A parameterized gamma-variate function can describe the temporal response of cerebral blood flow after Acetazolamide injection in Moyamoya patients

Abstract

Background: Quantitative modelling of the temporal response of cerebral blood flow (CBF) following Acetazolamide (ACZ) may be performed using a parameterized gamma-variate function with physiologically interpretable parameters. When combined with sequential CBF measurement before and after administration of ACZ, such a model could enable assessment of cerebrovascular reserve (CVR) while accounting for regional differences in the timing of the vascular response between steno-occlusive and unaffected territories. To evaluate this approach, we applied the model to patients with moyamoya disease (MMD). Methods: Twenty-four MMD patients underwent MRI including structural- and sequential multi-delay arterial spin labeling acquisitions before and after ACZ injection. Non-linear least squares curve fitting was performed on CBF values extracted from vascular regions. Root mean squared error (RMSE), absolute and relative CVR (CVRdiff and CVRrel) and time to maximum CBF response (tmax) were compared between the anterior circulation (affected) and posterior circulation (unaffected). Results: Twenty-one patients were evaluated (age range 10-61; 16 females), three was excluded due to technical issues. Median RMSE ranged between 2.8 and 3.8 ml/100 g/min. CVRrel was lower in anterior regions compared to posterior regions (mean differences 10.4-29.2%), whereas, CVRdiff differences were in general similar (0.8-6.9 ml/100 g/min). Median difference in tmax of 2.3-6.8 min was observed in affected compared to unaffected regions. Reversible intracerebral steal was found in two patients and successfully modelled using an extended gamma-variate function. Conclusion: This parameterized gamma-variate function framework provides physiologically interpretable parameters for modelling the CBF response to ACZ in MMD.