"I actually love single-walled carbon nanotubes; they're like the superheroes of material science." I feel the same way. This kid is genuinely excited and interested which, is extraordinary. I love seeing youngsters getting excited about science.
Unfortunately, this blog post and the one it claims as a source [1] are rather fluffy on details. Justin organized this information and presentation for the Intel Science and Engineering Fair (ISEF) [2]. More details about the process he discovered can be found on the ISEF 2012 profile page [3] and for those who don't want to follow the trail, I've reproduced it here:
"Pancreatic cancer is a devastating disease with a five-year survival rate of 5.5%. One reason for this is the lack of a rapid, sensitive, inexpensive screening method. A novel paper sensor is described that simply, rapidly and inexpensively screens for pancreatic cancer. Mia Paca cells overexpressing mesothelin, a biomarker for pancreatic cancer, were cultured; mesothelin was isolated, concentrated and quantified with ELISA. After optimization with the Western Blot assay, the antibody to human mesothelin was dispersed with single walled carbon nanotubes. This dispersion was used to dip-coat strips of filter paper, rendering the paper conductive. Optimal layering was determined using a scanning electron microscope. Cell media spiked with varying amounts of mesothelin was applied to the paper biosensor. Change in electrical potential was measured before and after application and a dose-response curve was constructed with an R2 value of 99.92%. In vivo tests on human blood serum obtained from healthy people and patients with chronic pancreatitis, PanIn, pancreatic cancer revealed the same trends.. The sensor’s limit of detection was found to be 0.156 ng/mL, satisfying the limit of 10 ng/mL, the level considered an overexpression of mesothelin consistent with pancreatic cancer. The sensor costs $3.00; 10 tests can be performed per strip. A test takes 5 minutes and is 168 times faster, 26,667 times less expensive, and 400 times more sensitive than ELISA, 25% to 50% more accurate than the CA10-9 test and is a sensitive, accurate, inexpensive, and rapid screening tool to detect mesothelin, a biomarker for pancreatic cancer."
Every time an ISEF article pops up, I always sigh out loud and begin a diatribe [1] about why the winner (invariably in medicine or biology) and the title (invariably "teenager cures XYZ") is misleading.
To my surprise, it looks like this guy actually did a lot of the research on his own. At the very least, it appears the ideas was authentically his own, and then he enlisted the help of a lab to accomplish it.
Short explanation: Paper covered with carbon nanotubes is conductive. This conductivity changes when cancer marker cells are present. The cancer markers cells are drawn in by their antibodies, which are pre-mixed with the carbon nanotubes.
Thanks for this. This is a great story but one of the type that is prone to being overhyped by writers whose angle is "man bites dog." As another commenter pointed out, the test is still very preliminary:
Whether or not the test is a smashing success is not the only point of inspiration. I think the field of medicine is extremely prone to disruption by clever, process-focused laymen, whether they are 15 or 30-years of age. The field of medicine is filled with brilliant people, but it's also held back by institutional tradition and authority, i.e. "that's just the way things are done."
Frequent readers of HN are probably familiar with Atul Gawande's article on the checklist, in which hospitals who mandated that surgeons and nurses follow a simple list of steps (literally, as simple as "wash your hands") were able to reduce the rate of needless post-surgical infections to zero.
http://www.newyorker.com/reporting/2007/12/10/071210fa_fact_...
I believe Merck's first big drug came from reading a research paper that had apparently been overlooked. In a book I've been reading about how companies develop drugs, a veteran of the industry wrote about how a potential multi-billion dollar drug was almost shelved because the company's analysts had forgotten to remove dummy data from their test-analysis routines...and that this kind of thing happens all the time, apparently.
And of course, most people are pretty familiar with the calcified information-systems of their doctors' offices. Those things remain in place because that's how they've been done in the past, by many other brilliant people, so why re-examine them?*
(to be fair, medical information systems are also hindered by medical privacy laws governing the copying/distribution of patient data)
Edit: Just after I posted this, I came across this NYT article about a boy who went to the hospital after cutting his arm diving for a basketball and died from infection-related complications. The upshot:
> In a statement, the hospital said that emergency physicians and nurses would be “immediately notified of certain lab results suggestive of serious infection, such as elevated band counts.” Rory Staunton’s bands, or a type of white blood cell, were nearly five times as high as a normal level.
> The hospital has developed a new checklist to ensure that a doctor and nurse have conducted “a final review of all critical lab results and patient vital signs” before a patient leaves, Lisa Greiner, a spokeswoman, said in the statement.
