AI Supported Detection of Cerebral Multiple Sclerosis Lesions Cuts Radiologic Reporting Times

By MedImaging International staff writers
Posted on 13 Nov 2024

Image: The study found a reduction in reporting time through AI-supported detection of cerebral MS lesions (Photo courtesy of mediaire)

Multiple Sclerosis (MS) is a common autoimmune disorder that affects the central nervous system. MRI plays an essential role in both diagnosing and monitoring disease progression and treatment responses. As such, reviewing cerebral MRIs for MS patients has become a routine part of clinical practice. To support reporting, numerous commercial software solutions for automated lesion detection have been introduced in the market. Now, a new study has highlighted the significant improvements in radiologic reporting for MS patients through the use of artificial intelligence (AI)-enhanced software. Published in the European Journal of Radiology, the study emphasizes significant reductions in reporting times and improved efficiency in detecting cerebral lesions.

In the study, researchers at the University Hospital of Schleswig-Holstein (Schleswig-Holstein, Germany) evaluated the effect of mediaire GmbH’s (Berlin, Germany) mdbrain software with AI-supported lesion detection on radiologic reporting. mdbrain is designed to automatically label, visualize, and quantify 3D MRI data of the head, thus automating the previously manual tasks of identifying, labeling, and calculating the volume of segmented brain structures in MRI images. The volumetric data, combined with image data, aids radiologists in detecting neurological disorders. The software currently offers modules for brain volumetry, lesion characterization, aneurysm detection, and tumor differentiation.

The study tested mdbrain’s effectiveness on 50 patient cases and found a substantial reduction in average reporting times for both initial and follow-up evaluations. Specifically, mdbrain shortened reporting time by 90 seconds for initial assessments and 75 seconds for longitudinal assessments—about one-third of the time required for reporting without AI assistance. The results also demonstrated that the software maintains high interrater reliability, matching the accuracy of traditional methods while significantly accelerating the diagnostic process. In addition to automating lesion detection, mdbrain also provides accurate lesion type differentiation, which is critical for diagnosing and tracking MS progression. This research shows how integrating mdbrain into radiology workflows can speed up MRI reporting, offering significant potential for enhancing MS clinical management, improving workflow efficiency, and enabling faster treatment decisions.