good at detecting cracked bones, it won't handle cranial problems, some types of tissues it won't distinguish very well, it requires good training to use it well, etc. You'll need more than one type of modality. You can shrink down ultrasound but you won't have such luck with MRI. Then when you're done you still need to be good at diagnostics and you're going to need a human for that; ie is that lump part of a spleen or is it a tumor, is that a lesion or a shadow, is the liver missing or am I just holding it wrong? Computer imaging just isn't that good yet, and when the imaging does get that good then you need the AI part to determine what the image means.

The "everything all in one device" is pure fiction. You're going to need multiple devices to gather the data, and then you upload it all to a big computer to analyze. The problem is that all this stuff exists in a hospital but what you need these smaller devices for is for when you're not at a hospital and you may not even have network connectivity.

Re:Simplify

[Krater76]

Figure out how to create a device that can detect common STDs and determine if the person constantly sneezing, has allergies or ebola, and you will be immensely rich. I have no idea if sneezing is a symptom of Ebola, I was trying to make a point.

I think the bleeding from all of your orifices might rule out allergies.

Maybe trying to solve it the wrong way?

[DavidR1991]

When I think of this kind of thing, I get the impression we're trying to solve the wrong problem. Would it make more sense to develop chips and systems that could be embedded _inside_ people? That way they could continuously monitor the person (somehow) and a 'tricorder' would simply extract data out of the systems inside the person

Old Joke

[Anne_Nonymous]

One day Bill complained to his friend that his elbow really hurt. His friend suggested that he go to a computer at the drug store that can diagnose anything quicker and cheaper than a doctor.

''Simply put in a sample of your urine and the computer will diagnose your problem and tell you what you can do about it. It only costs $10." Bill figured he had nothing to lose, so he filled a jar with a urine sample and went to the drug store. Finding the computer, he poured in the sample and deposited the $10. The computer started making some noise and various lights started flashing. After a brief pause out popped a small slip of paper on which was printed: "You have tennis elbow. Soak your arm in warm water. Avoid heavy lifting. It will be better in two weeks."

Later that evening while thinking how amazing this new technology was and how it would change medical science forever, he began to wonder if this machine could be fooled. He mixed together some tap water, a stool sample from his dog and urine samples from his wife and daughter. To top it off, he masturbated into the concoction. He went back to the drug store, located the machine, poured in the sample and deposited the $10. The computer again made the usual noise and printed out the following message:

"Your tap water is too hard. Get a water softener. Your dog has worms. Get him vitamins. Your daughter is using cocaine. Put her in a rehabilitation clinic. Your wife is pregnant with twin girls. They aren't yours. Get a lawyer. And if you don't stop jerking off, your tennis elbow will never get better."

* Cribbed from some dumb site

Well, there may be a way.

[jd]

Every molecule has a unique absorption frequency. So long as you can identify what absorbption bands are present - very very accurately - you're 99% of the way there. The other 1% requires you to create a catalog of such frequencies by scanning pure samples of pathogens.

A second approach would require nanotech and would be extremely slow. Basically, the idea would be to build a device that mimics the cell's mechanism for reading DNA strands and to maintain some sort of internal state that acted in the manner of a cryptographic hash. Once it has calculated the hash, you'd need some way of reading the value. Not sure how you'd do that part, or how you'd even retrieve the device. But again you're getting a value and hunting through a dictionary to see if it is present. If it is, the pathogen is there. If it isn't, it isn't.

Re:Well, there may be a way.

[ColdWetDog]

Every molecule has a unique absorption frequency. So long as you can identify what absorbption bands are present - very very accurately - you're 99% of the way there. The other 1% requires you to create a catalog of such frequencies by scanning pure samples of pathogens.

Huh? What kind of "absorption frequency" are we talking about here? IR spectroscopy (doesn't work well in complex samples), NMR (ditto), UV (ditto). Taking a human body and disassembling it into component molecules would give you 1) and enormous amount of static data that wouldn't help you much in a living organism 2) a pissed off sample, er, patient and 3) a big mess.

Problems in biology are rarely as simple as the presence or absence of something. It's how things interact.

A second approach would require nanotech and would be extremely slow. Basically, the idea would be to build a device that mimics the cell's mechanism for reading DNA strands and to maintain some sort of internal state that acted in the manner of a cryptographic hash. Once it has calculated the hash, you'd need some way of reading the value. Not sure how you'd do that part, or how you'd even retrieve the device. But again you're getting a value and hunting through a dictionary to see if it is present. If it is, the pathogen is there. If it isn't, it isn't.

Again, you seem to think that medicine is all about 'pathogens'. It's not. Lots of times the nasty little pathogens live happily ever after in your body. Occasionally they become pathogenic. Having a particular molecular ID isn't very helpful. You're also missing epigenetic phenomena (Covalent attachment to DNA from non genomic sources), the poorly understood but likely very important small RNA molecule interactions and a host of other things. Again the complex interactions between the organism, it's internal machinery, the external environment and time make reductionism in biology rather difficult. We really don't understand anything more complex than the tobacco mosaic virus and sub molecular detail.

Re:Well, there may be a way.

[pz]

Every molecule has a unique absorption frequency. So long as you can identify what absorbption bands are present - very very accurately - you're 99% of the way there.

I worked on a system like this for determining the concentrations of various pollutants in an exhaust stream from industrial smoke stacks. Even though we work only looking for a very small handful of gasses, it was a very hard problem and our solution never really worked very well.

