Teenage Genius Creates Ai Program that Diagnoses Breast Cancer

It seems that over the past few years America’s next generation of scientists don’t want to wait until they’re adults to make their mark in science.

Cases in point include Angela Zhang, a 17-year-old Cupertino, California high school student who won the top prize honor and $100,000 for an individual entry from the Siemens Foundation’s yearly high school science competition.

What did she do to deserve such praise? Zhang created a nanoparticle that kills cancer stem cells.

And then there’s the amazing Clara Lazen. At just 10-years-old, the Border Star Montessori School student designed a new molecule that some experts predict could lead to the creation of new batteries, pharmaceutical drugs, material science technology—even explosives.

Now, the newest teenage genius has made her mark by winning the coveted Google Science Fair’s grand prize in Palo Alto, California. Her award includes an all-expense paid field trip to the Galapagos Islands and a $50,000 college scholarship.  

Brittany Wenger from Sarasota, Florida took the top honors by creating a remarkably effective breast cancer diagnosing app incorporating a sophisticated neural networking program that mimics the way a human brain thinks. In essence, it’s an artificial intelligence program that learns and adapts to new data, teaching itself to improve the diagnoses it makes.

What she created gets smarter and smarter.

Cloud4Cancer 99 percent correct

Her particular program, called Cloud4Cancer, can recognize and track patterns of data that are too complex for the human brain to decipher. And as the program matures it gets better and better at doing its job.

Step-by-step, Wegner investigated the most likely neural networks that might be adopted to serve as the foundation for her goal. She sought to develop an effective AI program to reduce the number of invasive tests women who are tested for breast cancer undergo. The least invasive test harvests cells for a biopsy employing a process called fine needle aspirate. But the test is not highly conclusive, so many patients must undergo a second, even a third fine needle aspiration.

After thoroughly testing three neural networks already on the market, she decided to go with a Java-based program she designed. The program seemed superior to the others and was more easily adapted to the purpose of enhancing the success rate of the biopsy test.

Reporting on Wenger’s progress and methodology, livescience.com states that when she tested her own program “with 681 fine needle aspirate samples, her program gave correct diagnoses for 94 percent of the cases and correctly identified more than 99 percent of the cancerous cases.”

More impressive still the breast cancer diagnostic program’s analysis was determined to give inconclusive results only in about four percent of the tests. Livescience.com notes that Wegner reported on her project webpage that “Less than one percent of the answers were false negatives—benign diagnoses for lumps that were actually cancerous, a result she especially wanted to avoid.”

This bested the neural networks sold commercially. Their programs returned false negatives around five percent of the time.

Hospital ready

Not only was her diagnostic neural network viable, it proved better than anything currently available. Wegner also sees the AI program being modified to enhance tests for ovarian and prostate cancers.  

Speaking about the results of her project with ABC News, Wegner said, “I think it might be hospital ready.”

As the ABC reporter expressed amazement at Wegner’s accomplishment, the young scientist said she intends to focus on advanced computer science at college and eventually become a pediatric oncologist.

The university that accepts her as a student should count itself lucky.