MS follow-up is a change problem.

An MS follow-up read comes down to a handful of longitudinal questions. Have new lesions appeared? Have known ones enlarged? Has anything decreased or resolved? Could an apparent change be an artefact of registration, acquisition or image quality rather than disease? And how has the overall burden moved across the whole series of examinations?

Those questions span large three-dimensional studies acquired months or years apart, often on different scanners and under different protocols. Finding lesions on a single scan is the easier half of the job — segmentation is a largely solved conversation. The hard part, and the clinically decisive one, is reliable change assessment over time.

MS follow-up is not a segmentation problem. It is a change problem.

The planned system brings together baseline and follow-up comparison, longitudinal image alignment, MS lesion analysis, identification of candidate new and changing lesions, quality-control information, and a structured, auditable review of the longitudinal findings — one workflow shaped around the follow-up read, rather than a general tool with an MS label on it.

The radiologist remains responsible for the final interpretation. Lävi MS is not intended to make an autonomous MS diagnosis, and it will not make treatment decisions.

Lävi Delta comes first, and the ordering is deliberate. Delta establishes the general longitudinal foundation: image ingestion and provenance, registration, the quality-control and abstention architecture, the candidate-change workflow, the clinical review environment, auditability, and — just as importantly — the clinical-validation methodology itself.

Fig. 1 — Product architecturesolid = current programme · dashed = planned

Lävi MS then extends that foundation with disease-specific functionality for multiple sclerosis, under its own development, clinical-validation and regulatory pathway. Rather than pushing several regulated products through development at once, we validate the common platform through one product and build the second on proven ground.

Two things follow from this. Lävi MS is not an MS mode inside Lävi Delta — it is a separate product. And Lävi Delta is not a detour on the way to it — it is the platform the MS product will stand on.

The intended users are neuroradiologists and radiologists who routinely interpret MS follow-up MRI — in specialist MS centres and in hospital radiology departments. Neurologists may work with the structured outputs further along the clinical pathway, but the imaging interpretation itself remains a radiological task, and the product is designed around that.