Berlin’s Vara Gets First CE Mark for Autonomous Mammo AI

Berlin-based Vara has received the world’s first CE certification for an autonomous AI-enabled system that can screen mammograms without radiologist review. Announced on 4 September 2026, the Class IIb designation authorises normal-classified scans to be reported by AI alone, while any other result must still be read by at least one radiologist, marking a substantial regulatory milestone for population-scale AI diagnostics in Europe.

Standard European breast cancer screening protocol involves two radiologists interpreting each mammogram, creating substantial clinical and operational demand. Vara’s autonomous triage feature enables the AI to independently report scans it classifies as normal. To secure the CE mark, Vara developed ATMON (Autonomous-Triage Monitoring), a proprietary supervision system that meets the continuous human oversight requirement for high-risk AI systems under the European Union AI Act. ATMON monitors mammography hardware changes, system health and daily performance, automatically reverting a site to full radiologist reading if values fall outside preset limits.

Vara CEO and co-founder Jonas Muff positioned ATMON as essential for maintaining AI model performance once systems operate outside controlled study conditions. Vara’s clinical evidence base includes a Nature Medicine study published in 2024 comparing standard double reading by radiologists with AI-supported double reading across more than 460,000 women, which found a higher breast cancer detection rate among AI-supported readings. More than half of Germany’s organised breast screening programme currently runs on Vara, processing more than 250,000 screenings per month. ATMON is also being made available to existing and new customers even without switching to fully autonomous triage.

The commercial signal is a substantial validation for autonomous AI in high-stakes diagnostic imaging, positioning Vara as a regulatory pathfinder under the EU AI Act. Population-scale AI diagnostics is emerging as a significant new competitive category. Expect breast AI competitors Densitas, ScreenPoint, Kheiron, DeepHealth (RadNet), Volpara Health, iCAD, Therapixel and Lunit to sharpen their positioning as autonomous AI regulatory pathways mature.

For the latest updates and in-depth insights into the world of Medical Devices, including breakthrough treatments, industry trends, and regulatory news, contact Albert Baines today!