A groundbreaking multicenter study conducted at 12 screening centers in Germany has shown that the use of artificial intelligence (AI) in dual-read mammograms increases detection rates of breast cancer by 17.6%. The research, focused on women between 50 and 69 years old, compared traditional reading performance with AI assistance, showing a significant positive impact on screening programs.
Early detection of breast cancer, essential to improve survival rates, faces challenges such as the high workload of radiologists and the growing shortage of specialists. The majority of mammograms analyzed do not show signs of cancer, making the accuracy and efficiency of the diagnosis essential to avoid delays or errors.
Incorporating AI into mammography analysis seeks to address these challenges by reducing the burden on radiologists and improving both the sensitivity and specificity of screening.
Posted in ‘Nature Medicine‘, the observational study included 463,094 women, of whom 260,739 underwent AI-assisted analysis and 201,079 formed the control group with double standard reading.
Key findings included a increased cancer detection: The detection rate was 6.7 per 1,000 participants in the AI group, compared to 5.7 per 1,000 in the control group.
Reduction of false positives: Although the differences were slight, repetition rates (number of participants requiring additional testing) were slightly lower with AI: 37.4 per 1,000 versus 38.3 per 1,000.
Improved positive predictive value: The proportion of suspicious findings that turned out to be cancer was 17.9% with AI, compared with 14.9% in the control group. In cases requiring biopsy, the positive predictive value was 64.5% with AI, compared to 59.2% with standard methods.
Although false positive rates were similar between both groups, the researchers highlighted that AI’s ability to detect more cases of cancer without significantly increasing recurrence rates marks a crucial advance.
The success of this approach in a large-scale cohort reinforces the recommendation to further develop and use AI in mammography screening programs. With the growing need for technological solutions in healthcare, this study points to a promising path toward earlier and more accurate detection of breast cancer.
#improves #cancer #detection #mammography #study #women
