A major study in Sweden involving over 105,000 women found that artificial intelligence (AI) significantly improves breast cancer detection during mammography screening. The trial, known as MASAI, showed that AI-supported screening achieved an 80.5% sensitivity rate, a notable increase compared to the 73.8% sensitivity of traditional double reading by radiologists. This means AI helped detect 9% more cancers at the screening stage.Results from the randomized controlled trial were recently published in The Lancet journal.[trial+11]
AI Improves Cancer Detection and Reduces Radiologist Burden
The MASAI trial, conducted across four sites in Sweden from April 2021 to December 2022, enrolled 105,934 women for routine breast cancer screening.The median age of participants was about 54 years.Researchers compared AI-supported mammography with the standard method where two radiologists independently review images. The AI system not only improved sensitivity but also maintained an identical specificity of 98.5%, indicating it did not increase the rate of false alarms.The false positive rates were similar between both groups, with 1.5% for the AI-supported group and 1.4% for the control group.[trial+11]
A key benefit of the AI system was its ability to reduce the workload for radiologists by almost half, specifically a 44% reduction in screen-reading time.The AI tool helped by triaging mammograms, assigning low-risk cases to a single radiologist reading and high-risk cases to a double reading.It also highlighted suspicious findings to assist radiologists in their review.This efficiency gain is crucial, especially with growing shortages of breast radiologists.[news-medical+12]
Dr. Kristina LÃ¥ng, a lead author from Lund University in Sweden, highlighted the potential. "Widely rolling out AI-supported mammography in breast cancer screening programs could help reduce workload pressures among radiologists, as well as helping to detect more cancers at an early stage, including those with aggressive subtypes," Dr. LÃ¥ng said. She cautioned that implementation must be done carefully, using tested AI tools and with continuous monitoring.[theguardian+2]
Fewer Cancers Missed Between Screenings
Beyond initial detection, the trial also found that AI-supported screening led to a 12% reduction in interval cancers. Interval cancers are those diagnosed between scheduled screening rounds, often considered more aggressive because they grow quickly or were missed in the previous screening. In the AI-supported group, there were 1.55 interval cancers per 1,000 women, compared to 1.76 per 1,000 women in the standard screening group.[news-medical+7]
The AI-supported approach also detected fewer aggressive types of cancer. The study reported 16% fewer invasive cancers (75 versus 89) and 27% fewer aggressive sub-type cancers (43 versus 59) in the AI group compared to the control group. This suggests that AI helps in the early detection of clinically relevant breast cancers, potentially leading to better patient outcomes.[news-medical+8]
AI as a Support Tool, Not a Replacement
The researchers emphasized that the AI system is designed to support, not replace, healthcare professionals. Jessie Gommers, a PhD student from Radboud University Medical Centre in the Netherlands and first author of the study, noted that AI-supported mammography screening still requires at least one human radiologist to perform the screen reading. The AI acts as an assistant, helping radiologists make more accurate and efficient diagnoses.[news-medical+1]
The study does have some limitations. It was conducted in a single country, Sweden, and focused on one type of mammography device and one specific AI system. Also, the radiologists participating in the trial were described as moderately to highly experienced. Future research will need to explore how these findings translate to different healthcare settings, varying equipment, and diverse radiologist experience levels.[auntminnie+5]
Further studies will also investigate the long-term benefits and risks of using AI-supported mammography, including its cost-effectiveness in subsequent screening rounds. If these studies continue to show positive results, there could be a strong argument for widespread adoption of AI in breast cancer screening programs. This could lead to earlier detection of breast cancers, reduce radiologist workload, and ultimately improve patient care globally.[news-medical+6]
