Release specialist capacity
Reduce avoidable review and interaction time so scarce breast-imaging expertise can be directed toward complex cases, consultation and patient care.
Platform
Turn breast imaging complexity into scalable clinical capacity
BreastScreening-AI is building a clinician-controlled platform for mammography, ultrasound and MRI. It brings image review, AI assistance, explainable interaction and workflow measurement into one product strategy so providers can pursue more capacity, more consistent decisions and a measurable return on responsible AI adoption.
Business value
Stronger Breast Imaging Service
The platform is designed around the outcomes that matter to provider executives, clinical leaders and imaging teams: capacity, quality, standardization, adoption, visibility and scalable integration.
Strengthen quality control
Add a structured second-reader layer that surfaces potential disagreement while keeping the radiologist responsible for the final assessment.
Connect the imaging pathway
Create a common product layer across MG, US and MRI rather than purchasing isolated AI experiences for each point in the pathway.
Accelerate team adoption
Adapt explanations and interaction to different experience levels, supporting a more practical rollout across mixed-seniority clinical teams.
Make value measurable
Track time, decisions, disagreement, usability and trust alongside model performance to build a customer-specific operational and economic case.
Scale with governance
Introduce capabilities through controlled validation gates, with traceability, clinician override and evidence requirements built into deployment planning.
Product platform
Assisted Breast Imaging
The product direction combines six capabilities that can be purchased, validated and introduced progressively according to customer priorities and the authorized intended use.
Multimodal workspace
A coordinated experience for mammography, ultrasound, MRI and relevant case information, while tracking evidence maturity separately for every modality.
AI second reading
Classification and lesion-localization outputs provide an additional perspective while preserving clinician review and override.
Explainable assistance
Visible findings, contextual arguments and communication adapted to clinician experience support informed rather than automatic reliance.
Workflow measurement
Decision changes, time, usability, workload and interaction can be evaluated alongside model performance.
Traceable decisions
The intended workflow distinguishes the initial clinician assessment, AI recommendation and final clinician decision.
Progressive integration
Technical integration, human factors, clinical validation and economic evaluation are treated as separate deployment gates.
Why multimodality
Pathway
No modality answers every clinical question. A commercially useful platform must support how providers move from population screening to targeted characterization, risk-based assessment, staging and treatment planning. The detailed operational sequence will be developed further in the Workflow page.
Mammography: the screening foundation
Low-dose mammography is the established population-screening entry point and can reveal early changes, including calcifications. Its scale makes reading efficiency and consistent assessment commercially important, but dense tissue can obscure findings.
Ultrasound: targeted characterization
Ultrasound helps characterize a palpable or imaging-detected abnormality as solid, fluid-filled or mixed, adds real-time soft-tissue information and can guide biopsy. It complements rather than replaces mammography.
MRI: high-sensitivity problem solving
MRI provides detailed contrast-enhanced assessment for selected high-risk screening, disease-extent evaluation and further investigation of abnormalities. It adds information that may not be visible on MG or US, with higher cost and access constraints.
Clinical context: WHO, mammography, ultrasound and MRI. Modality selection remains specific to patient characteristics, risk factors, and clinical indication.
Evidence-backed traction
Operational and Quality Value
Controlled studies provide a quantitative foundation for customer validation. These figures are study outcomes, not guaranteed customer results or routine-care performance claims.
Clinicians across nine institutions
A peer-reviewed comparison of clinician-only and clinician-AI scenarios provides the platform's principal human-AI evidence base.
Read the studyPositive clinician response
Clinicians reported positive expectations and perceptive satisfaction, supporting the platform's adoption proposition.
Read the studyLower mean task time
Published means decreased by 69 seconds, from 377 to 308 seconds per experimental case, indicating a potential capacity lever.
Inspect the publicationLower false-positive proportion
False-positive classifications decreased from 54% to 27% in the controlled study.
Review the evidence contextLower false-negative proportion
False-negative classifications decreased from 6% to 2% under the experimental clinician-AI condition.
