Monitoring Brain Repair in Stroke Using Advanced Magnetic Resonance Imaging
2012; Lippincott Williams & Wilkins; Volume: 43; Issue: 11 Linguagem: Inglês
10.1161/strokeaha.111.649244
ISSN1524-4628
AutoresL. Sztriha, Ruth Tuura, Michel Modo, Gareth J. Barker, Steven Williams, Lalit Kalra,
Tópico(s)Functional Brain Connectivity Studies
ResumoHomeStrokeVol. 43, No. 11Monitoring Brain Repair in Stroke Using Advanced Magnetic Resonance Imaging Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissionsDownload Articles + Supplements ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toSupplemental MaterialFree AccessResearch ArticlePDF/EPUBMonitoring Brain Repair in Stroke Using Advanced Magnetic Resonance Imaging Laszlo K. Sztriha, PhD, Ruth L. O'Gorman, PhD, Michel Modo, PhD, Gareth J. Barker, PhD, Steven C.R. Williams, PhD and Lalit Kalra, PhD Laszlo K. SztrihaLaszlo K. Sztriha From the Departments of Clinical Neuroscience (L.K.S., L.K.) and Neuroimaging (G.B., S.W.), Institute of Psychiatry, King's College London, London, UK; MR Center, University Children's Hospital, Zurich, Switzerland (R.O.G.); Department of Radiology & McGowan Centre for Regenerative Medicine, University of Pittsburgh Medical Centre, Pittsburgh, PA (M.M.). , Ruth L. O'GormanRuth L. O'Gorman From the Departments of Clinical Neuroscience (L.K.S., L.K.) and Neuroimaging (G.B., S.W.), Institute of Psychiatry, King's College London, London, UK; MR Center, University Children's Hospital, Zurich, Switzerland (R.O.G.); Department of Radiology & McGowan Centre for Regenerative Medicine, University of Pittsburgh Medical Centre, Pittsburgh, PA (M.M.). , Michel ModoMichel Modo From the Departments of Clinical Neuroscience (L.K.S., L.K.) and Neuroimaging (G.B., S.W.), Institute of Psychiatry, King's College London, London, UK; MR Center, University Children's Hospital, Zurich, Switzerland (R.O.G.); Department of Radiology & McGowan Centre for Regenerative Medicine, University of Pittsburgh Medical Centre, Pittsburgh, PA (M.M.). , Gareth J. BarkerGareth J. Barker From the Departments of Clinical Neuroscience (L.K.S., L.K.) and Neuroimaging (G.B., S.W.), Institute of Psychiatry, King's College London, London, UK; MR Center, University Children's Hospital, Zurich, Switzerland (R.O.G.); Department of Radiology & McGowan Centre for Regenerative Medicine, University of Pittsburgh Medical Centre, Pittsburgh, PA (M.M.). , Steven C.R. WilliamsSteven C.R. Williams From the Departments of Clinical Neuroscience (L.K.S., L.K.) and Neuroimaging (G.B., S.W.), Institute of Psychiatry, King's College London, London, UK; MR Center, University Children's Hospital, Zurich, Switzerland (R.O.G.); Department of Radiology & McGowan Centre for Regenerative Medicine, University of Pittsburgh Medical Centre, Pittsburgh, PA (M.M.). and Lalit KalraLalit Kalra From the Departments of Clinical Neuroscience (L.K.S., L.K.) and Neuroimaging (G.B., S.W.), Institute of Psychiatry, King's College London, London, UK; MR Center, University Children's Hospital, Zurich, Switzerland (R.O.G.); Department of Radiology & McGowan Centre for Regenerative Medicine, University of Pittsburgh Medical Centre, Pittsburgh, PA (M.M.). Originally published25 Sep 2012https://doi.org/10.1161/STROKEAHA.111.649244Stroke. 2012;43:3124–3131Other version(s) of this articleYou are viewing the most recent version of this article. Previous versions: January 1, 2012: Previous Version 1 IntroductionThrombolysis and endovascular interventions have revolutionized stroke treatment, but many patients are excluded from such therapies, and residual disability is common.1 Emerging approaches to enhance poststroke brain repair may have no time constraints and are applicable to most stroke patients. Novel interventions to enhance brain repair include electromagnetic or robotic techniques, brain–computer interface, and restorative cell-based and pharmacological therapies.2–4 A major impediment to translation to patient care, however, is the lack of robust in vivo techniques to monitor the effects of such interventions in humans.3Noninvasive imaging of the human brain for multiparametric in vivo monitoring of poststroke recovery presents challenges. The clinical application of certain techniques such as positron emission tomography is frequently restricted by radiation exposure, limited resolution, high cost, or difficult access.5–7 Magnetic resonance imaging (MRI), however, is accessible, noninvasive, safe, and versatile, with high resolution, making this an ideal modality for multiparametric in vivo monitoring of stroke recovery. This review concentrates on MRI markers of stroke recovery in experimental models and, when available, in humans (Table).Table. Magnetic Resonance Options for Imaging Poststroke RecoveryTechniqueNeed for Exogenous Contrast AgentParameter MeasuredAvailability on Current Scanner PlatformAngiogenesisDSC-MRIYesCBF, CBVYesASLNoCBFYesSSCE-MRIYesAverage