A team of researchers from the University of Central Florida have developed an artificial intelligence system to detect cancerous tumors normally missed by radiologists, the university stated in a blog post.
The team has trained a computer to identify minute particles of lung cancer in CT scans, which are otherwise difficult for radiologists to pinpoint. The researchers claim that the system is about 95% accurate at identifying cancer versus 65% by humans, the blog added.
The AI uses an algorithm similar to facial-recognition software that scans thousands of faces to identify and match a particular pattern.
"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 tumors," said Rodney LaLonde, a doctoral candidate, in the blog post.
The system was developed at the university’s Center for Research in Computer Vision, which focuses on AI and its medical applications, and the research team was led by engineering assistant professor Ulas Bagci.
To train the computer to identify cancer tumors, the researchers developed the software by feeding it 1,000 CT scans, which were supplied by the National Institutes of Health in collaboration with the Mayo Clinic, the blog post stated.
According to Bagci, the team will implement the project in hospitals and will release the technology in the market within one or two years.
The team plans to present its finding at the annual medical imaging research conference, MICCAI, which will be hosted in Spain in September 2018.
Earlier this week, in a similar development, researchers at the Massachusetts Institute of Technology (MIT) reported that it has developed an in-body GPS system to treat cancer.