A recent study from Sweden suggests that artificial intelligence (AI) could aid in detecting certain aggressive types of breast cancers that may not be easily identified through traditional mammograms. The study revealed that when radiologists utilized AI to assist in reading mammograms, the detection rate of breast cancers improved, leading to a decrease in the number of advanced and aggressive cases.
Published in a recent issue of The Lancet, the study involved over 100,000 women who underwent mammography screenings. Half of the participants received AI support during their screenings, while the other half had their mammograms reviewed by two separate radiologists, a practice known as double reading commonly used in Europe but less so in Canada.
The research focused on interval cancers, which are invasive tumors that develop between routine mammograms and can be challenging to detect. The study showed a 12% reduction in interval cancers among the groups that utilized AI screening.
Dr. Kristina Lång, the lead author of the study and a breast radiologist at Lynn University in Malmo, Sweden, expressed optimism about the significant decrease in interval cancers observed. She highlighted the promising outcomes and emphasized the potential benefits of AI in breast cancer screenings.
In Canada, the Public Health Agency reported an estimated 30,500 breast cancer diagnoses in 2024, with a five-year survival rate of around 89%. Guidelines suggest that women over 40 can opt for mammograms every two to three years, and some provinces have initiated breast cancer screening at age 40, aligning with Sweden’s screening practices for individuals aged 40 to 74.
During the study, AI-supported mammograms were categorized based on risk level, streamlining the screening process. The results indicated that AI-assisted screenings led to fewer interval cancers compared to standard screenings, showcasing an improvement in overall screening efficacy.
While acknowledging the positive impact on interval cancer rates, Dr. Lång emphasized the need for long-term studies to assess the potential effects of AI-supported screenings on mortality rates. The study’s limitations included a single screening center and the absence of data on participants’ race and ethnicity.
Looking ahead, Dr. Lång mentioned that the next phase would involve evaluating the cost-effectiveness of AI in breast cancer screenings. Medical experts like Dr. Jean Seely from the Ottawa Hospital emphasized the cost-saving benefits of AI-assisted mammograms, highlighting their potential to not only save lives but also reduce healthcare expenditures.
Dr. Gerald Batist, an oncologist from Montreal, underscored the importance of investing in AI solutions for breast imaging, emphasizing the need for health ministers to prioritize innovation and technology in healthcare to keep pace with advancements in cancer detection.