Some might not see how awesome this is, but for example: My dad'd blood work came up perfect year after year, zero symptoms. Even 1 year before pancreatic cancer metastasize. There are already markers to detect it and it only takes adding blood to it ,but the price is prohibiting so it's not a standard test.
The idea of sets of super cheap markers which can are included all the time by default in blood tests is incredible.
Unfortunately, I've missed the mark where I can edit my comment, but I wanted to correct myself and say "Jack" organized this information; not "Justin."
>he’s just pioneered a new, improved test for diagnosing pancreatic cancer that is 90% accurate
Without more context, that claim is meaningless. Surely it cannot mean "this test gets the right answer 90% of the time when used on the general population". Nearly 99.99% of people do not have pancreatic cancer, so if I make a program that prints out the sentence "You don't have pancreatic cancer.", then my "test" would have a 99.99% accuracy rate against the general population.
One of the things I teach in a unit on information retrieval is the notion of precision and recall, and tradeoffs between them, and the idea that a single "accuracy" rate is almost never useful for these kinds of identification tasks.
I've often thought that such a unit would be a good idea for journalists and people in general; the language of reporting on and talking about a lot of scientific advances (medical in particular) is so damn imprecise about this stuff. Even if they report only one number, they could at least say which one it is. 90% could plausibly be either the precision or the recall here, or even F-measure or some other average of them, but as you point out, it's unlikely to be "accuracy" in the sense of (true positives + true negatives) / N, or it wouldn't be very impressive.
It must mean that it is 90% accurate in identifying cancer. If your test was given 100 people with PC it would be 100% wrong but this test would be 90% correct. That is how I am reading it anyway.
From the article... "Yes, he even got in trouble in his science class for reading articles on carbon nanotubes instead of doing his classwork."
This just irks me about our (U.S.) education system. I understand doing classwork is important. Yet here you have a young student taking interest in science and engineering and gets in trouble for it? That doesn't make sense to me at all. The work he's doing should be more than enough credit for his classwork. Take him aside from the class and foster this work and passion. He'll learn more science from this one project than doing 10x the amount of classwork in any syllabus. I see no reason why his science grade could be reflected on this work.
I expect the biggest problem is that giving teachers flexibility to give class credit for random projects outside the syllabus would quickly become the much easier route to passing the class. The more restrictions you put on the project, the more you take away from the students interest ("sorry, carbon nanotubes is not one of the allowable project materials") and the more likely you'll need to demand unreasonable beaucracy ("the published paper doesn't include the teachers name as lead author so it doesn't count. write it up again without plagerising").
Plus there are some benefits of a syllabus, such as making sure that fundamental knowledge is taught. If you (as a student) missed algebra because you were doing your investigation into fractals, you're going to find calculus much more difficult than you would otherwise.
Education is a competitive sport. You don't let Olympic sprinters waste time playing basketball, or they won't win any medals. If kids waste time learning about stuff that's not on the exam, they won't get the marks they deserve, and some other kid will get their place in university.
I like the quote but I think people throw it around far too loosely. It seems to be an impossibility or else every company founded by A class people (assumption: there must be at least SOME such companies) would only ever consist of A class people, but that clearly isn't the case. For sake of argument let's assume Google, Facebook and Microsoft (and likely many, many others) were all founded by A class people. Do these companies consist solely of A class people? If not someone, somewhere, must have hired a non A class person, thus the hirer themseves couldn't have been A class by the original statement. You can apply the recursive argument here eventually terminating in the founders :)
EDIT: Add Apple to the list of companies with (obvious) A class level founding teams. Funny oversight on my part since the quote is from Steve Jobs himself :)
I bet many of the other applications were also to top institutions. It wouldn't be a good fit for most labs even if a Nobel Laureate were the one wanting to do it.
I have a MD and a background in biochemistry, and can say most people I've met in the field of oncology don't nearly have the creativeness and audacity of this kid's idea. Essentially he's using the carbon nanotube materials, sending an electric current through it, and detecting minute voltage changes that occur when cancer proteins bind to the antibodies laced over the nanotubes.
The cost of using nanotubes to detect cancer, however, may be quite prohibitive. Think 500-1000 dollars per test. On a population screening level, I don't think most governments or insurance companies would go for it.
Because "we" are mediocre, our children follow after us, staring out of classroom windows while teachers prattle on about only-God-knows-what.
