Teenager unleashes computer power for cancer diagnosis

Brittany Wenger Image copyright Royal Society of Medicine
Image caption Brittany Wenger: Can artificial intelligence do a better job than humans at cancer diagnosis?

She is only 19 but has already experimented with neural networks, built prototype software to help doctors diagnose breast cancer, won a $50,000 college scholarship from Google and been invited to the White House to showcase her research.

And her ambition doesn't stop there. Brittany Wenger wants a dual career as a paediatric oncologist and research scientist.

The teenager from Sarasota, Florida, became interested in neural networks - a form of artificial intelligence that continuously learns and mimics the human brain - in high school.

She was building a neural network that could play soccer, her favourite game, when her cousin was diagnosed with breast cancer, inspiring her to put her talents to medical use.

Ms Wenger came up with the idea of creating an artificial intelligence software program to analyse data from a breast tissue biopsy.

"In the simplest terms, I am trying to teach the computer how to diagnose cancer," she says.

"I'm trying to teach the computer how to think so it can detect patterns that allow it to diagnose cancers easier and quicker."

Coding challenges

In London, to speak at the Royal Society of Medicine, she is passionate about the idea of fusing the concepts of computer science and biology.

She says a key message is to inspire others to give it a try.

"Coding is not easy," she says. "This experiment flopped three times before it was successful and I had to scrap the entire experiment."

She believes the great thing about science is that you learn just as much from failed experiments as successful ones.

"I think it was Thomas Edison who said, 'I have not failed. I've just found 10,000 ways that won't work'," she says.

Image copyright SPL
Image caption Cancer cell: Could diagnosis be quicker and easier?

"It's all a puzzle and it's exciting to find the pieces that fit together. And in that moment, that eureka moment, when you do figure it out, it's definitely worth all the hard work."

The breast cancer programme is being tested in two hospitals, in the US and Italy.

She is also applying the same sort of approach to leukaemia, with a cloud-based computer program to find patterns in a patient's genetic signature that can be used to predict relapse.

Tailored treatments

Now a first-year student at Duke University, North Carolina, studying biology, she eventually wants to become both a paediatric oncologist and a research scientist.

"I was that kid who never outgrew the why? phase," she says. "And through this means of science I can not only come up with my own questions, but I can find the answers."

Commenting on the role of computers in cancer diagnosis, Dr Emma Smith, senior science information officer at Cancer Research UK, said they were essential in modern research as they can store and process large amounts of data.

"This depth of knowledge has already led to big steps forward in diagnosing cancer and getting patients more tailored treatments.

"But any new technology needs to be thoroughly tested in clinical trials to prove it's as least as good as the current system before it can be used in the clinic."

Hear more on Science in Action BBC - Podcasts and Downloads - Science in Action

Related Topics

More on this story

Related Internet links

The BBC is not responsible for the content of external Internet sites