This kid is vaguely local to me, so I've been following the story on our local news outlets. Yes, he was rejected by a lot of different scientists, but there are LOTS of labs in this part of Maryland.
Here's a more detailed article from Forbes that describes the process better:
And his method has been tested and it works. The article states that a couple of drug companies are bidding on the rights to put his test strips into production.
Check out his wiki article for even more information.
Andraka cultured MIA PaCa cells, from a commercial pancreatic carcinoma cell line, which overexpress mesothelin, a biomarker for pancreatic cancer. The mesothelin was isolated, concentrated and quantified with ELISA. After optimization with the Western Blot assay, the human mesothelin-specific antibodies were mixed with single walled carbon nanotubes and used to coat strips of ordinary filter paper. This made the paper conductive. The optimal layering was determined using a scanning electron microscope. Cell media spiked with varying amounts of mesothelin was then tested against to the paper biosensor and any change in the electrical potential of the sensor strip (due to the changing conductivity of the nanotubes) was measured, before and after each application. Specifically, what happened was this:
“ The antibodies would bind to the mesothelin and enlarge. These beefed-up molecules would spread the nanotubes farther apart, changing the electrical properties of the network: The more mesothelin present, the more antibodies would bind and grow big, and the weaker the electrical signal would become. ”
A dose-response curve was constructed with an R2 value of 99.92%. Tests on human blood serum obtained from both healthy people and patients with chronic pancreatitis, pancreatic intraepithelial neoplasia (a precursor to pancreatic carcinoma), or pancreatic cancer showed a similar response. The sensor’s limit of detection sensitivity was found to be 0.156 ng/mL; 10 ng/mL is considered the level of overexpression of mesothelin consistent with pancreatic cancer. Andraka's sensor costs $3.00 and 10 tests can be performed per strip, taking 5 minutes each. The method is 168 times faster, 26,667 times less expensive, and 400 times more sensitive than ELISA, and 25% to 50% more accurate than the CA10-9 test.
Officials at Intel have said that Andraka's method is more than 90 percent accurate in detecting the presence of mesothelin.