Prediction of brain age using structural magnetic resonance imaging: a comparison of clinical utility of publicly available software packages.
Authors
Affiliations (5)
Affiliations (5)
- Centre for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet, & Stockholm Health Care Services, Region Stockholm, Sweden; Neurobiology Research Unit, Copenhagen University Hospital Rigshospitalet, Denmark; Department of Computer Science, University of Copenhagen, Denmark. Electronic address: [email protected].
- Neurobiology Research Unit, Copenhagen University Hospital Rigshospitalet, Denmark; Section of Biostatistics, University of Copenhagen, Denmark.
- Neurobiology Research Unit, Copenhagen University Hospital Rigshospitalet, Denmark; Department of Computer Science, University of Copenhagen, Denmark.
- Centre for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet, & Stockholm Health Care Services, Region Stockholm, Sweden.
- Centre for Psychiatry Research, Department of Clinical Neuroscience, Karolinska Institutet, & Stockholm Health Care Services, Region Stockholm, Sweden; Neurobiology Research Unit, Copenhagen University Hospital Rigshospitalet, Denmark.
Abstract
Brain age estimated from structural magnetic resonance images is commonly used as a biomarker of biological ageing and brain health. Ideally, as a clinically useful biomarker, brain age should indicate the current state of health and be predictive of future disease onset and detrimental changes in brain biology. In this preregistered study, we evaluated and compared the clinical utility, i.e., diagnostic and prognostic performance, of six publicly available brain age prediction packages using data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Baseline brain age differed significantly between groups consisting of individuals with normal cognitive function, mild cognitive impairment, and Alzheimer's disease for all packages, but with comparable performance to estimates of grey matter volume. Further, brain age estimates were not centred around zero for participants with normal cognition and showed considerable variation between packages. Finally, brain age was only weakly correlated with disease onset, memory decline, and grey matter atrophy within four years from baseline in individuals without neurodegenerative disease. The systematic discrepancy between chronological age and brain age among healthy subjects, combined with the weak associations between brain age and longitudinal changes in memory performance or grey matter volume, suggests that the current brain age estimates have limited clinical utility as a biomarker for biological ageing. This work was supported by a Longevity Impetus Grant from the Norn Group, the Karolinska Institutet Loo och Hans Ostermans Stiftelse, Gun och Bertil Stohnes Stiftelse, Stiftelsen Gamla Tjänarinnor, Stiftelsen Söderström - Königska and Åhlén-stiftelsen (243016). PPS was supported by a grant from the Swedish Brain Foundation (PD2024-0444) and the Åke Wibergs Stiftelse (M24-0117).