50k+
Oncological lesions in training
Model trained on more than 50,000 annotated oncological lesions from real-world data allowing it to generalize to other applications.
The AI-native oncology viewer
Raidium Read
Automating the oncology imaging workflow
A universal, interactive radiology cockpit powered by built-in foundation models — from detection to a structured report for increased productivity, starting with oncology.
Why Raidium
Shift the AI radiology paradigm
Designed to help radiologists move from a collection of narrow AI-solutions to a unified tool. Unlike traditional AI, our foundation models and agents understand the entire human anatomy.
The Tool Trap
Juggling narrow AI add‑ons that each solve only one task, exam type, or organ.
The Universal Cockpit
A single, promptable viewer that handles detection, segmentation and reporting across any workflow.
The Universal Cockpit
A single, promptable viewer that handles detection, segmentation and reporting across any workflow.
Features
Starting with an AI-powered oncology imaging workflow
Each step is automated by Raidium's foundation models, inside the radiologist's existing PACS viewer.
Raidium Read
DetectReal productivity gains
10x
Faster on lesion annotation
Measured speed-ups vs. manual workflow.
0.85 / 0.74
Semi-auto lesion DICE — MRI / CT
Oncopilot study, validated across brain, liver, lung, bone and more.
From our partners
Fewer than 8% of cancer patients worldwide are enrolled in a clinical trial. Part of the problem is structural: no physician can manually verify more than thirty eligibility criteria per trial option for every patient. Raidium is our imaging backbone. Their models evaluate RECIST 1.1 at the level of a radiologist and automatically flag events that would disqualify a patient from a trial. No other team in Europe combines that model quality with the clinical depth needed to make this project, MA-MTB, work in a real hospital setting.
Coming soon
Building a unified AI-native viewer
We are building toward a universal AI-native viewer that covers every radiology workflow.
Stage 1
Oncology workflow
Automated RECIST 1.1 measurement and follow-up, available today.
Stage 2
Universal AI-native viewer
One cockpit covering every radiology specialty and workflow.
Key benefits
Increase your productivity with Raidium
Every workflow improvement is built to give radiologists back time, without compromising on clinical rigor.
Increased reliability
Our state‑of‑the‑art models deliver clinical‑grade results and reduce inter‑reader variability while being on‑par with radiologists.
Time savings
By automating low‑value tasks, our workflow empowers radiologists to handle more cases and dedicate their expertise to complex interpretation.
A unified tool
One viewer for all your oncology imaging workflows, able to extend to additional specialties.
Raising the standard of care
Facilitates good practices so more patients receive standard of care oncology follow-up, wherever they are.
Ready to transform your oncology workflow?