2023-08-07 12:53:00
In France, 2.7 million women had a mammogram in 2021.
Recommended from the age of 50, this examination is recommended every two years until the age of 74. This helps to prevent breast cancer in time.
This examination consists of an x-ray of the chest in order to detect any abnormalities.
Once the images have been taken, a first reading is carried out by a first radiologist. To ensure the results, a second professional reads the pictures. This therefore requires twice the time, and therefore increases the workload of doctors.
A study Swedish
A recent study has shown that AI can be used as an alternative to the second reading required during a mammogram. The trial conducted by researchers at Lund University in Sweden was published in The Lancet Oncology.
“In our trial, we used AI to identify high-risk breast cancer screening exams, which were double-read by radiologists. The remaining exams were classified as low risk and were read by only one radiologist. In screen reading, radiologists used AI as a detection aid, in which it highlighted suspicious findings on images,” explained Kristina Lång, researcher and associate professor in diagnostic radiology at the University of Lund and consultant at Skåne University Hospital, who led the study.
Mammography of the future?
To conduct the study, the 80,003 participating women were randomly divided into two groups:
- 40,003 women in the intervention group who underwent AI-assisted screening
- and 40,030 in the control group who underwent a standard double read without AI assistance.
The results showed that the use of AI led to the detection of 20% more cancers compared to standard screeningwithout affecting false positives.
Another interesting result, the study showed that the screen reading workload of radiologists was reduced by 44%. This is good news because, as the study explains, there is a shortage of medical personnel in radiography centres. AI could then make up for this lack by allowing the staff on site to devote their time to other patients.
1691447181
#artificial #intelligence #diagnosis