[Magnetic resonance imaging in dementia]

Magnetic resonance imaging (MRI) plays a pivotal role in the diagnostic work-up of dementia. In addition to excluding secondary and potentially treatable causes, MRI enables the identification of characteristic neuroanatomical patterns that support the differentiation between major dementia entities. In daily clinical practice, the diagnostic value of MRI relies primarily on systematic visual assessment and semiquantitative evaluation of regional atrophy patterns and accompanying vascular pathology. In recent years, automated morphometric analyses and artificial intelligence (AI)-based approaches have gained increasing attention. While these techniques promise objective quantification of structural brain changes, their clinical utility on an individual patient level remains limited. High interindividual variability of brain volumes, technical dependencies, and normative database issues restrict their diagnostic accuracy and may lead to overinterpretation of numerical results if used outside the appropriate context. This review provides a clinically oriented overview of MRI in dementia imaging. Emphasis is placed on pattern-based visual assessment, the role of established semiquantitative rating scales and a critical appraisal of automated morphometry. A practical imaging algorithm is proposed that integrates visual analysis, rating scales and quantitative tools in a structured manner. The aim is to highlight the realistic clinical value of MRI and to emphasize its role as an interpretative rather than a purely numerical diagnostic modality.