Student creates an artificial "brain" to detect breast cancer inspired by cousin

"When my cousin got breast cancer, she was telling me about how difficult the diagnostic process was. I researched and found out about fine needle aspirates, which are the least invasive, quickest and cheapest [diagnostic] procedures that a woman can have. Right now, they’re wildly inconclusive, so a lot of doctors refuse to use them. Only about five hospitals in the United States still use the procedure. I was really interested in trying to revive them, and when I found out that the major problem was how difficult the patterns were for pathologists to diagnose, I wanted to create a tool to aid the doctors", said Brittany

As early as the seventh grade, Wenger was experimenting with advanced computer science and artificial intelligence programs. From those early experiments, Wenger has gone on to write computer programs that improve the accuracy of tests for breast cancer and leukemia, below, she explains how they work:
"Based on cell morphology, which is how the cells look, the program will look for patterns and try to determine whether a person has cancer or not. For example, multi-layered cells are an indicator that a person may have cancer. It’s a little more complicated than that, though, because masses will exhibit both characteristics of non-cancerous masses and cancerous masses, even though clearly one mass can only be cancer or not cancer. (...) The detector is meant to be an initial screening mechanism—it’s trying to go through and determine whether the cells are cancerous or not. If it returns a diagnosis that the cells are cancerous, further testing would be required to determine the specific type", the inventor described.

Adapted from: http://bit.ly/1jRVrEY

More info: https://cloud4cancer.appspot.com/
https://www.youtube.com/watch?v=FiUmjmOKlto