Several new studies on mdbrain were presented at this year’s neuroRAD – including volumetry H2H vs. Siemens Syngovia et al, influence of AI on diagnostic accuracy, and external validation of automatic lesion characterization.
Several new studies presented at neuroRAD 2021 underline the clinical added value of mdbrain. We would like to briefly present these here:
1.) Anna-Lena Mayer et al. (University Hospital Erlangen, Neuroradiology): AI-based volumetry algorithms to assist in epilepsy imaging diagnostics.
Based on studies of 24 patients with histologically confirmed epileptogenic hippocampal sclerosis, the precision of mdbrain’s volumetry solution was investigated. Atrophy was correctly detected with a specificity of 100% and a sensitivity of 96%, while the BrainMorphometry solution from the Siemens Syngovia suite used for comparison had a sensitivity of only 36%.
2.) Stefan Hock et al. (University Hospital Erlangen, Neuroradiology): Latest Artificial Intelligence Provides Fast, Accurate and Consistent Detection of Multiple Sclerosis Lesions
Based on blinded studies of 101 patients with clinically confirmed diagnosis of multiple sclerosis (MS) according to McDonald criteria, the extent to which mdbrain lesion characterization correctly detects and regionally classifies inflammatory lesions was investigated. The expert judgment of three neuroradiologists was used as the ground truth. At the same time, it was investigated whether the methodological step towards fully DeepLearning-based algorithms was beneficial. The results showed significant improvements especially in sensitivity in all regions, with still potential for improvement in the area of infratentorial lesions.
3.) Merita Aruci et al. (University Hospital Magdeburg, Neurology): Challenging Cases for WMH Segmentation Comparatively Processed by Seven Automated Methods
Based on blinded examination of 10 patients with cerebral small vessel disease (CSVD), 6 different automated methods for segmentation were investigated with respect to their accuracy – Freesurfer, PGS, SLS, LST, BIANCA and mdbrain. Manual segmentation by experienced neuroradiologists served as the ground truth. mdbrain came out on top in all metrics. The study shows that the applicability of mdbrain‘s lesion algorithm is not limited to MS lesions.
4.) Jan Rudolph et al. (LMU Klinikum, Radiology): Artificial intelligence substantially improves differential diagnosis of dementia – Added diagnostic value of rapid brain volumetry
Based on blinded examinations of 35 patients with clinically confirmed diagnosis of dementia (fronto-temporal dementia or dementia of the Alzheimer type), as well as 20 healthy controls, it was investigated to what extent mdbrain’s volumetry solution positively influences the clinical diagnostic certainty. At the same time, the dependence on the level of radiological training was determined. As a result, all groups benefited from the additional pair of eyes of the AI solution, whereby the effect was stronger the lower the level of training in radiology.
The recordings of the presentations can be found at the congress site: https://dgnr.conference2web.com