 |
| Study Google's AI system can improve early detection of breast cancer |
Researchers in the US and UK have confirmed that the artificial intelligence system developed by Google has proven equally effective for radiologists and that mammography can help predict the potential for breast cancer in women because it should have a bright future.
The study was published in the journal Nature on Wednesday. The study showed that artificial intelligence has the potential to improve the accuracy of breast cancer detection, which affects one in eight women worldwide.
The American Cancer Society reports that radiologists have found that 20% of breast cancers in mammograms are wrong and that half of women who have been tested for more than 10 years get false positives.
"The results of this study, which was developed by the Alphabet Department of Amnesty International and merged with Google Health last September," said co-author Maziar Itzmadi of Northwestern Medical University in Chicago. Significant progress has been made in early detection of breast cancer.
The team (including researchers from Imperial College University in London and the NHS) trained the system to identify thousands of breast cancer x-rays and then compared their expectations to 25,856 breasts in the UK. Actual X-ray results were compared to 3,097 from the United States.
Studies have shown that artificial intelligence systems with the same experience as radiologists can identify cancer, while reducing the number of false alarms in the American group by 5.7% and the number of false alarms in the American group by 5.7% 1, the British group 2%. Considering the wrong test normal, it also reduced the number of false negatives, dropping 9.4% in the American group and 2.7% in the British group.
These differences reflect the way breast mammograms are read. In the United States, only one radiologist reads results and takes tests twice a year. In the UK, tests are taken every three years and the radiologist reads the results. If different, ask about a third.