vessel size, MVDNoDCE-MRIYesKiNoGradient/spin echo BOLDNoDeoxygenated hemoglobinYesSWINoDeoxygenated hemoglobinYesCellsMRSNoConcentrations of metabolitesYesCell labelingYesProperties of contrast in cellsNoAxonal remodelingDTINoADC, FA, Q-space–based, and HARDI measurementsYesMEMRIYesParamagnetic manganese in neural networksNoBrain volumeT1-weighted volumetryNoRegional gray matter volumeYesADC indicates apparent diffusion coefficient; ASL, arterial spin labelling; BOLD, blood oxygenation level–dependent imaging; CBF, cerebral blood flow; CBV, cerebral blood volume; DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; DSC-MRI, dynamic susceptibility contrast-enhanced magnetic resonance imaging; DTI, diffusion tensor imaging; FA, fractional anisotropy; HARDI, high-angular resolution diffusion imaging; Ki, blood–brain barrier transfer constant; MEMRI, manganese-enhanced magnetic resonance imaging; MRI, magnetic resonance imaging; MRS, magnetic resonance spectroscopy; MVD, microvascular density; SSCE-MRI, steady-state susceptibility contrast-enhanced magnetic resonance imaging; SWI, susceptibility-weighted imaging.Biological Substrates of Neural RepairAngiogenesisThe peri-infarct cortex is a unique neurovascular niche, within which angiogenesis is closely and causally linked to neurogenesis through vascular growth factors and chemokines.2 Together with parenchymal astrocytes, angiogenic vessels facilitate synaptogenesis and axonal sprouting.2 Angiogenesis stimulated by cell-based or pharmacological interventions correlates with improved behavioral outcome.2 In rodents, capillary sprouting at the ischemic boundary leads to new vessel development between 2 and 28 days.8 Angiogenesis has been observed in the ischemic penumbra of humans 3 to 4 days after stroke, and higher cerebral blood vessel density has been associated with improved survival.9 Angiogenic vessels are permeable during the early stages of development and become less leaky as they mature, potentially allowing new vessels to be distinguished by imaging techniques.10NeurogenesisIn experimental stroke, focal ischemia increases neurogenesis in distinct parts of the ipsilateral subependymal zone, and neuroblasts migrate from the subependymal zone to the ischemic boundary regions, where they exhibit phenotypes of mature neurons (Figure 1).11–13 It has been shown in rat models that migration of immature cells from the subependymal zone occurs over the first 2 weeks but can be sustained for several months.14 Neurogenesis after stroke has been associated with functional recovery in mice.15 In humans, postmortem studies have demonstrated active proliferation of astrocytes, neuroblasts, migrating neuroblasts, and immature neurons in the ipsilateral subependymal zone at 10 days after stroke.16–18 Cells expressing markers of newly born neurons have been detected in the ischemic penumbra of humans, where they cluster around blood vessels.16,19Download figureDownload PowerPointFigure 1. Imaging of transplanted and endogenous neural stem cells (NSC). A, Mouse NSC labeled with a bimodal gadolinium-rhodamine (GRID) contrast agent can be detected both by magnetic resonance imaging (MRI) and fluorescence histology. The overlay of both images confirms good correspondence between the areas of transplanted cells on both imaging modalities. The blue arrows indicate the area of CA1 with transplanted cells, whereas the yellow arrows point to the injection tract. B, In rats with global ischemic damage (15 minutes of 4-vessel occlusion), transplanted NSC (in red because of rhodamine chelate from contrast agent) migrate into the CA1 cell layer of the damaged hippocampus (NeuN+ neurons marked in green), corresponding to the hippocampal region on the magnetic resonance image (blue arrows). Detection of endogenous neurogenesis, however, constitutes a major challenge because only a small proportion of cells within the dentate gyrus (white arrows) are proliferating neural stem cells or newly generated neurons. C, A similar challenge lies in the in vivo detection of neurogenesis in the subependymal zone (SEZ; white arrows) next to the lateral ventricle. The SEZ is a very dense cellular region as can be seen here with DAPI (blue) labeling of every cell nucleus, delineating a thin line along the lateral ventricle. Although this zone is normally only a few cells deep, some parts of SEZ expand because of a stroke (area delineated with orange line), and remaining within ≈100 µm in thickness. In vivo ventricular motion as well as the dramatic variation in tissue type will be major challenges in the detection metabolic changes within this specific region. However, cells migrating into the peri-infarct area (astrocytes stained with GFAP are visible here in green) potentially can affect the metabolite profile of this area, although again the numbers infiltrating this area are relatively small compared with the overall