What working with us looks like.

Early hospital collaboration happens in a research and development context only — not routine clinical use. Collaborating sites can take part in workflow evaluation, technical and deployment evaluation, and, where separately agreed, clinical validation.

What a collaboration typically requires: ethics or governance approval as applicable, an appropriate data-processing or data-sharing agreement, retrospective imaging data, hospital IT involvement, and defined radiologist reader time.

No software licence fee is charged for an agreed early research collaboration. Each party bears its own internal personnel and governance costs unless otherwise agreed in writing.

Participation in the formal validation programme is subject to the applicable study protocol and is not implied by being an early collaborating site. And to repeat what the footer already says: Lävi Delta must not be used for clinical decision-making — nothing about a collaboration changes that.

The intended product deployment is on-premise, inside the hospital network: studies arrive over standard DICOM, are processed without operator interaction, and standard DICOM objects return into the existing PACS workflow. In that deployment, no image data needs to leave the institution.

We do not publish hardware figures yet. The requirements are not frozen, and a published number that changes is worse than no number — deployment specifics are worked through with hospital IT as part of each collaboration.

Current practice, stated as fact: on-premise processing as the intended product deployment architecture; role-based access control and multi-factor authentication in the reader environment; audit logging of reader actions where implemented.

In preparation — listed as intent, not as achievement: alignment work with MDCG 2019-16 (the EU guidance on medical-device cybersecurity); independent penetration testing; a formal update and patch-management policy; and verified product-wide encryption at rest and in transit. None of these are claimed as complete until their technical verification is.

The full producer-to-reader chain has been executed against a full production-form deployment of the pipeline, image store and reader workstation — including verifying that rejected cases stay rejected. It was not a routine clinical deployment at a hospital site, and we will not imply that it was. Details are on the technology page; what is and is not proven is on the evidence page.

If your department reads longitudinal brain MRI and the programme above interests you, write about clinical collaboration. A neuroradiologist answers.

Status as of 2026-08-21