The article was updated at 09.45 on January 11, 2019.
The doctors behind the study created an algorithm based on more than 60,000 simple images of cervices collected in Costa Rica. From the beginning of the 1990s, 9,400 women were followed for periods of up to 18 years.
When the researchers used automated scans of the images, the hit was clearly better at identifying cell changes that were really developed for cancer or precursors of cancer, the study published in the Journal of the National Cancer Institute. This is in comparison with the fact that human experts have assessed the images.
In the case of women aged 25-49, with whom the risk of suffering is greatest, AI technology was even safer.
The research involved a screening technique in which healthcare professionals apply diluted acetic acid to the cervix, where cellular changes can appear as white. The method is mainly used in countries with less money to carry out broad sample checks, which are considered safer.
In the next few years, doctors want to mix patients around the world with the further testing of AI technology.
Back to E-health