volume of the peri-infarct area. Neuronal cells (NeuN in red) are dramatically decreased even in the peri-infract area, and very little cell replacement occurs because of endogenous neurogenesis.Axonal RemodelingIt has been shown in rats that neurons in the surviving peri-infarct cortex are induced to express a growth-associated genetic profile that enhances axonal sprouting and mediates the formation of new connections.20 In experimental stroke, axonal sprouting occurs in a growth-permissive zone outside the glial scar within the peri-infarct cortex, and also up to several millimeters away from the infarct.21 Significantly increased axonal connections, originating from the peri-infarct motor cortex, have been demonstrated during recovery in rodent and primate models, correlating well with behavioral outcome.22,23Imaging of Biological SubstratesAngiogenesisAlthough direct depiction of microvessels with magnetic resonance angiography is not feasible because of its relatively low spatial resolution, there are many MRI parameters that can be used to monitor angiogenesis and vascular remodeling after stroke indirectly (Table).24,25Hemodynamic parameters affected by angiogenesis, such as cerebral blood volume (CBV) and cerebral blood flow (CBF), can be measured with perfusion techniques such as dynamic susceptibility contrast (DSC)-enhanced MRI or arterial spin labeling (ASL).25 DSC-MRI calculates cerebral hemodynamic parameters from the time course of signal changes induced by the first passage of an intravenously injected paramagnetic contrast agent.26 DSC-MRI has demonstrated significantly increased CBV in the recovering ipsilesional cortex of rodent stroke models, which correlated well with increased vascular density on histology.27,28 The ASL method applies radiofrequency pulses to alter the magnetization of water protons in arterial blood, thereby generating an endogenous intravascular tracer.29 Using ASL, a significantly enhanced CBF has been demonstrated in the perilesional cortex in rat models that correlated with vascular density.28,30The low specificity of DSC-MRI or ASL for angiogenesis may be overcome by steady-state susceptibility contrast-enhanced MRI, which allows estimation of blood volume, vessel size, and microvessel density.25 The ratio of changes in gradient-echo to spin-echo relaxation rate (ΔR2*/ΔR2) induced by a high-molecular-weight intravascular contrast agent provides an estimate of average vessel size after accounting for echo time, contrast concentration, and the main magnetic field.31 The ratio Q≡ΔR2/(ΔR2*)2/3 characterizes microvascular density.32 In a rat model of transient cerebral ischemia, steady-state susceptibility contrast-enhanced MRI showed decrease in vascular density and increase in vessel size attributable to strengthened collateral circulation in the reperfused cortex at days 1 and 3, confirmed on histology, followed by an increase in vascular density likely attributable to angiogenesis at days 14 and 21.33Increased permeability of the blood—brain barrier around newly formed vessels can be assessed in humans and animal models with dynamic contrast-enhanced MRI, which measures the time course of contrast-induced changes in the T1 relaxation time constant attributable to contrast agent extravasation into tissue.34 The blood—brain barrier transfer constant, Ki, can be estimated by locally assessing the dynamics of the signal intensity of a time series of T1 weighted magnetic resonance images with tracer kinetic models.35 In a rat stroke model, Ki in the ischemic boundary correlated with histologically confirmed angiogenesis after cell-based treatment.27 The increase of Ki, peaking at 2 to 3 weeks, colocalized with elevated CBF and CBV at 6 weeks after cell therapy. In another study, Ki elevation was detected 1 to 3 weeks after stroke in sildenafil-treated animals compared with 1 to 6 weeks in controls, suggesting that therapy accelerates angiogenesis, and this is measurable with imaging.36Newly developed vasculature rich in venous blood can be detected with blood oxygenation level–dependent MRI.36 This exploits the local magnetic field disturbances induced by the relatively high magnetic susceptibility of deoxygenated hemoglobin, which provides T2 contrast on spin-echo images and a larger T2* contrast on gradient echo images.37 Susceptibility-weighted imaging, a variant of T2*-weighted MRI that includes phase information, can further enhance the contrast of the areas with increased magnetic susceptibility.24,38 In a rat model of stroke, perilesional angiogenesis was successfully identified on T2* and susceptibility-weighted imaging maps.36 These techniques also detected an accelerated angiogenetic profile with sildenafil, and regions exhibiting T2* shortening spatially matched with those with elevated Ki.36 T2* shortening in the ischemic hemisphere spatiotemporally corresponded with increased CBF on ASL in rats.30MRI