Imagine a smart assistant for doctors that has already 'seen' millions of medical scans and knows exactly what to look for. Instead of a human spending hours labeling every single image by hand, this AI learns from the patterns on its own. It helps doctors spot disease markers quickly and accurately, making medical tests and drug trials much faster.
Radiologists are overwhelmed by increasing workloads due to aging populations, and the development of precision medicine is slowed by the high cost and time required to manually label medical data for AI.
The Curia foundation model, a multimodal AI trained on 130 TB of data, and an interactive human-machine interface for automated radiology segmentation and reporting.
If you are a CRO dealing with slow and expensive patient screening for drug trials — this project developed a multimodal foundation model that creates scalable biomarkers. This makes clinical trials faster and more cost-effective by automating the identification of specific patient traits.
If you are a clinic dealing with an enormous workload of radiology scans due to an aging population — this project developed an interactive AI platform that performs automatic segmentation and annotation. This relieves radiologists from repetitive tasks and provides decision support.
If you are a software provider dealing with the high cost of manually labeling data for AI training — this project developed a self-supervised learning model trained on 130 TB of data. This allows for the creation of biomarkers without the time-consuming effort of manual labeling.
Based on available project data, the company intends to offer the AI radiology platform as a Software as a Service (SaaS) to hospitals and pharma companies, though specific pricing tiers are not listed.
Yes, the platform uses a foundation model that is scalable across different segments. It has already been evaluated across six clinical domains and specifically mentions applications in NASH, liver diseases, and oncology.
Based on available project data, the technology is developed by Raidium, an SME, but specific licensing terms or patent numbers are not provided in the summary.
The project features a human-centric interface that allows radiologists to train the AI using their own experience while performing automatic segmentation and reporting tasks.
The project period is from 2024-11-01 to 2026-10-31, indicating the development and integration phase is currently active.
The project is led by a single partner, Raidium, a French SME. This 100% industry-led structure suggests a strong commercial drive and agility, as there are no university or research institute partners slowing down the transition to market.
Contact Raidium (France) regarding their Curia foundation model
Contact us to explore licensing or partnership opportunities with Raidium.