When we are inspired and and challenge our kids to do more, they come up with s* like this. Age is no barrier! Bravo.
Great story. You have to admire this young man's drive, determination, and obvious intellectual precocity. He'll surely go far (he already has).
I also think it's inspirational b/c it points up the importance of synthetic scientific research done by folks who are so to speak 'out of the fold'. We need institutional science, clearly, but we also need people who can take a step back and look at the big picture and see how things fit together.
A new pancreatic cancer test is absolutely newsworthy on its own (especially if this one proves out in broader studies). The fact that it was done by a 15-year old makes it even more impressive. It's not unreasonable to applaud a young person for being way ahead of the curve.
You must not have much experience being around teenage boys. An improved test for pancreatic cancer is not something normally coming forth from that demographic.
Just so it's clear, these were individual labs at universities that rejected his request. There are a lot of legitimate reasons they may not have wanted to take him on, including:
-Not enough physical space in lab - most wet labs are crowded with 2-3 people per bench
-No one with enough time to supervise a high school student bumbling around the lab. You basically need a grad student or post-doc babysitting new people (HS student or new grad, doesn't matter) since they don't know where anything is or how to do anything
-Not enough money. Funding has to come from somewhere, and if they don't have money to throw at this particular project they won't.
-Not the focus of the lab. I would hazard a guess that the HS student emailed anyone that was remotely involved in pancreatic cancer or material science. Realistically, there are probably a handful of labs that do this sort of research and the kid just didn't know which ones to email...so he emailed everyone.
-Other. Maybe the professor is teaching? On sabbatical? Already has ten undergrads?
The lab he worked in seems very well-suited for this project. I find it unlikely that there are really 200 labs in the Baltimore/Washington area that are. So I'm guessing much of the reaction is "good for this kid, it never would have worked here."
I've actually read about this before, but still never found any technical details about how the test works. My inner conspiracy theorist worries that a pharmaceutical corporation has bought this idea up.
I was interested to see what criteria were used to measure performance, but I haven't found a paper written about it. The most detailed information I could find was from here:
Is there a law for "something"-year-old does "something" headlines?
I see them all the time - and I hate them because I cynically predict that what ever it is that they are about to tell me is going to be either trivial, false or hyped.
I don't blame the kids - I guess it's just to get the clicks/attention for the news organisations.
If your age is being touted as the reason we should care about what you have done - then the thing that follows is probably bullshit, trivial or hyped.
Have you ever heard "34"-year-old tech CEO launches new multi billion dollar product line?
I hate to sound like a dick, but it bugs me that MAKE claims him as one of their own by arbitrarily labeling him a "Maker." He's not a maker, he's a research scientist.
[+] [-] randomdrake|13 years ago|reply
Unfortunately, this blog post and the one it claims as a source [1] are rather fluffy on details. Justin organized this information and presentation for the Intel Science and Engineering Fair (ISEF) [2]. More details about the process he discovered can be found on the ISEF 2012 profile page [3] and for those who don't want to follow the trail, I've reproduced it here:
"Pancreatic cancer is a devastating disease with a five-year survival rate of 5.5%. One reason for this is the lack of a rapid, sensitive, inexpensive screening method. A novel paper sensor is described that simply, rapidly and inexpensively screens for pancreatic cancer. Mia Paca cells overexpressing mesothelin, a biomarker for pancreatic cancer, were cultured; mesothelin was isolated, concentrated and quantified with ELISA. After optimization with the Western Blot assay, the antibody to human mesothelin was dispersed with single walled carbon nanotubes. This dispersion was used to dip-coat strips of filter paper, rendering the paper conductive. Optimal layering was determined using a scanning electron microscope. Cell media spiked with varying amounts of mesothelin was applied to the paper biosensor. Change in electrical potential was measured before and after application and a dose-response curve was constructed with an R2 value of 99.92%. In vivo tests on human blood serum obtained from healthy people and patients with chronic pancreatitis, PanIn, pancreatic cancer revealed the same trends.. The sensor’s limit of detection was found to be 0.156 ng/mL, satisfying the limit of 10 ng/mL, the level considered an overexpression of mesothelin consistent with pancreatic cancer. The sensor costs $3.00; 10 tests can be performed per strip. A test takes 5 minutes and is 168 times faster, 26,667 times less expensive, and 400 times more sensitive than ELISA, 25% to 50% more accurate than the CA10-9 test and is a sensitive, accurate, inexpensive, and rapid screening tool to detect mesothelin, a biomarker for pancreatic cancer."
