Spatial Patterns of Cerebral Blood Flow in Alzheimer's Disease Identified by the Subtype and Stage Inference Algorithm.

INTRODUCTION

Much research has focused on the deposition of amyloid and tau proteins in the Alzheimer's disease (AD) brain, but many amyloid and tau models assumed a single spatial progression of amyloid and tau accumulation. We estimated the changing patterns of an indirect biomarker, i.e., the cerebral blood flow (CBF), in AD, and we discuss the pathological process of AD.

METHODS

The participants were 341 patients who visited our hospital's outpatient department for memory loss (146 males, 195 females): 115 diagnosed with AD, 176 diagnosed with mild cognitive impairment, and 50 diagnosed with subjective cognitive decline. For the evaluation of disease-related changes in their CBF, the patients underwent 99mTc-ethyl cysteinate dimer single-photon emission computed tomography scans. We differentiated the subtypes of CBF in AD by using a machine-learning algorithm called the "Subtype and Stage Inference (SuStaIn)"algorithm.

RESULTS

When we divided the data into two groups, the SuStaIn algorithm identified two different CBF subtypes: the typical AD pattern and a cortical pattern with hippocampal sparing.

CONCLUSION

We observed two subtypes of the pattern of change in the CBF of individuals with AD, and these subtypes were highly similar to previous findings derived from SuStaIn algorithm applied differing neuroimaging modalities. Such subtyping derived from CBF imaging might have clinical utility in the treatment of AD.