<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=1683513&amp;fmt=gif">
Skip to content
Contact us

VAI-B: AI Validation Framework for Breast Cancer Screening: Standardizing Performance Benchmarks

Karolinska Institut-KTH-Linköping University-Lund University

Industry

Academia

Challenge

Multiple vendors are currently offering artificial intelligence (AI) computer-aided systems for triage, detection, diagnosis, and risk prediction of breast cancer based on screening mammography. There is an imminent need to establish validation platforms that enable fair and transparent testing of these systems.

Results

The VAI-B project now has a growing body of real-world evidence to showcase its medical AI, with more observational, feasibility, and qualitative studies already underway.

Key Product

Research

medtecharena.se/vaib/
+100.000
Mammographies
44.419
Patients
20
Hospitals
4
AI vendor systems
vaib

VAI-B

VAI-B involves many Swedish Institutions like: 

  • Karolinska Institute is one of the world’s foremost medical universities, located in Stockholm, Sweden. It is internationally recognized for its cutting-edge research and high-quality education in medicine and health sciences. 
  • Lund University is a top Swedish research institution known for its high-impact, interdisciplinary work across fields like medicine, engineering, and environmental science.
  • Linköping University
  • KTH
  • Medtech4Health

Barriers to Trustworthy AI in Breast Cancer Screening

The rapid emergence of AI tools for breast cancer screening — offering triage, detection, diagnosis, and risk prediction — has outpaced the development of standardized validation frameworks. Each vendor provides proprietary performance claims, but there is no universal, fair, or transparent way to externally validate these AI systems using independent datasets. The major challenge lies in:

  • Ensuring consistent benchmarking of AI systems across multiple vendors.

  • Maintaining data privacy when working with sensitive radiological and clinical information.

  • Supporting scalable evaluations that reflect real-world clinical conditions across regions and hospitals.

How Our AI Validation Platform Became the Trusted Choice for Clinical-Grade Breast Cancer Screening

Collective Minds Research came as the only available solution  for the project porpuse.

Transforming AI Evaluation: Faster Development, Safer Clinical Integration

VAIB stands out with exceptional data quality and diversity, addressing the critical need for representative datasets. This aligns perfectly with today’s customer and regulatory authorities demands to compare different algorithms side by side and on representative datasets, enabling informed decisions. 

The platform is now equipped to:

Support faster AI development cycles by providing objective performance data to vendors.

Enable safer adoption of AI systems in hospitals through evidence-based evaluations tailored to local imaging data.