AI mammography screening cuts aggressive breast cancer diagnoses by 12% in landmark trial

AI-Supported Mammography Delivers Breakthrough Results in Breast Cancer Detection

The first randomized controlled trial evaluating AI in mammography screening has revealed that AI-supported mammography significantly improves breast cancer detection while reducing aggressive interval cancers. The MASAI trial, conducted across four sites in Sweden between April 2021 and December 2022, enrolled 105,934 women and compared AI-assisted screening against the European standard of double reading by radiologists 1

1

2

2

.

Kristina Lång from Lund University, the study's lead author, described the findings as "extremely thrilling," noting this marks the largest trial to date examining AI use in cancer screening 3

3

. Published in The Lancet, the results demonstrate that AI can enhance early detection of clinically relevant breast cancers and may improve outcomes for women participating in screening programs 2

2

.

How AI-Supported Mammography Works in Practice

The AI system, developed by biotech firm ScreenPoint Medical and trained on more than 200,000 mammography examinations from 10 countries, analyzes scans and assigns a risk score ranging from 1 to 10 based on visual patterns 1

1

5

5

. Scans receiving scores of 1 to 9 are assessed by one experienced radiologist, while those scoring 10—indicating cancer is most likely present—receive double reading by two radiologists. The AI also highlights suspicious areas to support radiologists in their analysis 3

3

.

This approach stands in contrast to standard European screening, where each mammogram is assessed by two radiologists without AI assistance. The trial participants underwent digital breast tomosynthesis, also known as pseudo-three-dimensional mammography 2

2

.

Interval Cancers Reduced by 12% with AI Support

Interval cancers—aggressive tumors that develop rapidly between screening appointments—pose a significant challenge in breast cancer screening. These cancers tend to be particularly aggressive and more likely to spread elsewhere in the body 1

1

. During the two-year follow-up period, the AI-supported group experienced 1.55 interval cancers per 1,000 women compared to 1.76 per 1,000 in the standard screening group, representing a 12% reduction 3

3

.

Jessie Gommers, first author and PhD student at Radboud University Medical Centre in the Netherlands, explained that interval cancer rates serve as an important measure reflecting both screening program effectiveness and test performance. "Interval cancers are often more aggressive and are associated with poorer outcomes than cancers detected at screening," Gommers noted 2

2

.

Cancer Detection Increases Without Rising False Positives

The trial found AI support increased cancer detection by 29% compared to standard screening—a remarkable improvement that didn't come at the cost of accuracy 2

2

. In the AI-supported group, 81% of cancer cases were detected at screening versus 74% in the control group, a 9% increase 3

3

4

4

.

Critically, false positives remained similar in both groups at 1.5% for the AI-supported arm and 1.4% for standard screening 2

2

3

3

. This indicates AI-supported screening didn't increase unnecessary recalls, which can cause anxiety and additional diagnostic procedures for women. The AI-supported group also showed 16% fewer invasive cancers, 21% fewer large tumors, and 27% fewer aggressive sub-type cancers 2

2

4

4

.

Radiologist Workload Reduced by 44% While Maintaining Human Oversight

Previously published interim results from the MASAI trial revealed that AI-supported screening reduced radiologist workload by 44%—a significant benefit given widespread staffing pressures in healthcare 2

2

5

5

. However, researchers emphasize that AI serves as a support tool for radiologists rather than a replacement. Lång stressed that women participating in screening want "a human in the loop," and she agrees with this approach 1

1

.

Final recall decisions remained entirely with radiologists, with AI helping them avoid overlooking suspicious regions 2

2

. Gommers noted that while the study doesn't support replacing healthcare professionals with AI, the results justify using AI to ease substantial pressure on radiologists, enabling these experts to focus on other clinical tasks and potentially shorten waiting times for patients 3

3

.

Implementation Requires Caution and Continuous Monitoring

Despite the promising results, Lång cautioned that introducing AI in healthcare must be done carefully. "Introducing AI in healthcare must be done cautiously, using tested AI tools and with continuous monitoring in place to ensure we have good data on how AI influences different regional and national screening programmes and how that might vary over time," she stated 4

4

.

The study has limitations that warrant further research. It was conducted in southwest Sweden using a single AI system, so additional research in diverse populations and with other AI tools is needed 2

2

. The trial didn't assess whether AI-supported screening performs better in certain ethnic groups, though ongoing trials in the UK will help address this question 1

1

. Fiona Gilbert at the University of Cambridge noted that countries need to see the impact on their own populations, where people are screened at different frequencies and represent different ethnicities 1

1

.

What This Means for Breast Cancer Screening Programs

With more than 2.3 million women diagnosed with breast cancer globally in 2022, according to the World Health Organization, and breast cancer remaining the leading cause of death in women aged 35 to 50, improvements in screening could save significant numbers of lives 4

4

5

5

.

Lång expects the AI approach to be rolled out across southwest Sweden within a few months based on these results 1

1

. However, it will likely take about five years for other countries to complete similar trials justifying broader implementation. Cost-effectiveness also needs establishment, though some estimates suggest AI assistance may be worth investing in if it cuts interval cancer rates by at least 5%—a threshold this trial exceeded 1

1

.

Simon Vincent, chief scientific officer at Breast Cancer Now, emphasized the significance: "This first trial underlines the huge potential of AI to support radiologists in breast cancer screening. Screening is a vital tool for early detection, and the sooner the disease is found, the better chance of successful treatment" 4

4

. Trials launched in the UK will be important in determining the safest and most effective way to use these tools to find more cancers early.