Estimating cerebrospinal fluid biomarkers using brain perfusion SPECT in Alzheimer’s disease

Objective Alzheimer’s disease (AD) involves amyloid-β (Aβ) and phosphorylated tau (p-tau) pathology. Cerebrospinal fluid (CSF) biomarkers reflect these changes but require invasive lumbar puncture, whereas amyloid positron emission tomography (PET) is less invasive but expensive and typically limited to a single reimbursed scan. 2-Deoxy-2-[ ¹⁸ F]-fluoro-D-glucose (FDG)-PET is less invasive than lumbar puncture; however, it is not covered by health insurance in Japan. In contrast, brain perfusion single photon emission computed tomography (SPECT) is widely available, repeatable, and cost-effective compared to FDG-PET. We investigated the correlation between regional relative cerebral blood flow (rCBF) measured by SPECT and CSF biomarkers, and evaluated whether rCBF reductions could predict CSF biomarker levels. Methods In this retrospective study, 88 patients underwent Mini-Mental State Examination (MMSE), MRI, N -isopropyl- p -[ ¹²³ I] iodoamphetamine (IMP)-SPECT, and CSF biomarker assessments. SPECT data were normalized to cerebellar counts and co-registered to MRI. Voxel-wise analyses identified the regions where decreased rCBF correlated with CSF biomarkers. Simple regression evaluated correlations between the standard uptake value ratios (SUVRs) of posterior cingulate (PC), bilateral parietal cortices, MMSE, age, sex and biomarker levels. Multiple regression models incorporated the three SUVRs, MMSE, and age. Predicting validity was assessed using correlation coefficient (r), coefficient of determination (r²), and Bland–Altman analysis. Results Voxel-wise analysis revealed positive correlations between CSF Aβ42 level and CBF in PC and angular gyrus, while CSF t-tau and p-tau correlated negatively with parietal hypoperfusion. CSF Aβ40 showed no significant correlations. Simple regression demonstrated weak correlations, such as CSF Aβ42 with right parietal cortex (r = 0.47). Multiple regression yielded moderate predictability for CSF Aβ42 (r² = 0.42), whereas other biomarkers were poorly predicted. Conclusion SPECT revealed AD-typical hypoperfusion patterns and demonstrated modest potential to estimate CSF Aβ42 levels, but not CSF Aβ40, t-tau, or p-tau. Although SPECT cannot substitute CSF biomarker measurements or amyloid/tau PET, SPECT may serve as a pre-screening tool to identify patients requiring definitive biomarker testing.