# Advanced neuroimaging assessment of neurodegenerative dementia syndromes: A framework for comprehensive multimodal FDG-PET, MR-perfusion, and MR-diffusion analysis

## TL;DR

This paper introduces a new imaging method combining PET and MRI to better diagnose different types of neurodegenerative dementia.

## Contribution

The study proposes a novel multimodal framework integrating FDG-PET, MR-perfusion, and MR-diffusion for improved dementia subtype differentiation.

## Key findings

- White matter FA changes were observed near regions with GM metabolic and perfusion alterations.
- Multimodal SVM achieved 81-94% accuracy in distinguishing NDS subtypes and SCD.
- Combining FDG-PET and MRI could enhance differential diagnosis of neurodegenerative dementia.

## Abstract

•New approach integrating metabolic, perfusion, and microstructural imaging in NDS.•FA was altered in WM adjacent to regions of GM alterations (rSUV, rCBF)•Multimodal SVM reached 81% to 94% accuracy for separating NDS from SCD.•Advanced MRI may complement PET for differential diagnoses in future.

New approach integrating metabolic, perfusion, and microstructural imaging in NDS.

FA was altered in WM adjacent to regions of GM alterations (rSUV, rCBF)

Multimodal SVM reached 81% to 94% accuracy for separating NDS from SCD.

Advanced MRI may complement PET for differential diagnoses in future.

The reference for imaging in neurodegenerative dementia syndromes (NDS) is [18F]fluorodeoxyglucose ([18F]FDG) positron emission tomography (PET). Advanced multimodal magnetic resonance imaging (MRI) may complement PET-derived measures for differential diagnostics.

The aim of the present study was to evaluate a new methodological approach integrating metabolic, perfusion, and microstructural parameters from simultaneous FDG-PET/MRI to reveal different imaging signatures for different NDS subtypes.

66 patients with NDS (Alzheimer’s disease: 28; behavioral frontotemporal dementia: 10; semantic variant primary progressive aphasia (PPA): 8; non-fluent variant PPA: 11; logopenic variant PPA: 9) and 10 subjects with subjective cognitive deficits (SCD) underwent combined FDG-PET/MRI with pseudo-continuous arterial spin labeling (pCASL) and diffusion tensor imaging (DTI). Standardized uptake values (SUV), cerebral blood flow (CBF), and fractional anisotropy (FA) were used to separate a respective NDS subgroup from all other NDS subgroups and from SCD based on a support vector machine (SVM) applied on region of interest (ROI)-based parameters.

White matter alterations directly adjacent to the regions of alterations in SUV and CBF were identified. The SVM analysis on ROI-based parameters reached accuracies between 81% and 94% for separating an NDS subgroup from all other NDS subgroups and from SCD.

Multimodal neuroimaging by combination of FDG-PET and MRI shows potential to further advance the diagnostic spectrum in NDS. Since particularly early diagnosis of NDS remains key for effective treatment and management of patients with NDS, the present framework appears promising to be developed further until it aligns and integrates with clinical routine.