In:
Alzheimer's & Dementia, Wiley, Vol. 17, No. S4 ( 2021-12)
Abstract:
Magnetic resonance imaging (MRI) has become important in the diagnostic work‐up of neurodegenerative diseases. icobrain dm, a CE‐labeled and FDA‐cleared automatic tool for brain volumetry using clinical MRI scans, has shown great potential in differentiating cognitively healthy controls (HC) from Alzheimer’s disease (AD) dementia (ADD) patients in selected research cohorts (Struyfs H, et al. (2020)).This study examined the diagnostic value of icobrain dm for AD in routine clinical practice, and compared it to FreeSurfer, one of the standard tools in neuroimaging research. Method The Belgian Dementia Council (BeDeCo) initiated a retrospective, multi‐center Belgian study (REMEMBER). Through automated volumetric analyses of whole brain, gray matter, cortical gray matter, frontal, parietal, occipital and temporal cortices, hippocampal volumes and lateral ventricles on real‐world clinical brain MRI T1w (n=820) images across the AD continuum, icobrain dm’s (v.4.4.0) ability to differentiate diagnostic groups in the AD continuum vs. HC was compared to FreeSurfer (v.6.0). Using icobrain dm’s automated volumetric output, we also investigated which combined brain volumes contributed most to establish an AD diagnosis and could differentiate between different stages of the disease. Result The final study population consisted of subjects from eight Belgian memory clinics, with HC (n=89), subjective cognitive decline (SCD, n=89), mild cognitive impairment (MCI, n=350), and ADD (n=250) (Table 1). icobrain dm outperformed FreeSurfer in processing time (15‐30 minutes vs. 9‐32 hours), robustness (0 failures vs. 67 failures due to complete abortion of analysis during topology corrections) and diagnostic performance (AUC, sensitivity and specificity) in univariate differentiation between HC and ADD patients (Figure 1, Table 2). Stepwise backward regression models showed improved diagnostic accuracy for pairwise group differentiations when combining multiple brain volumes computed by icobrain dm, with highest performance obtained for distinguishing HC from ADD patients (AUC=0.914; Specificity 83.0%; Sensitivity 86.3%) (Table 3). Conclusion Automated volumetry has a diagnostic value for ADD diagnosis in routine clinical practice. Our findings indicate that icobrain dm ’ s automated volumetry can improve diagnostic certainty in a real‐world clinical setting with a significantly reduced processing time, robustness and diagnostic performance, as compared to FreeSurfer.
Type of Medium:
Online Resource
ISSN:
1552-5260
,
1552-5279
Language:
English
Publisher:
Wiley
Publication Date:
2021
detail.hit.zdb_id:
2201940-6
Bookmarklink