[1] - http://www.fastcoexist.com/1680194/meet-the-15-year-old-who-...
[2] - http://www.societyforscience.org/isef/
[3] - http://apps.societyforscience.org/intelisef2012/project.cfm?...
[+] [-] polyfractal|13 years ago|reply
To my surprise, it looks like this guy actually did a lot of the research on his own. At the very least, it appears the ideas was authentically his own, and then he enlisted the help of a lab to accomplish it.
[1] http://hackerne.ws/item?id=3329605
Edit: Because HN is killing old permalinks: http://hackerne.ws/item?id=3328995
[+] [-] brittohalloran|13 years ago|reply
[+] [-] danso|13 years ago|reply
http://www.pancan.org/section_research/strategic_research_pr...
Whether or not the test is a smashing success is not the only point of inspiration. I think the field of medicine is extremely prone to disruption by clever, process-focused laymen, whether they are 15 or 30-years of age. The field of medicine is filled with brilliant people, but it's also held back by institutional tradition and authority, i.e. "that's just the way things are done."
Frequent readers of HN are probably familiar with Atul Gawande's article on the checklist, in which hospitals who mandated that surgeons and nurses follow a simple list of steps (literally, as simple as "wash your hands") were able to reduce the rate of needless post-surgical infections to zero. http://www.newyorker.com/reporting/2007/12/10/071210fa_fact_...
I believe Merck's first big drug came from reading a research paper that had apparently been overlooked. In a book I've been reading about how companies develop drugs, a veteran of the industry wrote about how a potential multi-billion dollar drug was almost shelved because the company's analysts had forgotten to remove dummy data from their test-analysis routines...and that this kind of thing happens all the time, apparently.
And of course, most people are pretty familiar with the calcified information-systems of their doctors' offices. Those things remain in place because that's how they've been done in the past, by many other brilliant people, so why re-examine them?*
(to be fair, medical information systems are also hindered by medical privacy laws governing the copying/distribution of patient data)
Edit: Just after I posted this, I came across this NYT article about a boy who went to the hospital after cutting his arm diving for a basketball and died from infection-related complications. The upshot:
> In a statement, the hospital said that emergency physicians and nurses would be “immediately notified of certain lab results suggestive of serious infection, such as elevated band counts.” Rory Staunton’s bands, or a type of white blood cell, were nearly five times as high as a normal level.
> The hospital has developed a new checklist to ensure that a doctor and nurse have conducted “a final review of all critical lab results and patient vital signs” before a patient leaves, Lisa Greiner, a spokeswoman, said in the statement.
http://www.nytimes.com/2012/07/19/nyregion/after-rory-staunt...
[+] [-] checoivan|13 years ago|reply
The idea of sets of super cheap markers which can are included all the time by default in blood tests is incredible.
[+] [-] randomdrake|13 years ago|reply
[+] [-] samstave|13 years ago|reply
Yep this test is gong nowhere.
[+] [-] mistercow|13 years ago|reply
Without more context, that claim is meaningless. Surely it cannot mean "this test gets the right answer 90% of the time when used on the general population". Nearly 99.99% of people do not have pancreatic cancer, so if I make a program that prints out the sentence "You don't have pancreatic cancer.", then my "test" would have a 99.99% accuracy rate against the general population.
[+] [-] blahedo|13 years ago|reply
I've often thought that such a unit would be a good idea for journalists and people in general; the language of reporting on and talking about a lot of scientific advances (medical in particular) is so damn imprecise about this stuff. Even if they report only one number, they could at least say which one it is. 90% could plausibly be either the precision or the recall here, or even F-measure or some other average of them, but as you point out, it's unlikely to be "accuracy" in the sense of (true positives + true negatives) / N, or it wouldn't be very impressive.
[+] [-] ShawnBird|13 years ago|reply
[+] [-] larrydag|13 years ago|reply
This just irks me about our (U.S.) education system. I understand doing classwork is important. Yet here you have a young student taking interest in science and engineering and gets in trouble for it? That doesn't make sense to me at all. The work he's doing should be more than enough credit for his classwork. Take him aside from the class and foster this work and passion. He'll learn more science from this one project than doing 10x the amount of classwork in any syllabus. I see no reason why his science grade could be reflected on this work.
[+] [-] polymatter|13 years ago|reply
Plus there are some benefits of a syllabus, such as making sure that fundamental knowledge is taught. If you (as a student) missed algebra because you were doing your investigation into fractals, you're going to find calculus much more difficult than you would otherwise.
