A team of researchers working in the fields of engineering and medicine at the University of Central Florida (UCF) has recently developed a new artificial intelligence system capable of spotting often-missed cancerous tumours.
“I believe this will have a very big impact,” said group leader and Engineering Assistant Professor Ulas Bagci. “Lung cancer is the number one cancer killer in the United States and if detected in late stages, the survival rate is only 17 percent. By finding ways to help identify earlier, I think we can help increase survival rates.”
The breakthrough was achieved by training the system to detect small bits of lung cancer in computerised tomography (CT) scans using the most optimal technique, which sometimes eludes radiologists given the same task with obvious consequences to the survival rates of lung cancer patients.
Employing a method similar to that used in training facial-recognition software, the research team fed the system over 1,000 CT scans collected in the database of the U.S. National Institutes of Health.
AI beats radiologists in spotting lung cancer. Image credit: Senior Airman Julianne Showalter, U.S. Air Force via afrc.af.mil, CC0 Public Domain.
Over time, the system had learned to ignore what’s irrelevant, i.e., nerve cells and other tissues, zero-in on lung cells and classify them as either normal or abnormal, the latter being an indication of potential malignancy.
“We used the brain as a model to create our system… You know how connections between neurons in the brain strengthen during development and learn? We used that blueprint, if you will, to help our system understand how to look for patterns in the CT scans and teach itself how to find these tiny tumours,” said Rodney LaLonde from UCF.
In trials conducted by the team, the system managed to successfully identify cancerous lesions 95 percent of the time, which is much better than its human counterparts, who got it right approximately 65 percent of the time.
Bagci is currently looking for partners to commence clinical trials after which – given success – the system could be as little as one or two years away from being deploying in hospital settings.
“I think we all came here because we wanted to use our passion for engineering to make a difference and saving lives is a big impact,” LaLonde says.
A paper detailing the invention (a pre-print version of which is available at arxiv.org) will be presented at the 21st International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2018) in Granada, Spain in September 2018.
Sources: paper, digitaljournal.com, today.ucf.edu.