Skip to main content

Vara's AI Gets Green Light for Autonomous Breast Cancer Screening

In a world first, Berlin-based healthcare AI company Vara has received CE certification under the EU's Medical Device Regulation (MDR) for its breast cancer screening AI. This isn't just another AI approval—it's the first time an AI system has been cleared to autonomously triage mammograms in breast cancer screening programs.

What does that mean in practice? The AI can now directly exclude clearly normal mammograms from radiologists' review. Only cases that aren't clearly normal will still require a human eye. This model is designed specifically for organized population screening, not for symptomatic patients or diagnostic exams.

Why this matters

European screening programs typically have two radiologists read each mammogram. But here's the catch: about 97% of those images are normal. Meanwhile, the number of radiologists is shrinking while the screening population keeps growing. It's a classic supply-demand crunch, and Vara's AI aims to ease the pressure.

But here's the reassuring part: the certification isn't just about the AI model itself. It's tied to an accompanying ATMON autonomous triage monitoring system. This system continuously tracks equipment performance, system status, cancer detection rates, and recall rates. If any metric goes beyond predefined thresholds, it automatically switches back to full radiologist review. So there's a safety net built in.

The evidence behind the approval

The certification leans heavily on the PRAIM study, a large-scale trial conducted between July 2021 and February 2023 across 12 sites in Germany. It involved 463,094 women—a staggering number. The results were promising: the cancer detection rate in the AI-assisted group was 6.7 per 1,000, compared to 5.7 per 1,000 in the control group. That's a relative increase of 17.6%.

However, it's worth noting that the study tested AI-assisted double reading, not the autonomous triage model that just got certified. So while the results are encouraging, they don't directly validate this specific workflow. Still, Vara is confident that the monitoring system and the underlying AI are up to the task.

What this means for the future

This certification could be a game-changer for breast cancer screening programs struggling with workforce shortages. By letting AI handle the obvious normals, radiologists can focus their expertise where it's needed most. And with the built-in safety checks, there's a layer of accountability that should ease concerns about AI going rogue.

Of course, this is just the beginning. As AI continues to integrate into healthcare, we'll likely see more such approvals. But for now, Vara is leading the charge, and the rest of the world is watching closely.

Key Points

  • Vara's AI is the first to receive CE certification for autonomous triage in breast cancer screening.
  • The AI can exclude clearly normal mammograms from radiologist review, while other cases still require human verification.
  • The certification is supported by the PRAIM study involving over 463,000 women, showing a 17.6% relative increase in cancer detection.
  • The ATMON monitoring system ensures safety by automatically reverting to full radiologist review if thresholds are exceeded.