Review the evidence contextFaster personalized interaction
Mean task time decreased from 166.12 to 124.02 seconds with assertiveness-based communication; p = 0.005 and reported r = 0.49.
Read the CHI studyCommercial interpretation: the results support further evaluation of productivity, quality and adoption value. They do not yet establish customer ROI, fewer biopsies, improved cancer detection or realized cost savings.
Modeled market impact
What a 69-second capacity gain could mean at scale
The scenarios below apply the published experimental mean-time difference to annual review volumes. They illustrate the scale of addressable operational value; they are not a total-addressable-market estimate, forecast, price recommendation or claim of realized savings.
Per 1 million annual reviews
A 69-second reduction applied to one million reviews equals approximately 12 full-time-equivalent years of capacity, assuming 1,600 productive hours per FTE.
At 10 million annual reviews
The same scenario equals approximately 120 FTE-years of potential capacity across a large screening system or multi-market provider network.
Illustrative annual capacity value
At 10 million reviews and an assumed fully loaded specialist-hour value of €75–€150, released capacity would have this modeled gross value before implementation costs.
At 50 million annual reviews
A broader cross-system scenario equals approximately 599 FTE-years of potential capacity. Actual eligible volumes and workflow effects would require country-level validation.
Broader-system capacity value
This applies the same €75–€150 hourly assumption to 50 million reviews. It represents capacity value, not cash savings, revenue or net economic benefit.
Global diagnoses in 2022
WHO estimates 2.3 million women were diagnosed and 670,000 died from breast cancer in 2022, underscoring the global need for timely, scalable breast-care pathways.
Review WHO contextTransparent model
Capacity hours = annual reviews × 69 seconds ÷ 3,600. FTE-years = capacity hours ÷ 1,600. Gross capacity value = capacity hours × assumed €75–€150 per specialist hour.
The model excludes software, integration, infrastructure, training, governance and change-management costs. It also excludes downstream benefits from fewer false positives or false negatives because the controlled study proportions cannot yet be translated responsibly into avoided recalls, biopsies, delayed diagnoses or treatment costs.
Go-to-market fit
Value Proposition
The platform is relevant where breast imaging quality, specialist capacity, multimodal coordination and responsible AI adoption are strategic priorities.
Hospitals and imaging centers
Potential value lies in workflow capacity, second-reading support, quality measurement and a structured route for evaluating AI before wider adoption.
Multi-site healthcare groups
A common platform can support standardized evaluation, cross-site performance monitoring and consistent governance across different teams.
PACS and imaging partners
The multimodal and explainable-interaction strategy offers a differentiated application layer for approved imaging ecosystems and integration programs.
Clinical research programs
The platform can support prospective validation, reader studies, human-factors testing and health-economic evidence generation.
Adoption path
Convert product evidence into customer value
Healthcare customers need more than an algorithm. Adoption depends on integration, clinical validation, governance, training and a value case grounded in local workflow and cost data.
Current assets
- Peer-reviewed clinician studies with cohorts of 31, 45 and 52 participants.
- Research across mammography, ultrasound and MRI.
- Protected multimodal and clinician-adaptive intellectual property.
- Project-level TRL 5 progressing toward TRL 6.
Required before scaled adoption
- Locked intended use and product configuration.
- Independent multicenter and prospective clinical validation.
- Quality management, regulatory authorization and post-market controls.
- Customer-specific budget impact and return-on-investment evidence.
Commercial position
BreastScreening-AI remains research software under progressive product, human-factors and clinical evaluation. Commercial discussions should focus on validation partnerships, integration planning and evidence generation rather than claims of current clinical authorization.
Discuss a validation partnershipHuman firstProduct principle
The platform is designed to strengthen clinical judgement. A qualified clinician retains responsibility for interpretation and patient management.
Evaluate the platform in your clinical environment
Explore the intended workflow or contact the team about validation, integration and evidence partnerships.
Explore workflow Contact the team