Publications on AI-assisted radiology and mammography.
A curated reading list of peer-reviewed work behind our approach, plus a live feed of the latest matching publications from PubMed.
Publications on AI in radiology and mammography
A curated reading list of peer-reviewed work on AI-assisted breast imaging, plus a live feed of the most recent matching publications on PubMed.
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
Lång K, Josefsson V, Larsson A-M, et al. · The Lancet OncologyRandomised, controlled trial in Swedish population screening showing AI-supported single reading was non-inferior to standard double reading for cancer detection while reducing radiologist screen-reading workload by ~44%.
International evaluation of an AI system for breast cancer screening
McKinney SM, Sieniek M, Godbole V, et al. · NatureGoogle Health / DeepMind model evaluated on UK and US screening datasets reduced false positives and false negatives versus radiologists, and generalised across populations.
Stand-alone artificial intelligence for breast cancer detection in mammography: comparison with 101 radiologists
Rodriguez-Ruiz A, Lång K, Gubern-Merida A, et al. · Journal of the National Cancer InstituteStand-alone AI achieved cancer detection performance comparable to the average of 101 radiologists across nine reader studies.
Prospective implementation of AI-assisted screen reading to improve early detection of breast cancer
Dembrower K, Crippa A, Colón E, et al. · Nature MedicineProspective implementation in a Swedish screening programme showed AI triage increased cancer detection rate while flagging cases for additional review.
Detection of breast cancer with mammography: effect of an artificial intelligence support system
Rodríguez-Ruiz A, Krupinski E, Mordang J-J, et al. · RadiologyAdding an AI decision-support system to radiologist reading improved cancer detection performance (AUC) without increasing reading time.
Robust breast cancer detection in mammography and digital breast tomosynthesis using an annotation-efficient deep learning approach
Lotter W, Diab AR, Haslam B, et al. · Nature MedicineDeep learning model trained with weak supervision generalised across mammography and digital breast tomosynthesis (DBT), matching expert readers across multiple sites.
External evaluation of 3 commercial artificial intelligence algorithms for independent assessment of screening mammograms
Salim M, Wåhlin E, Dembrower K, et al. · JAMA OncologyIndependent evaluation of three commercial AI systems on a Swedish screening cohort; the best-performing system matched the average radiologist.
Mammographic breast density assessment using deep learning: clinical implementation
Lehman CD, Yala A, Schuster T, et al. · RadiologyDeep learning model for BI-RADS density assessment deployed clinically at a large academic centre, with high agreement with radiologists and consistency over time.
A deep learning mammography-based model for improved breast cancer risk prediction
Yala A, Lehman C, Schuster T, et al. · RadiologyMirai precursor: deep learning model that uses mammograms to predict 5-year breast cancer risk, outperforming the Tyrer-Cuzick model.
Toward robust mammography-based models for breast cancer risk
Yala A, Mikhael PG, Strand F, et al. · Science Translational MedicineMirai: locally trained risk model validated across multiple international sites for 1- to 5-year breast cancer risk prediction.
Multi-Institutional Validation of a Mammography-Based Breast Cancer Risk Model
Yala A, Mikhael PG, Lehman C, et al. · Journal of Clinical OncologyMulti-site validation of the Mirai risk model demonstrating consistent discrimination across diverse populations.
Screening performance of abbreviated versus full-protocol breast MRI in women at increased risk of breast cancer: a systematic review and meta-analysis
Geuzinge HA, Bakker MF, Heijnsdijk EAM, et al. · European RadiologyMeta-analysis comparing abbreviated and full breast MRI protocols for screening, relevant to AI-assisted triage of supplemental imaging.
Supplemental MRI screening for women with extremely dense breast tissue
Bakker MF, de Lange SV, Pijnappel RM, et al. (DENSE Trial) · New England Journal of MedicineDENSE randomised trial: supplemental MRI in women with extremely dense breasts reduced interval cancers, motivating AI-driven density and risk stratification.
Artificial intelligence in radiology
Hosny A, Parmar C, Quackenbush J, et al. · Nature Reviews CancerFoundational review of AI applications in radiology covering detection, segmentation, classification, and clinical translation.
FDA-cleared artificial intelligence and machine learning-based medical devices and their 510(k) predicate networks
Muehlematter UJ, Daniore P, Vokinger KN · The Lancet Digital HealthAnalysis of the regulatory landscape for AI/ML-enabled medical devices cleared by the FDA, including imaging-focused devices.
Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices
U.S. Food & Drug Administration · FDA.govOfficial, continually updated list of FDA-authorised AI/ML-enabled medical devices, including a large share of radiology applications.
Effect of artificial intelligence-based triaging of breast cancer screening mammograms on cancer detection and radiologist workload: a retrospective simulation study
Raya-Povedano JL, Romero-Martín S, Elías-Cabot E, et al. · RadiologySimulation of AI-based mammography triage showed potential to safely reduce radiologist workload by ~70% while maintaining cancer detection.
Frequently asked questions
No. Urgent American Radiology Services builds AI-assisted workflows that prioritize, summarize, and surface findings, but radiologists make the final interpretation. Every AI output is reviewable, overridable, and audit-logged.
Urgent American Radiology Services is a workflow and orchestration layer. Clinical AI components used inside Urgent American Radiology Services-orchestrated workflows can include FDA-authorized devices. Urgent American Radiology Services does not claim FDA clearance for the orchestration platform itself unless a specific component clearance is documented.
Initial focus is screening mammography support, chest X-ray triage and prioritization, and reporting automation. Additional modalities are evaluated based on partner-AI clinical evidence and integration fit.
Urgent American Radiology Services integrates over DICOM, HL7 v2, FHIR, and standard PACS/RIS interfaces. We work alongside your existing systems rather than replacing them, and we map structured outputs back to your reporting tools.
Clinical workflows are designed for HIPAA-aligned operation: encryption in transit and at rest, role-based access control, audit logging, data minimization, and Business Associate Agreement readiness. The marketing site never collects PHI.
We start with a workflow assessment, define eligibility rules and integration points, run a configuration and validation phase, and then go live with monitoring and quality analytics. Most deployments are phased rather than big-bang.
Every AI output is treated as decision support. Radiologists confirm, edit, or override findings, and overrides are captured for QA and continuous improvement. Workflow steps and escalation paths are configurable per site.
We track turnaround time, worklist throughput, recall and follow-up adherence, override rates, and quality-program metrics. Centers receive dashboards and can export their own data.
Urgent American Radiology Services partners with independent imaging centers and community-health organizations to expand access to screening mammography and follow-up navigation in underserved U.S. regions.
Request a workflow assessment. We will review your modalities, volume, integrations, and goals, and propose a deployment plan tailored to your operation.