Artificial Intelligence (AI) has significantly improved the detection of breast cancer during routine mammography, according to the results of a landmark randomised controlled trial published this week.

The study — the first completed trial of its kind — demonstrated that AI-assisted screening helps radiologists identify more breast cancer cases, potentially easing workforce pressures in health systems and improving patient outcomes.

World-First Trial Highlights

Involving more than 100,000 women across Sweden between April 2021 and December 2022, the trial compared traditional mammogram readings with

• AI-supported analysis. Participants were randomly assigned to:
AI-assisted mammography, where a radiologist reviewed scans with support from an AI system, and

• Standard practice, where two radiologists independently read each scan.

The results showed that:

• The AI-assisted group detected 9% more breast cancer cases during screening than the control group.

• Over the following two years, there was a 12% lower rate of interval cancers — cancers diagnosed between scheduled screenings, which are often more aggressive.

• The rate of false positives — scans incorrectly flagged as suspicious — remained similar between both groups.

Experts say earlier detection of breast cancer is critical for improving treatment outcomes and reducing mortality, particularly in populations where access to specialised radiologists is limited.

Implications for Global Health

Lead author Dr Kristina Lang of Sweden’s Lund University emphasised that the findings support expanded use of AI-assisted mammography in screening programmes worldwide.

“Widely rolling out AI-supported mammography can help reduce workload pressures among radiologists and help detect more cancers at an earlier stage,” Dr Lang said, urging that implementation be cautious and monitored continuously to ensure performance remains robust across diverse populations.

The study was published in the prestigious The Lancet medical journal, marking a milestone in the integration of AI into routine clinical practice.

Expert Reactions and Cautions

Medical professionals welcomed the trial’s results but stressed that AI should support — not replace — human radiologists.
Jean-Philippe Masson, head of the French National Federation of Radiologists, noted that while AI adds valuable insight, the radiologist’s experience remains essential to confirm diagnoses.

Background: Breast Cancer and Screening Challenges

Breast cancer remains one of the most common cancers worldwide, with millions of new diagnoses each year. Early detection through screening programmes helps reduce late-stage presentations and improve survival rates.

Moving Forward

As healthcare systems grapple with workforce shortages and growing demand for diagnostic services, the trial’s findings could encourage adoption of AI tools that bolster capacity without compromising quality.

For further details and access to the full study results, visit the The Lancet publication and related medical journal coverage.