Abstract 137: Diaschisis Profiles In The Hemodynamic Response To Hemodynamic Stimuli: Implications For Cerebellar Normalization In Patients With Steno-occlusive Disease

Abstract

Introduction: Cerebellar hemispheres are commonly used as global references for internal normalization in hemodynamic analysis, including cerebrovascular reactivity (CVR) experiments. We have reported blood oxygen level-dependent (BOLD) MRI coupled acetazolamide for CVR; however, the influence of crossed cerebellar diaschisis (CCD) is understudied in this setting. Importantly, poor signal-to-noise traditionally limits BOLD-CVR to estimation of arbitrary exam conclusions (CVR end ). We developed a custom computational pipeline for dynamic CVR analysis, permitting unprecedented, dynamic maximal CVR (CVR max ) estimates. We describe CCD in a population of patients with chronic unilateral steno-occlusive disease (uSOD) using CVR max . Methods: 32 acetazolamide-challenge BOLD MRI exams in 23 patients with uSOD were examined using a custom denoising and preprocessing strategy for CVR time-signal analysis. BOLD pre and BOLD post were defined as the average of the respective first and last one-minute BOLD data. CVR end was calculated as %CVR=100*( BOLD post – BOLD pre )/ BOLD pre . CVR max was the maximal, denoised per-voxel value. CVR end and CVR max maps were produced. Cerebellar hemispheres ipsilateral and contralateral to the diseased cerebral hemisphere were visually compared for CCD. Pearson correlations were calculated to study CVR max and CVR end in cerebral and cerebellar hemispheres. Results: CCD could be observed in CVR end and CVR max maps (Figure 1). Significant correlations were observed between healthy cerebral and contralateral cerebellar hemispheres (r=0.5992, p=0.0011 for CVR max , r=0.5034, p=0.0054 for CVR end ). CVR max (r=0.4114, p=0.0215) but not CVR end (r=-0.085, p=0.6611) showed correlations between diseased cerebral and contralateral cerebellar hemispheres. Conclusion: We present novel CVR max estimation from dynamic BOLD-MRI, revealing CCD otherwise unrecognized by conventional BOLD-CVR.