NeuroImage: Clinical (Jan 2022)

Concordance of regional hypoperfusion by pCASL MRI and 15O-water PET in frontotemporal dementia: Is pCASL an efficacious alternative?

  • Tracy Ssali,
  • Lucas Narciso,
  • Justin Hicks,
  • Linshan Liu,
  • Sarah Jesso,
  • Lauryn Richardson,
  • Matthias Günther,
  • Simon Konstandin,
  • Klaus Eickel,
  • Frank Prato,
  • Udunna C. Anazodo,
  • Elizabeth Finger,
  • Keith St Lawrence

Journal volume & issue
Vol. 33
p. 102950

Abstract

Read online

Background: Clinical diagnosis of frontotemporal dementia (FTD) remains a challenge due to the overlap of symptoms among FTD subtypes and with other psychiatric disorders. Perfusion imaging by arterial spin labeling (ASL) is a promising non-invasive alternative to established PET techniques; however, its sensitivity to imaging parameters can hinder its ability to detect perfusion abnormalities. Purpose: This study evaluated the similarity of regional hypoperfusion patterns detected by ASL relative to the gold standard for imaging perfusion, PET with radiolabeled water (15O-water). Methods and materials: Perfusion by single-delay pseudo continuous ASL (SD-pCASL), free-lunch Hadamard encoded pCASL (FL_TE-pCASL), and 15O-water data were acquired on a hybrid PET/MR scanner in 13 controls and 9 FTD patients. Cerebral blood flow (CBF) by 15O-water was quantified by a non-invasive approach (PMRFlow). Regional hypoperfusion was determined by comparing individual patients to the control group. This was performed using absolute (aCBF) and CBF normalized to whole-brain perfusion (rCBF). Agreement was assessed based on the fraction of overlapping voxels. Sensitivity and specificity of pCASL was estimated using hypoperfused regions of interest identified by 15O-water. Results: Region of interest (ROI) based perfusion measured by 15O-water strongly correlated with SD-pCASL (R = 0.85 ± 0.1) and FL_TE-pCASL (R = 0.81 ± 0.14). Good agreement in terms of regional hypoperfusion patterns was found between 15O-water and SD-pCASL (sensitivity = 70%, specificity = 78%) and between 15O-water and FL_TE-pCASL (sensitivity = 71%, specificity = 73%). However, SD-pCASL showed greater overlap (43.4 ± 21.3%) with 15O-water than FL_TE-pCASL (29.9 ± 21.3%). Although aCBF and rCBF showed no significant differences regarding spatial overlap and metrics of agreement with 15O-water, rCBF showed considerable variability across subtypes, indicating that care must be taken when selecting a reference region. Conclusions: This study demonstrates the potential of pCASL for assessing regional hypoperfusion related to FTD and supports its use as a cost-effective alternative to PET.

Keywords