Interval cancer, sensitivity, and specificity comparing AI-supported mammography screening with standard double reading without AI in the MASAI study: a randomised, controlled, non-inferiority, single-blinded, population-based, screening-accuracy trial

Gommers J, Hernström V, Josefsson V et al. Interval cancer, sensitivity, and specificity comparing AI-supported mammography screening with standard double reading without AI in the MASAI study: a randomised, controlled, non-inferiority, single-blinded, population-based, screening-accuracy trial. Lancet. 2026; 407(10527): 505-514. doi: 10.1016/S0140-6736(25)02464-X. 

The Mammography Screening with Artificial Intelligence (MASAI) trial recently reported its third analysis in The Lancet (January 2026)evaluating interval cancer rates in AI-supported mammography versus standard double-read mammography screening.

In this population-based randomised controlled trial, over 105,000 women in southwest Sweden were allocated 1:1 to AI-supported screening (with triage to single or double reading) or standard double reading. Interval cancers were defined as those diagnosed between screening rounds or within two years of a negative screen.

Interval cancer rates were non-inferior with AI-supported screening compared with standard double reading (1.55 vs 1.76 per 1000; ratio 0.88, 95% CI 0.65–1.18), confirming that AI does not increase missed cancers. Importantly, interval cancers in the AI group were less likely to be invasive, large (T2+), or biologically aggressive (non-luminal A), suggesting earlier detection.  Screening performance was improved with AI.  Sensitivity was significantly higher (80.5% vs 73.8%; p=0.031), while specificity was unchanged at 98.5% in both groups, showing no increase in false positives or unnecessary recallsThese findings build on earlier MASAI analyses demonstrating a 44% reduction in reading workload1 and improved cancer detection rates2. Together, they indicate that AI-supported mammography is clinically safe, more sensitive, and operationally efficient compared with standard double reading.  However, before widespread adoption in programmes such as the NHS Breast Screening Programme, further evidence is needed on cost-effectiveness, the risk of overdiagnosis, and long-term outcomes including mortality reduction.

Summary author: Miss Mhairi Mactier, ST5 General Surgery, NHS Greater Glasgow & Clyde

References: 

1. Lång K, Josefsson V, Larsson AM et al. 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. Lancet Oncol. 2023; 24(8): 936-944. doi: 10.1016/S1470-2045(23)00298-X. 
2. Hernström V, Josefsson V, Sartor H et al. Screening performance and characteristics of breast cancer detected in the Mammography Screening with Artificial Intelligence trial (MASAI): a randomised, controlled, parallel-group, non-inferiority, single-blinded, screening accuracy study. Lancet Digit Health. 2025; 7(3) :e175-e183. doi: 10.1016/S2589-7500(24)00267-X.