Perfusion as a biomarker of brain dysfunction in dementia (AD, DLB, FTD/PPA, PDD): comparison of CT, MRI (ASL/DSC/DCE), SPECT, and PET with interpretive pitfalls - a narrative review.

The term 'brain perfusion' is applied in clinical practice to a family of neuroimaging techniques that measure, in reality, quite different physiological quantities in the brain circulation. CT perfusion (CTP) and DSC-MRI track a contrast bolus to derive semi-quantitative haemodynamic parameters. ASL-MRI estimates cerebral blood flow (CBF) without contrast, but the result is sensitive to arterial transit time, haematocrit, and the patient's haemodynamic state on the day of the scan. DCE-MRI quantifies blood-brain barrier (BBB) permeability, a property of the neurovascular unit rather than a flow measurement. Perfusion SPECT provides a relative, normalisation-dependent CBF map that is distorted by cortical atrophy. [¹⁸F]FDG-PET reflects synaptic glucose metabolism, but not blood flow. Because these techniques answer different physiological questions, their results are not interchangeable, and applying a threshold or pattern derived from one modality to interpret another is methodologically unsound - yet this conflation occurs with regularity in clinical practice. This narrative review synthesises the clinical applications, diagnostic performance, and interpretive pitfalls of each technique in Alzheimer's disease (AD)/mild cognitive impairment (MCI-AD), dementia with Lewy bodies (DLB)/Parkinson's disease dementia (PDD), frontotemporal dementia (FTD)/primary progressive aphasia (PPA), and mixed dementia with vascular pathology. Recognised limitations include the narrative study design and the small number of head-to-head multi-modal studies in pathologically confirmed cohorts. A comparative table and practical minimum reporting elements are provided.