Building on more than 20 million existing imaging studies, today we're announcing new data capabilities, including EHR, pathology, labs, and cardiac data, to support the next generation of medical AI.
The announcement, on the opening day of HLTH Europe in Amsterdam, marks a major expansion of our role in the medical AI ecosystem.
Atlas currently gives AI developers access to more than 20 million medical imaging studies, including radiology reports and longitudinal datasets that can show the same patient’s scans across multiple time points. With today's expansion, we're adding support for additional data types, including EHR data, pathology results, and cardiac data such as ECG.
These new data types will be introduced progressively through Atlas, with the goal to give AI developers a scalable, responsible, and searchable infrastructure for accessing multimodal healthcare data as availability grows.
The expansion reflects a broader shift in medical AI as the field moves from narrow single-point solutions toward broader foundation models and more advanced clinical AI systems, developers increasingly need access to data that reflects the complexity of real-world care. Imaging remains essential, but many clinically valuable AI applications require additional context, including clinical notes, patient history, laboratory data, pathology results, and other structured or unstructured clinical information.
“Imaging is already one of the strongest applications for medical AI,” said Pranav Rajpurkar, PhD, Associate Professor at Harvard Medical School and Co-Founder of a2z Radiology AI. “By adding multimodal context, Gradient can help developers build more useful, representative, and generalizable systems, while addressing one of a field's biggest data access challenges.”
“Medical AI is moving fast, but the next generation of models will only be as good as the data behind them,” said Josh Miller, CEO of Gradient Health. “Imaging has been our foundation, and it remains central to what we do, but healthcare is multimodal by nature, a scan rarely tells the whole story on its own. By expanding Atlas to support multimodal data, we are helping AI developers build systems with richer clinical context, while giving healthcare organizations a secure and responsible way to contribute to innovation. Better data means better AI, and better AI should mean better access to advanced care for more patients.”
Our existing imaging infrastructure already supports a wide range of AI development use cases, including model training, testing and evaluation, regulatory submissions, and foundation model development. The addition of multimodal data will also improve search and cohort-building for imaging use cases, helping developers identify more clinically relevant datasets by combining imaging findings with other contextual signals.
This expansion also creates a larger opportunity for healthcare data partners. By working with Gradient Health, healthcare organizations can generate value from de-identified data assets while contributing to the development of more representative and clinically useful AI.
Learn more, and start a 7-day free trial, via our website:
.png?upscale=true&width=1120&upscale=true&name=Gradient%20Health%20Logo%20Pack%202026_Gradient%20Health%20Logo%20-%20Dark%20Outlined%20Stacked%20(1).png)
