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:
There are two aspects to the reported work. The first is the use of carbon nanotube biosensors incorporating antibodies. Google shows they've been around for a while and are reportedly more sensitive than previous methods. The second is using the levels of mesothelin to diagnose pancreatic cancer, this is less widely established, but was also recently reported by a different lab (http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjourna...).
Putting the two together seems to be the work of Jack Andraka.
The fact that a 15 year old digested that information and put the two together is impressive in itself. Who cares if he was standing on the shoulders of giants? He bothered to gaze towards the horizon rather than at his own navel.
I agree. Also, it's possible that his work will actually directly lead to a test for pancreatic cancer, though that would take more work to establish, obviously.
