HOPPR Releases Chest Radiography Model with Fine-Tuning and Inference API Access

The HOPPR Marie Curie Chest Radiography Model enables partners to fine-tune and deploy binary classifiers for chest X-ray images using their own data with API access, structured outputs, and usage billing.

CHICAGO, July 17, 2025 /PRNewswire/ -- HOPPR, a secure AI development platform for medical imaging, today announced the commercial release of its HOPPR Marie Curie Chest Radiography Foundation Model and fine-tuning API for binary classification, alongside inference API access and a new usage billing portal. This first commercial launch marks a key milestone in delivering developer infrastructure for building medical imaging AI models.

DeepHealth, an AI innovator and PACS platform provider, is the first commercial partner to adopt the model and API-driven fine-tuning workflow. As part of a limited release and prior to commercial launch, DeepHealth fine-tuned multiple models and deployed a customized chest radiography classifier within weeks, demonstrating the platform's adaptability and real-world potential.

"We were looking for next-gen solutions that perform as well or better than today's Convolutional Neural Networks (CNNs) in a diverse, global organization like RadNet. We did a broad-based evaluation with models in the market and selected HOPPR as our development tool to fine-tune and deploy chest X-ray models with confidence, speed, and transparency," said Sham Sokka, Chief Operating and Technology Officer at DeepHealth, one of HOPPR's key strategic partnerships. "HOPPR's fine-tuning rapidly gave us a foundation model to move quickly and significantly reduce our development costs and improve our operational quality and effectiveness."

The Marie Curie Chest Radiography Foundation Model is built on a large-scale Vision Transformer (ViT) architecture and optimized for binary classification in chest X-ray imaging. Through a single API, developers can fine-tune the model using their own labeled datasets and receive structured outputs and prediction scores via the inference API.

"Our goal is to give medical imaging developers a powerful and flexible model they can make their own without the heavy cost, complexity, and long timelines of building infrastructure from scratch," said Dr. Khan Siddiqui, CEO and Co-founder of HOPPR. "With access to our fine-tuning and inference APIs and integrated billing portal, this release gives developer teams what they need to go from labeled data to deployment with speed, control, and transparency."

Unlike static AI marketplaces with prebuilt applications, HOPPR provides a secure development environment for customizing foundation models. Built within a quality management system, the platform is designed to support traceable, reproducible development workflows. HOPPR also provides model training artifacts to support partner regulatory submissions. 

"We invested in HOPPR because they're redefining how AI imaging applications are built and deployed for physicians," said Lawrence K. Cohen, CEO at Health2047, a venture studio founded by the American Medical Association. "This launch validates HOPPR's platform approach, giving imaging developers the speed, control, and infrastructure they need to build AI solutions ready for real-world clinical integration."

While this release is optimized for teams with existing data and fine-tuning expertise, future platform updates will expand access to HOPPR's proprietary data, cohort-building tools, and model validation capabilities, broadening support for customers who do not have data or infrastructure in place.

HOPPR is also collaborating with research institutions, including the University of Miami, to explore potential clinical and research use cases. Upcoming updates will expand to new modalities and introduce language-based interactions.

About HOPPR
HOPPR is a secure AI development platform for medical imaging that enables organizations to build, deploy, and scale AI solutions with speed and flexibility. Built for developers and imaging vendors, HOPPR offers trusted tooling, models, access to curated data with known provenance, and an integrated quality management system to support the development of AI applications, bridging the gap between innovation and real-world deployment.  For more information, visit www.hoppr.ai.

View original content to download multimedia:https://www.prnewswire.com/news-releases/hoppr-releases-chest-radiography-model-with-fine-tuning-and-inference-api-access-302507380.html

SOURCE HOPPR