[+] [-] wisty|13 years ago|reply
[+] [-] rogerbinns|13 years ago|reply
Here is a wonderful 11 minute talk by Ken Robinson http://www.youtube.com/watch?v=zDZFcDGpL4U
[+] [-] lifeisstillgood|13 years ago|reply
199 rejections, and the Professor that decided to let a 15 year old try a wild idea is at ... Johns Hopkins.
America may be doing a lot of things foolishly, but mentoring great talent for the future seems taken care of
[+] [-] ryanmolden|13 years ago|reply
I like the quote but I think people throw it around far too loosely. It seems to be an impossibility or else every company founded by A class people (assumption: there must be at least SOME such companies) would only ever consist of A class people, but that clearly isn't the case. For sake of argument let's assume Google, Facebook and Microsoft (and likely many, many others) were all founded by A class people. Do these companies consist solely of A class people? If not someone, somewhere, must have hired a non A class person, thus the hirer themseves couldn't have been A class by the original statement. You can apply the recursive argument here eventually terminating in the founders :)
EDIT: Add Apple to the list of companies with (obvious) A class level founding teams. Funny oversight on my part since the quote is from Steve Jobs himself :)
[+] [-] michaelhoffman|13 years ago|reply
[+] [-] drchiu|13 years ago|reply
I have a MD and a background in biochemistry, and can say most people I've met in the field of oncology don't nearly have the creativeness and audacity of this kid's idea. Essentially he's using the carbon nanotube materials, sending an electric current through it, and detecting minute voltage changes that occur when cancer proteins bind to the antibodies laced over the nanotubes.
The cost of using nanotubes to detect cancer, however, may be quite prohibitive. Think 500-1000 dollars per test. On a population screening level, I don't think most governments or insurance companies would go for it.
[+] [-] politician|13 years ago|reply
[+] [-] drharris|13 years ago|reply
[+] [-] spyder|13 years ago|reply
[+] [-] martinshen|13 years ago|reply
[+] [-] nzeribe|13 years ago|reply
[+] [-] benthumb|13 years ago|reply
I also think it's inspirational b/c it points up the importance of synthetic scientific research done by folks who are so to speak 'out of the fold'. We need institutional science, clearly, but we also need people who can take a step back and look at the big picture and see how things fit together.
[+] [-] marknutter|13 years ago|reply
[+] [-] bcrescimanno|13 years ago|reply
[+] [-] twistedanimator|13 years ago|reply
[+] [-] snowwrestler|13 years ago|reply
Why would this tire you out, though?
[+] [-] majorapps|13 years ago|reply
[+] [-] polyfractal|13 years ago|reply
-Not enough physical space in lab - most wet labs are crowded with 2-3 people per bench
-No one with enough time to supervise a high school student bumbling around the lab. You basically need a grad student or post-doc babysitting new people (HS student or new grad, doesn't matter) since they don't know where anything is or how to do anything
-Not enough money. Funding has to come from somewhere, and if they don't have money to throw at this particular project they won't.
-Not the focus of the lab. I would hazard a guess that the HS student emailed anyone that was remotely involved in pancreatic cancer or material science. Realistically, there are probably a handful of labs that do this sort of research and the kid just didn't know which ones to email...so he emailed everyone.
-Other. Maybe the professor is teaching? On sabbatical? Already has ten undergrads?
[+] [-] michaelhoffman|13 years ago|reply
[+] [-] molossus|13 years ago|reply
[+] [-] hirenj|13 years ago|reply
http://www.pancan.org/section_research/strategic_research_pr...
Essentially, this seems pretty damn early as far as results go. There will probably be a few years before we know it actually works or not.
[+] [-] gavinlynch|13 years ago|reply
[+] [-] confluence|13 years ago|reply
I see them all the time - and I hate them because I cynically predict that what ever it is that they are about to tell me is going to be either trivial, false or hyped.
I don't blame the kids - I guess it's just to get the clicks/attention for the news organisations.
If your age is being touted as the reason we should care about what you have done - then the thing that follows is probably bullshit, trivial or hyped.
Have you ever heard "34"-year-old tech CEO launches new multi billion dollar product line?
No. You read "Google launches Project Glass".
[+] [-] dreadsword|13 years ago|reply
[+] [-] lhartwich|13 years ago|reply
[+] [-] vtry|13 years ago|reply
[+] [-] saddino|13 years ago|reply