Evidence
Multistage AI-Driven Workflow Improves General Radiologist Screening Mammography Performance to the Level of Fellowship-Trained Breast Imagers: Real-World Evidence in >500,000 Patients
RSNA: Podium Presentation
McCabe et al. 2025
Interpretive performance in screening mammograms (SMGs) varies widely. An AI-driven workflow was deployed at scale, and performance before and after implementation was compared between radiologists with and without fellowship training in breast or women’s imaging. The Multistage AI-Driven workflow significantly improved CDR and PPV1 for General radiologists to levels comparable with Fellowship-trained radiologists.
Large-Scale Deployment of a Multistage AI-Driven Workflow Increases Detection of Deadlier Breast Cancers
RSNA: Poster Presentation
Louis et al. 2025
This study assesses the cancer subtypes detected in a large-scale deployment of a multistage AI-driven workflow compared to the standard of care (SOC). Employing the multistage AI-driven workflow significantly improved clinical outcomes in terms of CDR in a cohort of over 2400 cancers, with the majority of cancers detected being clinically relevant and no corresponding increase in the proportion of DCIS, showing benefits for screening mammography.
Equitable Impact of an AI-driven Breast Cancer Screening Workflow in Real-World US-Wide Deployment
Nature Health
Louis et al. Nov 2025
Artificial intelligence shows promising results for improving early breast cancer detection and overall screening outcomes. Here the AI-Supported Safeguard Review Evaluation (ASSURE) study evaluates an AI workflow on digital breast tomosynthesis exams from women across four states to optimize early cancer detection. Implementation of the AI workflow improved screening effectiveness with equitable benefits.
AI Driven Safeguard Review Process Helps Detect Aggressive Breast Cancers
RSNA: Poster Presentation
Kim et al. 2024
An AI-driven safeguard review process was implemented prospectively, and its custom-built AI algorithm was used to flag the most suspicious screening DBT exams that had not been recalled by the initial interpreting radiologist. An expert breast imaging specialist performed a second, safeguard review of the 2,296 flagged exams. This resulted in the detection of 41 additional cancers, mostly invasive, 22.0% of which were deemed aggressive.
Impact of a Categorical AI System for Digital Breast Tomosynthesis on Breast Cancer Interpretation by Both General Radiologists and Breast Imaging Specialists
Radiology Artificial Intelligence
Kim et al. Feb 2024
The performance of 18 general radiologists and breast imaging specialists (9 generalists, 9 specialists) was evaluated with and without the aid of a custom-built categorical AI system. The categorical AI system helped improve overall radiologist interpretation performance of DBT screening mammograms for both general radiologists and breast imaging specialists and across various patient subgroups and breast cancer characteristics.
Mammography AI in 147 Clinics Results in Increased Cancer Detection Rate
Leeann Louis, PhD
AI software leverages information in DBT to help identify cancers with limited or no visibility on FFDM
Brittaney Everitt, MS
An AI-driven peer review quality-assurance process to improve cancer detection: Real world experience in a community practice
Mireille Aujero, MD
Categorical artificial intelligence helps general radiologists and breast imaging specialists improve cancer detection performance on DBT mammograms for women of all races
Jiye G. Kim, PhD
Radiologists using categorical AI for screening mammography improve more than double reading alone
Hyunkwang Lee, PhD
Real world evidence supports robustness of AI categories for screening mammography
Bryan Haslam, PhD
Robust Breast Cancer Detection in Mammography and Digital Breast Tomosynthesis Using an Annotation-Efficient Deep Learning Approach
Nature Medicine
Lotter et al. Jan 2021
DeepHealth’s deep-learning algorithm achieves state of the art performance in mammogram classification. The AI model was compared to 5 readers in a reader study of 131 index cancers and 154 confirmed negatives. The model showed robust and generalizable performance, reporting an Area Under the Curve (AUC) of 0.945 and outperforming all radiologists with a sensitivity 14% higher than the average radiologist sensitivity.