can be used to measure hemodynamic parameters, microvascular density, blood–brain barrier permeability, and deoxygenated hemoglobin, all of which are potential markers of angiogenesis and vascular remodeling. However, MRI is not without limitations and MRI techniques differ in their availability and suitability for clincal use. MRI measurement algorithms also are based on relatively basic biophysical and mathematical models that may be inaccurate under complex or altered conditions seen with stroke.25DSC-MRI is arguably the most widely available perfusion MRI technique in the clinical setting, but DSC-MRI methods are not fully quantitative, and thus provide information with regard to the relative CBF and CBV.39 Blood oxygenation level–dependent methods also are widely available and easily implemented but are again able to provide only a qualitative assessment of deoxyhemoglobin concentration or T2*. ASL methods are arguably more quantitative and in recent years have become increasingly available, although most ASL methods do not provide a measure of CBV.39 Both DSC-MRI and ASL are sensitive to errors arising from an increase in arterial transit time to the peri-infarct tissue, and DSC-MRI perfusion values also can be confounded by dispersion of the tracer input function or T1 changes from blood–brain barrier disruption.39 In addition, whereas ASL and DSC-MRI offer good sensitivity to perfusion changes during stroke recovery, alterations in CBF and CBV may reflect vasodilation or metabolic changes (Figure 2), so these perfusion measures may not offer a high specificity for angiogenesis per se. Similarly, an increase in Ki (from dynamic contrast-enhanced MRI) or a shortening of T2* may arise because of other vascular pathology.25 Steady-state susceptibility contrast-enhanced MRI measures of microvascular density and vessel size arguably offer the greatest specificity for angiogenesis but currently suffer from limited availability, especially within the clinical setting.Download figureDownload PowerPointFigure 2. Group-average cerebral blood flow (CBF) maps using arterial spin labeling (ASL) in 6 patients with a cortical ischemic stroke in the left middle cerebral artery territory. At 3 weeks after stroke (A), hypoperfusion in the affected hemisphere (arrows) and crossed cerebellar diaschisis (arrowheads) is demonstrated. At 15 weeks, perfusion (B; displayed with the same threshold as in A) is restored in the contralesional cerebellum (arrowheads) and is increased in the contralesional hemisphere (arrows).NeurogenesisThere are inherent limitations of resolution and sensitivity that preclude the direct visualization of endogenous neurogenesis. The unique metabolic properties of newly generated cells, however, can serve as surrogate markers. Magnetic resonance spectroscopy (MRS) potentially allows the in vivomeasurement of marker compounds associated with different cell types. The metabolites most frequently detected in the human brain include choline, creatine, glutamate, glutamine, lactate, myo-inositol, and N-acetyl aspartate (NAA), but there are at least 25 other less common metabolites that can be assessed by MRS.40 Proton (1H) MRS offers better sensitivity compared with other nuclei (phosphorus or carbon) and is more likely to detect minor changes in metabolites.40In vitro nuclear MRS, a correlate of in vivo MRS, can identify cultured neurons and glial and meningeal cells by their metabolic properties.41 Cultured neural stem cells exhibit high levels of choline and myo-inositol, and low concentrations of creatine, a profile different from embryonic stem cells or mature cells of the central nervous system.42 A spectroscopic peak at 1.28 parts per million, found in cultured cells and the rat and human hippocampi, has been reported as being unique to neural progenitor cells, although this specificity subsequently has been questioned.43–47A variety of metabolites in the peri-lesion area can inform on cellular repair mechanisms. Both single-voxel and multiple-voxel MRS techniques have been used to noninvasively measure in vivo alterations of metabolite profile during stroke recovery in both human studies and in animal models. 