
Artificial intelligence is almost twice as effective at determining the degree of aggressiveness of a rare form of cancer than comparing scanners.
According to a new study, AI can distinguish details invisible to the naked eye and is 82% accurate compared to laboratory tests that are 44% accurate.
In the article published in Lancet Oncology, scientists used a well-established radiomics technique to identify signs invisible to the naked eye of retroperitoneal sarcoma, which appeared on scans of 170 patients in the connective tissue of the back of the abdomen.

Scientists at the Royal Marsden Hospital and Cancer Research Institute say it could improve treatments and benefit thousands of people a year. They are also excited about the potential it has to distinguish types of cancer early.
AI is showing great promise in diagnosing breast cancer and reducing the duration of treatment.
Computers can be loaded with vast amounts of information and trained to identify patterns and make predictions, solve problems and learn from their mistakes.
With this data, the AI algorithm was able to determine the tumor grade of 89 European and American patients using scanners, and the result was much more accurate than biopsies.