Digital Decision Support tool for MR-diagnostics in Multiple Sclerose (MS), objectives and functionality of the Product:

Our goal is to develop, clinically validate and implement a Product based on AI algorithms, in collaboration with the contracted Supplier, to be used as a clinical decision support tool in the diagnosis and monitoring of MS patients, to improve quality, diagnostic opportunities and efficiency.

The Product should fulfill at least the following requirements on functionalities:

  • The Product should be able to initiate analysis of MR images from the PACS of Karolinska University Hospital Region Stockholm's "Bild- och Funktionstjänst" (BFT).
  • The Product should be designed for brain MRI analysis and compatible with MR images from all clinical MRI scanners available at Karolinska University Hospital.
  • The Product should be able to automatically segment MS lesions T2-weighted FLAIR (hyperintensities) and T1-weighted images (hypointensities) with high precision.
  • The Product should be able to automatically segment contrast-enhancing MS lesions on T1-weighted images with gadolinium-based contrast agents with high precision.
  • The Product should be able to predict contrast-enhancement in MS lesions based on conventional non-enhanced MRI with high accuracy. The predicted contrast-enhancement should be visualized topographically and accompanied by a measure of its probability (on the scale 0-100%).
  • The Product should be able to analyze multiple timepoints of brain MRI and be able to compare differences in lesion count and lesion volume longitudinally.
  • The Product should be able to automatically produce subtraction imaging between timepoints (if multiple timepoints exist) to visualize new and changed lesions.
  • The Product should be able to automatically quantify global (grey matter, white matter) and regional tissue volumes (for example basal ganglia) of structures in the brain both cross-sectionally and longitudinally.
  • The Product's analysis results should be robust in terms of repeatability (same scanner) and reproducibility (different scanners).
  • The Product's analysis results should be easily accessible in PACS as a structured report (containing at least lesion count and volume of T1-hypointensities, T2- hyperintensities, contrast-enhancing lesions, predicted contrast-enhancement and other volumetric measurements).
  • The Product's analysis results in the form of a structured report should also be exportable in a tabulated format (such as .csv) for transfer to patient charts, patient administrative systems (PAS) and quality registers.
  • The Product's volumetric analysis results should be exportable to an external media in the form of anatomical masks (binary and/or probabilistic).

In addition, the client wishes to keep open for further development of functionalities based on perceptions of future potential needs and clinical benefit in the area. The development of these functionalities, and possibly other functionalities proposed by the Supplier, shall be agreed between the Client and the Supplier in accordance with the attached agreement, and may include the following:

  • Corresponding functionality, as specified in the requirements above, for spinal MRI.
  • Creation of preliminary radiological reports.
  • Medical triage of the order of reporting of MRIs.
  • Prediction of contrast-enhancement in brain and/or spinal MRI without contrast agents for other neurological disorders.

The product may be used in publicly funded neuroradiological activities with access to Region Stockholm's common medical image data storage (Image and Functional Service - BFT); besides Karolinska eg SÖS, Södertälje, S: t Göran, Sabbatsberg, Farsta, Täby.

For more information follow the links the boc below. In case there are un-clarities w r t the published procurement documents, there is an additional opportunity to set up meetings during the week 8th-12th of July. Please contact Peter Losman, procurement officer,

Dead-line for submission is 30st of August.

Karolinska University Hospital is procuring innovation of another AI decision support tools

The Center for Innovation

About the Center for Innovation

Current innovation partnerships