1H/13C MRS showed significant decrease in NAA, choline, and glutamate/glutamine turnover with increased glutamine and lactate at 24 hours in the perilesional areas of the rat brain, followed by normalization of NAA, choline, glutamine, and glutamate/glutamine turnover at 3 weeks, paralleling neurological improvement.48 These alterations in the metabolite profile indicate substantial recovery of the initially impaired neural and glial functioning in the lesion border zone over time.48 In humans, acute stroke is associated with elevated lactate and reduced NAA in the infarct zone, whereas the peri-infarct regions are characterized by lactate levels similar to the core but with NAA concentrations significantly higher than inside the infarct.49,50 In a longitudinal multivoxel MRS study of 51 stroke patients, NAA concentrations decreased both in infarcted and ipsilesional normal tissue over the first 2 weeks and remained lower than baseline at 3 months, indicating neuronal loss.51 A biphasic lactate peak was thought to represent initial tissue hypoxia followed by infiltration by inflammatory cells with a high rate of anaerobic glycolysis.51 In another study, higher myo-inositol concentrations in the spared ipsilesional and contralesional primary motor cortices of chronic stroke patients were thought to reflect astrocyte-mediated poststroke plasticity rather than gliosis.52MRI contrast labeling offers an alternative approach for in vivo imaging of cellular elements during recovery.5,53,54 Animal studies have used superparamagnetic (iron oxide–based) particles and paramagnetic (gadolinium- or manganese-based) agents, some of which are approved for human use.5 Paramagnetic agents generate a hyperintense signal on T1-weighted MRI, whereas cells harboring even a single micron-size iron oxide particle can be seen as hypointense spots on T2 and T2*-weighted MRI, enabling precise anatomic localization and tracking of cell migration.55 Neural stem cells labeled in vitro successfully have been detected in vivo in animal stroke models by means of MRI after transplantation (Figure 1).5,56,57 A patient receiving an intravenous injection of iron-oxide microbead-labeled bone marrow stem cells 2 weeks after an ischemic stroke had development of hypointensity compatible with iron (stem cell) deposition in the infarct, as detected with T2* MRI 4 days after the injection.58 Animal studies also indicate that endogenous neuronal stem/progenitor cells also can be labeled using cerebroventricular injections of MRI contrast agents, but their application in humans may be limited.55 Labeling can alter cellular functional characteristics, and may not allow reliable differentiation between viable and dead cells; uptake of contrast by other cell types such as phagocytes cannot be excluded.5,53,57,59MRI approaches to detect cellular changes during stroke recovery include the identification of progenitor cells by their unique metabolic profile or by labeling with contrast agents. (1H) MRS is available as an option on many commercial (human) MRI scanners, although more interactive optimization is often required than is usual for MRI scanning, and considerable local expertise is necessary to ensure high-quality data. MRS has low sensitivity, which is a major limitation if there are very few cells to detect, as is the case in the peri-lesion area harboring sparsely seeded neural progenitors. The low sensitivity also means that large imaging voxels are required, which leads to partial volume effects and makes longitudinal follow-up more difficult because the relative concentrations of different components within a voxel may change over time. Furthermore, there is no one-to-one mapping of metabolites to cell types, which affects specificity.Axonal Remodeling and Cortical ChangesDiffusion tensor imaging (DTI) delineates anatomic connectivity of white matter pathways and can detect tract disruption.60 DTI provides 2 scalar measures, the apparent diffusion coefficient and fractional anisotropy (FA), which characterize the magnitude of water diffusion and the degree of anisotropy for each voxel (Figure 3).24 In addition, careful interpretation of axial (parallel to the long axis of the fiber) and radial (perpendicular) diffusivity provides measures of axonal and myelination status, respectively.24,61Download figureDownload PowerPointFigure 3. Fractional anisotropy (FA) maps are easily obtainable on current scanners. FA values are decreased in the ipsilesional internal capsule (arrow) in a patient with left hemispheric ischemic stroke.Experimental stroke in rats has been characterized by decreased FA in perilesional areas at 3 days, followed by return to control values after 9 weeks.62 In rat models of stroke, treatment with neural progenitor cells, sildenafil, or erythropoietin has been associated with increased elevation of FA in peri-infarct areas at 5 to 7 weeks, which has correlated with reorganization of axons and myelin on histology, and with improved functional outcome.30,63,64 In humans, a decrease in FA of the ipsilesional corticospinal tract has been detected during the first 3 months after ischemic stroke.65 Early fiber tract degeneration has paralleled impaired recovery66 and has predicted long-term deficits.65 During a period of up to a few years after stroke, human studies (including some using specific interventions such as intonation-based speech therapy or brain–computer interface training)67,68 have observed increasing FA, connectivity and fiber number in lesioned and nonlesioned white matter, correlating with better functional outcome.67–71 DTI evaluation of motor tract integrity in chronic