Raidium Read is being used at Moffitt Cancer Center for research and clinical-trial work before expected FDA 510(k) clearance, TNW reported.

The deployment does not make Raidium Read a fully cleared clinical product.

It places the AI radiology viewer inside an oncology research setting where radiologists compare lesions across sequential scans and where repeatability can matter as much as speed.

Moffitt is using the platform after replacing legacy radiomics applications.

Oncology imaging workflows often require clinicians to measure the same lesions across multiple time points, compare prior and current scans and keep trial measurements consistent across readers.

Raidium built the viewer around Curia, its proprietary foundation model, rather than adding a separate AI assistant to an existing PACS viewer.

TNW reported that the model was trained on more than 200 million CT and MRI slices from 150,000 exams.

Curia performs organ-agnostic automated RECIST measurements across multiple time points.

RECIST is the standard method used to track tumour response to treatment, and Raidium says its system reduces inter-reader variability by a factor of three.

That performance claim remains company-provided in the TNW account.

The article did not include independent benchmark methodology, patient-outcome data or a completed FDA clearance decision.

The viewer reviews large imaging sets, segments lesions in different anatomical regions and compares earlier lesion records with new follow-up scans.

The operating claim is not just lesion detection.

It is that a radiology workflow can carry structured lesion evidence from one scan to the next without forcing staff to rebuild the measurement trail manually.

Raidium CEO and co-founder Paul Herent said standard PACS viewers have resisted evolution for two decades.

Moffitt radiologist Dr. Cesar Lam said the system could support research work that recently would have been difficult to imagine.

Those comments point to the same workflow problem from two sides: software vendors want AI-native image review, while cancer centres need tools that can support trial evidence without disrupting clinical imaging systems.

Raidium says the system requires no backend integration, a claim that would lower deployment friction if it holds across hospital environments.

The next proof is whether Raidium can move from a research and trial installation to cleared clinical use with independent validation, named commercial deployments and measurable gains in radiology workflow quality.