← Signals
Recordedrangescience

A randomised screening trial reports what AI-supported mammography actually changes

Sweden's MASAI trial randomised over 100,000 women between AI-supported single reading and standard double reading. A clinical safety analysis appeared in The Lancet Oncology in 2023; the cancer-detection analysis reported a 29% higher screen-detected cancer detection rate and a 44.2% reduction in screen-reading workload — 48,444 fewer readings — with seven more false positives between the arms; and the full results in The Lancet reported a non-inferior interval-cancer rate, higher sensitivity, unchanged specificity, and fewer interval cancers with unfavourable characteristics. This is the strongest human-consequence evidence in the pack, and it is worth noting how much it cost to obtain: a population-scale randomised trial running for years, in one country, for one modality.

Observed — measured or witnessed, with the observation cited.

Subjects

Evidence

Partly verified — Core facts are sourced; some optional detail is deliberately absent.