It may be a new weapon in the fight against this nasty disease.
Breast cancer is the most common cancer in women. Despite significant improvements in treatment methods, around 500,000 women worldwide die from the disease every year. Population surveys should help to detect breast cancer early so that the number of women who die from it is reduced. However, an enormous number of women must be screened during these population screening programs. And radiologists could use some help with this. Certainly in some countries with a shortage of radiologists who focus on breast cancer. And perhaps that artificial intelligence can offer a solution.
ANIMALS
In the fight against cancer, researchers also look at how often the disease occurs in some animals. And it shows that elephants are quite resistant to the disease . Whales, bats and naked moles also rarely get cancer. By understanding why they are less likely to get cancer, it may be possible to find ways to stop the disease in people as well.
Mammography
During the population screening, digital X-rays are made of the breast. This is also called mammography. In the new study published in the Journal of the National Cancer Institute, the researchers wondered if computers would be able to recognize breast cancer in these photos. And so the researchers compared the performance of an AI computer with that of 101 radiologists.
Exams
A total of 2,652 exams were assessed, of which 653 had breast cancer mammograms. The exams were assessed by both the AI computer and 101 independent radiologists. And the results show that the AI computer is not inferior to an experienced radiologist. It appears that both the computer and the radiologists provide a 95% reliable diagnosis. This means that the AI computer can match the radiologist when it comes to breast cancer detection.
“It was interesting to see that AI systems can reach the level of radiologists,” says researcher Ioannis Sechopoulos. “And not only that, they match the performance of experienced radiologists who spend a significant amount of their time reading mammograms.” Although the findings are promising, the researchers state that even more research is needed before AI computers are actually used.