Researchers running the MASAI trial at four Swedish hospitals tested whether an AI system could stand in for the second reader. Nearly 80,000 women were randomly assigned to AI-supported screening or the standard double reading.
The AI-supported arm found 20 percent more cancers than standard double reading, 41 additional cases, without raising the false positive rate. It also cut the radiologists' total reading workload by roughly 44 percent, freeing up months of specialist time.
This was a safety analysis, not the final word. The trial's real target, whether the extra cancers caught early actually lower the rate of interval cancers that appear between scheduled screenings, will not be known until the full 100,000-participant group has been followed for two years. Detecting more is not automatically the same as saving more lives.
Read the original study
Lång et al., The Lancet Oncology, 2023 · doi.org/10.1016/S1470-2045(23)00298-X
Same mammograms, a fifth more cancers found.
Source
- Lång, K., Josefsson, V., Larsson, A.-M., Larsson, S., Högberg, C., Sartor, H., Hofvind, S., Andersson, I., & Rosso, A. (2023). Artificial intelligence-supported screen reading versus standard double reading in the Mammography Screening with Artificial Intelligence trial (MASAI): a clinical safety analysis of a randomised, controlled, non-inferiority, single-blinded, screening accuracy study. The Lancet Oncology, 24(8), 936–944. https://doi.org/10.1016/S1470-2045(23)00298-X