The early detection and follow-up of neurodegenerative disorders, like Dementia or Multiple Sclerosis, is still a challenge. Even the (unaided) eye of the well-trained expert Neuroradiologist is not always able to detect subtle signs and changes. In these cases, a quantitative volume assessment of different brain regions and an automatic detection and characterization of brain lesions can be extremely valuable – particularly if analyzed in terms of their dynamics. Such reliable, quantitative reports can also be used by the treating Neurologist in order to optimize treatment.
To develop those kinds of expert systems, we collaborate with mediaire GmbH in Berlin. This collaboration aims to develop innovative software solutions for the automated analysis of MRI scans to support radiologists in their daily routine and improve the quality of their reports. Perspectively, we want to enable the automatic output of the most likely diagnosis including potential alternative diagnoses powered by artificial intelligence (AI). A unique feature of mediaire’s approach is the completely local integration of the software, removing the need for patient data to leave the practice. This allows for a maximum protection of sensitive patient data.
Our contribution to this project is to review the analyses and reports generated by the computer system and to counsel the company on software optimization.