HOPPR Releases 2D Mammography Foundation Model to Accelerate AI Innovation in Breast Imaging

New Vision Transformer-based (ViT) model supports fine-tuning across multiple imaging tasks, reinforcing the importance of innovation during Breast Cancer Awareness Month.

CHICAGO, Oct. 21, 2025 /PRNewswire/ -- HOPPR, a secure development platform for AI medical imaging, today announced the release of its HOPPR™EB 2D Mammography Foundation Model, designed to help developers accelerate building AI applications for breast imaging.

The HOPPR™EB 2D Mammography Foundation Model demonstrates proven fine-tuning performance across binary and multi-class classification applications such as cancer detection, breast density classification, and device identification. The model's design also enables adaptation to other downstream tasks within mammography and related imaging domains. This model offers a powerful starting point for teams looking to build AI solutions for research and operational workflows in mammography.

"We've been exploring ways to accelerate readiness of AI in breast imaging," said Sham Sokka, Chief Operating and Technology Officer at DeepHealth. "HOPPR's mammography foundation model should give us a flexible infrastructure to adapt it to our workflow needs. It's a meaningful step forward in accelerating development, readiness, and real-world use."

The announcement coincides with Breast Cancer Awareness Month, underscoring the importance of expanding access to AI development infrastructure that supports innovation in early detection and breast health.

The HOPPR™ EB 2D Mammography Foundation Model was trained through a multi-stage process that includes self-supervised learning and expert distillation. When fine-tuned for classification tasks, it is delivered with parameter-efficient LoRA adapters and a lightweight MLP head, returning structured predictions in standardized JSON format for easy integration into development pipelines.

Key Features:

• Supports fine-tuning for multiple use cases: Enables fine-tuning for a range of tasks from study-level and breast-level classification to image-level feature detection.
• High performance: Internal testing shows ROC-AUC of 0.92 for cancer detection, 0.94 for breast density classification, and 0.99 for pacemaker detection.
• Developer-friendly infrastructure: Secure, HIPAA-compliant fine-tuning and inference via API, usage-based billing, and traceability.

Unlike prebuilt AI applications, HOPPR's approach gives developers control over model fine-tuning and inference through a single secure API using their own labeled DICOM data. Built under a quality management system aligned with ISO 13485, SOC 2, and HITRUST standards, the platform provides secure data handling, workflow orchestration, and structured model outputs in one secure environment.

"Foundation models are changing the pace of innovation in imaging AI, but only if they're accessible, adaptable, and built with real-world deployment in mind," said Dr. Khan Siddiqui, CEO and Co-founder of HOPPR. "With this release, we're giving developers the infrastructure to move quickly with transparency, traceability, and control from day one."

The release expands HOPPR's growing model library, featuring the HOPPR™ MC Chest Radiography Foundation Model, API-based fine-tuning, and usage-based billing within its secure AI development platform.

Join the Webinar: Breaking the Mold: Why Imaging AI Needs a New Playbook

HOPPR and DeepHealth will host a live webinar exploring how foundation models are reshaping imaging AI at enterprise scale. This 35-minute webinar will include a live demo, insights from AI and clinical leaders, and a Q&A session.

Date: Thursday, October 23, 2025
Time: 11:00 a.m. EDT / 8:00 a.m. PDT / 5:00 p.m. CEST
Speakers:

• Khan Siddiqui, MD, Founder and CEO, HOPPR
• Niccolò Stefani, Business Leader Population Health & Clinical AI, DeepHealth and Moderator of the session
• Jorrit Glastra, Head of Technology Clinical AI, DeepHealth
• Jason Sinner, MD, Medical Director, RadNet Management, Inc.

Register now: https://www.hoppr.ai/breakingthemoldwebinar

About HOPPR

HOPPR is a purpose-built development platform that accelerates AI in medical imaging. Built for developers, innovators, and imaging vendors, HOPPR offers ViT-based foundation models, API-based fine-tuning, and a QMS-aligned environment to support traceable AI development workflows. With curated datasets, structured outputs, and transparent documentation, HOPPR's secure and privacy-centric patent-pending technology helps bridge the gap between AI innovation and workflow integration. For more information, visit www.hoppr.ai.

SOURCE HOPPR