Magnetic Resonance Imaging of Stroke

Introduction

Neuroimaging methods have become indispensable tools for diagnosis of tissue status and aid in treatment decision making after stroke. Especially magnetic resonance imaging (MRI), which allows noninvasive and longitudinal measurement of multiple (early) biomarkers of brain tissue injury, can provide important insights in stroke lesion development. In addition to its diagnostic potential in the clinic, MRI offers a valuable method for in vivo studies on stroke pathophysiology, treatment, and recovery in experimental animal models.

For the theory and principles behind MRI, we refer to a large variety of available textbooks and websites that describe (the basics of) MR physics, contrast mechanisms, and pulse sequences. In brief, MRI takes advantage of the natural abundance of protons and their intrinsic magnetic properties for the generation of signals that reflect the properties and state of underlying tissue(s); signals that can be employed for the construction of images that contain anatomical or functional information. To that aim, MRI protocols can be sensitized to proton density, MR relaxation times T ₁ , T ₂ , and T ₂ ^∗ , diffusion, perfusion, and flow of water molecules in tissues. In the following paragraphs, we will describe in brief how different MRI protocols can be applied to assess brain injury after stroke. Due to space limitations we can only refer to a few review papers, but these include references to relevant original publications.

T ₁ -, T ₂ -, and T ₂ ∗-Weighted MRI

Brain tissue damage after stroke may be identified with standard proton density MRI and T ₁ -, T ₂ -, or T ₂ ∗-weighted MRI protocols . Higher proton density, and T ₁ and T ₂ values after stroke are typically caused by increased interstitial water or vasogenic edema, which usually occurs at subacute to chronic stages of infarct progression. A clinically frequently applied sensitive method for detection of cerebral ischemic lesions is fluid-attenuated inversion recovery (FLAIR) MRI, which depicts tissue areas with prolonged T ₂ while suppressing MRI signal from cerebrospinal fluid .

Hemorrhages, in which deposits of hemosiderin cause local magnetic field distortions leading to T ₂ ∗ shortening, are detectable (as hypointensities) with gradient-echo T ₂ ∗-weighted MRI or susceptibility-weighted MRI . Bleedings are usually preceded by blood–brain barrier (BBB) disruption, which may be detected with T ₁ -weighted MRI after intravenous injection of paramagnetic gadolinium (Gd)-containing contrast agent . Tissue accumulation of paramagnetic contrast agent after leakage across the BBB leads to local T ₁ shortening, giving rise to an increase in T ₁ -weighted signal intensity. BBB permeability can be quantified with dynamic contrast-enhanced (DCE) MRI and tracer kinetic analysis, which enables calculation of a blood-to-brain transfer constant (K _trans or K _i ) . Accumulation of intravenously injected contrast agent in brain parenchyma early after stroke has been found to be predictive of subsequent hemorrhagic transformation, conceivably as a result of (thrombolysis-induced) reperfusion injury .

Dynamic acquisition of T ₂ ⁽ ∗ ⁾ -weighted MRI sensitized to paramagnetic deoxygenated hemoglobin, i.e., blood oxygenation-level dependent (BOLD) MRI, allows measurement of hemodynamic activity, for example, linked to cerebrovascular reactivity or neuronal activity—the latter forms the basis of functional MRI (fMRI) . Cerebral autoregulation can be assessed with BOLD MRI in combination with a vasodilatory challenge, induced by injection of the vasodilator acetazolamide or by CO ₂ inhalation. This normally leads to increased BOLD MR signal intensity, which may be reduced or absent in tissue affected by stroke . With BOLD fMRI of the hemodynamic response to task-related neuronal activation (resulting in increase in local blood oxygenation) or “resting-state” neuronal activity one can obtain information on the poststroke brain’s functional state and connectivity (although the BOLD response may be significantly affected by altered vascular reactivity or neurovascular coupling after stroke). The application of fMRI to assess brain (re)organization, by measuring (changes in) spatiotemporal patterns of neuronal signals/responses, in relation to stroke recovery is a growing area of research.