French startup Raidium deploys AI radiology viewer in US
Paris-based Raidium has deployed an AI-native imaging platform at a leading US cancer center, proving European tech can upend entrenched medical infrastructure.
Raidium, a startup with bases in Paris and Silicon Valley, has deployed its imaging platform at Moffitt Cancer Center, a top American oncology research institution. The system, Raidium Read, replaces Moffitt's older radiomics applications and is currently limited to clinical trials and research use. The company expects to secure FDA 510(k) clearance for broader clinical application before the end of 2026.
The platform is powered by Curia, a proprietary foundation model trained on more than 200 million CT and MRI slices from 150,000 exams. Rather than adding artificial intelligence tools to existing Picture Archiving and Communication Systems (PACS), Raidium built the viewer entirely from scratch with the model embedded at its core. This allows the system to perform automated, organ-agnostic RECIST measurements—the standard method for tracking how tumors respond to treatment—across multiple time points.
This automation addresses a major bottleneck in oncology. Radiologists traditionally spend significant time manually pulling measurements from prior studies to compare against new imaging, a workflow that is slow and highly inconsistent between different readers. By automatically detecting and segmenting lesions across anatomical regions, then mapping this historical data against new follow-up scans, Raidium Read reduces inter-reader variability by a factor of three. Furthermore, the system requires no backend integration, allowing for faster deployment than traditional software installations.
For European health-tech developers, the Moffitt deployment signals a viable route into the highly protected American healthcare market. While artificial intelligence has demonstrated the ability to outperform biopsies in grading rare cancers, most of those tools remain research prototypes. Raidium's strategy of rethinking the foundational software—treating an error-prone clinical task as a reasoning problem, much like Corti's Symphony AI did for medical coding—offers a more practical path to clinical integration. Securing a leading US research institution as an initial customer provides a strong proof of concept that could attract further investor and commercial interest.
“For twenty years, the standard PACS viewers have resisted evolution,” said Paul Herent, Raidium’s CEO and co-founder. The early results at Moffitt support his view. Dr. Cesar Lam, a radiologist at the center, noted that the platform enables research projects that “would have seemed impossible not too long ago.”
