Background and Objectives: Early diagnosis of vasospasm after subarachnoid hemorrhage (SAH) can prompt aggressive treatment and improve neurological outcomes. Transcranial Doppler (TCD) is the only diagnosis modality that is noninvasive and available bedside. The purpose of this study was to numerically evaluate the relevance of various blood velocity indices in detection of unbalanced cerebral blood flow due to vasospasm, and to improve the accuracy of diagnosis using TCD.
Methods: We employed a well-validated numerical model of arterial blood flow coupled with a sophisticated intracranial model to generate a cerebral blood flow database. Anterior (MCA and/or ACA) and posterior (PCA and/or basilar arteries) vasospasms were considered under normal and impaired cerebral autoregulation conditions. For each case, mean blood velocities and their ratios between ipsilateral and contralateral, downstream and upstream, and anterior and posterior arteries were monitored during the progress of vasospasm.
Results: Blood velocities at vasospastic arterial segments demonstrated non-monotonic behavior, i.e. the velocities increased initially with mild and moderate vasospasm, however further vasospasm leads to decreasing values. This may lead to false-negative decisions clinically. Blood flow rates, however, decreased monotonically at the affected arteries. Blood velocities upstream of the vasospastic artery decreased in proportion to the blood flow rates (e.g. for MCA vasospasm, 30% and 20% reduction at ICA and CCA). For all vasospasm locations considered, normalization of velocities by upstream and contralateral velocities provided more robust detection. Moreover, the improvements were most compelling in cases with impaired cerebral autoregulation.
Conclusions: The velocity indices and diagnosis strategy proposed in this study can improve the accuracy of TCD diagnosis for cerebral vasospasm. These indices are particularly effective in cases of severe vasospasm where traditional indices (e.g. absolute velocities, Lindegaard index) become problematic.