{"entity":{"id":"paper-screentrustcad-dembrower-lancet-digit-health-2023","kind":"paper","name":"Artificial intelligence for breast cancer detection in screening mammography in Sweden: a prospective, population-based, paired-reader, non-inferiority study (ScreenTrustCAD)","aka":[],"tldr":"Among 55,581 women screened in Stockholm, one radiologist working with Lunit's software found 261 cancers against 250 for two radiologists, and the software on its own found 246, so the AI could stand in for a second reader without missing cancers.","summary":"Primary publication of ScreenTrustCAD (NCT04778670), a prospective, population-based, paired-reader, non-inferiority study at Capio Sankt Göran Hospital, Stockholm. Consecutive women aged 40 to 74 without breast implants in population screening were included. The primary outcome was screen-detected breast cancer within three months of mammography; the primary analysis assessed non-inferiority (margin 0.15 relative reduction in diagnoses) of double reading by one radiologist plus AI against standard double reading by two radiologists, with single reading by AI and triple reading by two radiologists plus AI as further comparisons.\n\nFrom 1 April 2021 to 9 June 2022, 58,344 women underwent screening and 55,581 were included. 269 women (0.5 percent) had screen-detected cancer on an initial positive read. One radiologist plus AI was non-inferior to two radiologists (261 versus 250 detected cases; relative proportion 1.04, 95 percent CI 1.00 to 1.09). AI alone (246 versus 250; 0.98, 0.93 to 1.04) and two radiologists plus AI (269 versus 250; 1.08, 1.04 to 1.11) were also non-inferior. The authors conclude that replacing one radiologist with AI gave a 4 percent higher non-inferior detection rate and suggest controlled implementation with risk management and real-world follow-up. Funded by the Swedish Research Council, the Swedish Cancer Society, Region Stockholm and Lunit.","asOf":"2026-09-24","links":[{"label":"Lancet Digital Health 2023","url":"https://doi.org/10.1016/S2589-7500(23)00153-X"},{"label":"PubMed","url":"https://pubmed.ncbi.nlm.nih.gov/37690911/"},{"label":"ClinicalTrials.gov NCT04778670","url":"https://clinicaltrials.gov/study/NCT04778670"}],"tags":[],"related":[],"cancers":["breast-hr-positive","tnbc"],"sections":[],"technologies":["mammography","radiology-ai-screening"],"targets":[],"drugs":["lunit-insight-mmg"],"companies":["lunit"],"institutions":["karolinska"],"pathways":[],"terms":[],"trials":["screentrustcad"],"people":[],"bottlenecks":[],"keyPapers":[],"journals":["lancet-digital-health"],"dependsOn":[],"notes":[],"journal":"The Lancet Digital Health","year":2023,"doi":"10.1016/S2589-7500(23)00153-X","pmid":"37690911","authors":"Dembrower K, Crippa A, Colón E, et al.","paperType":"observational","findings":["One radiologist plus AI detected 261 cancers vs 250 for two radiologists (relative proportion 1.04, 95 percent CI 1.00 to 1.09); non-inferior.","AI alone detected 246 (relative proportion 0.98, 0.93 to 1.04) and two radiologists plus AI 269 (1.08, 1.04 to 1.11); both non-inferior.","55,581 of 58,344 screened women were included; 269 (0.5 percent) had screen-detected cancer."],"whatItMeans":"Screening programmes short of radiologists can read this as prospective evidence that a well-validated AI reader can take the second reader's seat without lowering cancer detection. It is a paired-reader study in one hospital, not a randomised trial, so MASAI and real-world follow-up carry the argument further.","caveats":["Paired-reader design in a single hospital; the same women were read under every strategy rather than randomised between them.","The lower confidence bound of the primary relative proportion is 1.00, so the result sits at the edge of showing more detection rather than simply non-inferiority.","Interval cancers and overdiagnosis were not part of the primary analysis; Lunit part-funded the study."],"changedPractice":false,"participants":55581},"route":"/key-papers/paper-screentrustcad-dembrower-lancet-digit-health-2023/","neighbours":{"cancer":[{"id":"breast-hr-positive","kind":"cancer","name":"HR-positive / HER2-negative breast cancer","route":"/cancers/breast-hr-positive/"},{"id":"tnbc","kind":"cancer","name":"Triple-negative breast cancer (TNBC)","route":"/cancers/tnbc/"}],"technology":[{"id":"radiology-ai-screening","kind":"technology","name":"AI in radiology","route":"/technologies/radiology-ai-screening/"},{"id":"mammography","kind":"technology","name":"Mammography & tomosynthesis","route":"/technologies/mammography/"}],"drug":[{"id":"lunit-insight-mmg","kind":"drug","name":"Lunit INSIGHT MMG","route":"/drugs/lunit-insight-mmg/"}],"company":[{"id":"lunit","kind":"company","name":"Lunit","route":"/companies/lunit/"}],"institution":[{"id":"karolinska","kind":"institution","name":"Karolinska University Hospital","route":"/institutions/karolinska/"}],"trial":[{"id":"screentrustcad","kind":"trial","name":"ScreenTrustCAD","route":"/trials/screentrustcad/"}],"journal":[{"id":"lancet-digital-health","kind":"journal","name":"The Lancet Digital Health","route":"/journals/lancet-digital-health/"}]}}