It suffered because of the problem that while every molecule has a unique spectrum of absorption frequencies (saying it is only one frequency is highly over-simplified), unless you have molecules in near isolation (ie, in highly rareified gasses) and the molecules are nice and small, the absorption spectra are quite broad. Now throw in the fact that in a living creature, you have gazillions of different molecules, many of which will be unknown, and the problem becomes intractable because it's underconstrained.

perspective

[Hazel Bergeron]

1. Star Trek had radio telephones, not "cell phones" - they've been around since before WW2: dumb tool leaving the thinking to humans;

2. The Styalator tablet input device was produced in 1957: dumb tool leaving the thinking to humans;

3. The MRI was fairly recent, but PET (and Star Trek didn't distinguish) applied to medical imaging was discussed by Sweet and Brownell in 1953: dumb tool leaving the thinking to humans;

4. The tricorder could be considered a combination of imaging, sensors and an expert system: attempt to replace human judgement with AI.

Unsurprisingly, one of these things is glaringly missing from everyday modern life.

Re:perspective

[Hazel Bergeron]

Re cellphones: they are so called because they use a network of cells, which is precisely what you don't get when you're visiting other planets. Instead you get the single base station - the ship - and perhaps peer-to-peer, IOW classical radiotelephone.

Re Styalator: Well, yes, until the advent of LCD in the late '60s, merging tablet and screen was going to be a problem. The Dynabook, contemporary with Star Trek, was the realisable conceptual equivalent.

Re the expert system: they're already available in the form of MYCIN and descendants. Perhaps all that's interesting in the Tricorder is the apparent ability to quickly take vital stats without contact, which could of course be useful for pros and laymen.

Re:perspective

[1u3hr]

Star Trek's communicators functioned far more like cell phones than old military radiotelephones. They were personal devices that everyone kept on them all the time, and could be used to contact other individuals at will who also wore them, without respect to distance. Sound like anything we all carry these days?

They look vaguely like cell phones. They operate like walkie talkies. Cell phones require a network of cells to work. Communicators, like walkie talkies, can directly contact other units. Cell phones are useless on alien planets (unless Doctor Who has jailbroken it for you.)

Now's your chance?

[pushing-robot]

If you've ever watched Star Trek and said, 'Hey, I could build that,' then why the fuck haven't you?

Re:Now's your chance?

[whoda]

Because I didn't want to do it for Billions in profit.
But, doing it for $10 Million and just for fun sounds like a great idea.

Re:Now's your chance?

[syousef]

If you've ever watched Star Trek and said, 'Hey, I could build that,' then why the fuck haven't you?

Because I sobered up?

Tricorder Misdiagnosed

[Lead Butthead]

I don't believe the tricorders as presented in varies incarnation of Star Trek TV shows/movies are actually capable of diagnosing any ailment; each device is merely a collection of high precision sensors. The physician holding the device is the one that is making the diagnoses base on the data presented by the device.

Re:Just ask about vegetables eaten and vitamin D

[Theotherguy_1]

And get probably 75% of medical issues diagnosed and cured, as they are mostly nutritional deficiencies... :-)

Not true. Most are pathogenic infectious diseases and accidental injuries. Most in the west are chronic diseases like heart disease and cancer. These both have links in genetics, and yes, even diet (in the case of heart disease).

As is a good night's sleep, friends, family, a connection to that which is beyond us, meaningful work, daily exercise walking and such, and that kind of stuff.

Evidence please? I Know that sleep can have adverse effects on health, but it certainly won't cure "75% of diseases.'

And obviously avoid smoking, excessive alcohol, and obvious environmental toxins at work and play.

Please tell me what an "obvious environmental toxin" is.

The focus on magic bullets is unfortunate.

No. It isn't. It happens to work! Please study the history of modern medicine.

As is a focus on diagnosing things like cancer, heart disease, and diabetes that are mainly signs of vegetable deficiency disease and lack of vitamin D (and to a lesser extent those other issues).

This is a complete lie. Please stop spreading your nonsensical blather on the internet. Heart disease caused by "vegetable deficiency?" Cancer caused by lack of vitamin D? Do you live in the 15th century or something?

Most health rests on the basics. It's true that there are exotic genetic diseases and so on, but what causes the most chronic misery and early death in the industrialized words is these basic nutritional (and sunlight) problems.

Please provide evidence.

You see, what you seem to be promoting is called "holistic medicine." It was a cute practice of the ancient Greeks which sustained itself for many centuries. It was also, unfortunately, mostly wrong. We've made great advances in medicine since then. Please don't attempt to drag us back to the dark ages.

What are the rules?

[PCM2]

So for only $10 millions dollars you can buy a device that is worth billions. Yeah right.

What are the rules of the contest? Is there any language that says you can't file patents for your invention -- or parts of it -- before submitting? $10 million would be a nice chunk of seed capital.

Re:Simplify

[The Wild Norseman]

I think the bleeding from all of your orifices might rule out allergies.

Yes, but it still leaves open the possibility you're listening to a Justin Beiber album.

Viral diagnoses

[dachshund]

Yes, in all seriousness --- come up with a device that can instantly diagnose a bacterial or viral infection (specifically: is this swine flu, bird flu, or a cold) and you'll save a hell of a lot of lives when a pandemic comes round.

Even if it only works with a tiny number of pre-determined pathogens, that would be huge.