Particle.news
Download on the App Store

AI-Backed Mammograms Cut Later Breast Cancer Diagnoses by 12% in Large Swedish Trial

Researchers say validated tools with human oversight could ease radiologist workload, though adoption should be cautious and supported by monitoring and cost-effectiveness studies.

Overview

  • Published in The Lancet, the MASAI randomized trial enrolled more than 100,000 women in Sweden (2021–2022) to compare AI-supported screening with standard double reading.
  • Over two years, interval cancers occurred at 1.55 per 1,000 women with AI support versus 1.76 per 1,000 with standard reading, a 12% reduction.
  • In the AI arm, 81% of cancers were found at the screening visit compared with 74% in the control group, with 16% fewer invasive, 21% fewer large, and 27% fewer aggressive sub-type cancers subsequently diagnosed.
  • Specificity was maintained with similar false-positive rates (1.5% in the AI group vs 1.4% in controls), and radiologists retained final responsibility as the system triaged low-risk cases to single reading.
  • Earlier MASAI analyses reported a 44% cut in reading workload; authors and commentators call for careful rollout, replication in other settings (including large UK trials), and full economic evaluation, with some Swedish regions preparing near-term implementation.