Blood-brain barrier permeability measurement by biexponentially modeling whole-brain arterial spin labeling data with...

Blood-brain barrier (BBB) permeability assessment remains of ongoing interest in clinical practice and research. Transitions between intravascular (IV) and extravascular (EV) gray matter (GM) compartments may provide information regarding the microstructural status of the BBB. Due to different transverse relaxation times (T₂ ) of water protons in vessels and GM, it is possible to determine the compartment in which these protons are located. This work presents and investigates the feasibility of a simplified analytical approach for compartmentalizing the proportions of magnetically marked water protons into IV and EV GM components by biexponentially modeling T₂ -weighted arterial spin labeling (ASL) data. Numerous model assumptions were used to stabilize the fit and achieve in vivo applicability. Particularly, transverse relaxation times of IV and EV water protons were determined from the analysis of two supporting T₂ -weighted ASL measurements, utilizing a monoexponential signal model. This stabilized a two-parameter biexponential fit of ASL data with T₂ preparation (PLD = 0.9/1.2/1.5/1.8 s, TE_T2Prep  = 0/30/40/60/80/120/160 ms), which thereby robustly provided estimates of the IV and EV compartment fractions. Experiments were conducted with three healthy volunteers in a 3 T scanner. Averaged over all subjects, the labeled water protons inherit T_2,IV  = 200 ± 18 ms initially and adapt T_2,EV  = 91 ± 2 ms with a longer retention time in cerebral structures. Accordingly, the EVlocated ASL signal fraction rises with increasing PLD from 0.31 ± 0.11 at the shortest PLD of 0.9 s to 0.73 ± 0.02 at the longest PLD of 1.8s. These results indicate a transition of the water protons from IV to EV space. The findings support the potential of biexponential modeling for compartmentalizing ASL spin fractions between IV and EV space. The novel integration of monoexponential parameter estimates stabilizes the two-compartment model fit, suggesting that this technique is suitable for robustly estimating the BBB permeability in vivo.