Quantification of ASL Perfusion MRI in MCI Patients Using Single-Post Labeling Delay Model-Fitting Approach

Abstract

Purpose: Neuroimaging techniques hold many promises to detect Alzheimer Disease (AD) at early stages before clinical symptoms fully develop, suggesting decreased regional Cerebral Blood Flow (CBF). Perfusion deficiencies are present from very early clinical phases of AD, i.e. Mild Cognitive Impairment (MCI) and persist well into the latest stages, demonstrating a pattern of increased hypo-perfusion with the disease development. Accurate quantification such as quantification model and noise reduction method is necessary to achievement of good results in insufficient Post Labeling Delay (PLD) time.
Materials and Methods: Arterial Spin Labeling (ASL) is a non-invasive MRI technique to extract brain regional CBF, which in recent years gained wide acceptance for its value in clinical and neuroscience applications. In the present work, 44 participants in 2 groups were imaged (normal control and MCI) using single-Post Label Delay (single-PLD) model-fitting ASL Perfusion at 1.5T. Images were de-noised with Independent Component Analysis (ICA) algorithm and then preprocessing such as motion correction, distortion correction, normalization, and tissue segmentation analyzed with SPM12. Calibration image that needed to calculate absolute perfusion was estimated from the data by fitting a curve to the control images in the dataset. Absolute CBF values were finally extracted from the kinetic model quantification.
Results: Perfusion decreases in the right precuneus (28%), right inferior partial cortex (22%) and right middle frontal cortex (20%) cortex in MCI subjects.
Conclusion: According to the results, ASL-MRI is able to calculate perfusion changes associated with MCI.