stroke patients can predict motor function and behavioral gains from robotic therapy, motor practice, transcranial direct current, or epidural motor cortex stimulation.72–77DTI methods incorporating a single tensor fit to the diffusion at each voxel show an anomalous overall lowering of FA where white matter fiber tracts cross, despite the presence of highly oriented tissue.24 Newer q-space–based and high-angular resolution diffusion imaging techniques (eg, diffusion spectrum imaging, q-ball, and persistent angular structure MRI) provide model-independent analysis to obtain multiple tensors per voxel and thereby extract information on complex tissue structure, including crossing fiber tracts.60,78–80 These techniques are particularly relevant in detecting randomly oriented crossing axons during early-stage axonal remodeling in stroke recovery.60 In addition, changes in axonal density associated with stroke recovery can be measured using MRI diffusion entropy, a technique not dependent on axonal orientation.60Manganese-enhanced MRI is a novel tool for in vivo assessment of neuronal networks.81 Manganese-enhanced MRI is based on the detection of manganese (Mn2+), which enters active neurons and is transported axonally and trans-synaptically.81 Paramagnetic manganese shortens the T1 relaxation time and increases signal intensity on T1-weighted MRI. In a rat stroke model, decreased manganese enhancement in the ipsilateral sensorimotor network at 2 weeks after stroke was followed by enhanced connectivity on manganese-enhanced MRI and histology at 10 weeks, which correlated with behavioral recovery.82MRI also allows the measurement of cortical thickness and gray matter density or volume.83,84 In chronic stroke patients, atrophy of gray matter in remote brain regions (likely a consequence of impaired connectivity) predicted a lesser extent of motor improvement from subsequent constraint-induced movement therapy or epidural motor cortex stimulation.72,85 Increases in gray matter have been demonstrated in compensatory brain regions of chronic stroke patients86 and after constraint-induced movement therapy in parallel with functional improvements.87 The mechanism of this increase in gray matter is likely to be multifactorial, involving both neural and non-neural elements.86In summary, DTI and manganese-enhanced MRI are potent in vivo techniques to study white matter structure, and MRI also provides tools to measure dimensions of gray matter. DTI is available either as standard or as an option on most commercial MRI scanners that also routinely provide volumetric T1-weighted measurements. However, there are potential limitations. The interpretation of tractography data requires experience and knowledge of white matter anatomy. Complex analytic methods are computationally demanding and require longer processing times.88 Several tractography algorithms have been described, but there is no consensus on the best method to use. Only few studies have validated tractography with dissection or histology, and knowledge of the relationship of magnetic resonance measures and specific axonal events is limited. DTI cannot differentiate anterograde from retrograde connections or determine whether a pathway is functional.88 Manganese is neurotoxic at high concentrations or after chronic exposure, which limits manganese-enhanced MRI-based neuronal tracing in human studies.89General Considerations for HumanApplication and Future DirectionsNoninvasive MRI biomarkers may provide deeper insights into specific neural events underlying stroke recovery. Furthermore, they may prove helpful to estimate prognosis and to identify patients who may benefit from specific interventions.90,91 The use of MRI biomarkers as surrogate end points may facilitate the screening of novel therapeutic interventions before large clinical trials. Considerations for surrogate markers should include feasibility of translation from animal to human studies, causal role in the disease process, and capacity to capture the net effect of therapy on clinical outcome.91 MRI may not be the best brain-mapping tool for every aspect of stroke recovery. Positron emission tomography has a higher potential to characterize regional brain metabolism and receptor and transporter distribution and density;6 transcranial magnetic stimulation has advantages in evaluating the physiological properties of neural populations,92 whereas magnetoencephalograpy offers superior temporal resolution.93The use of MRI to monitor recovery in humans presents additional challenges. There are well-known contraindications to MRI, and the administration of contrast agents may cause complications, particularly if administered on multiple occasions.94,95 The most feasible clinical research protocols for the study of large patient populations in a longitudinal fashion would use noncontrast techniques that are available on current scanner platforms (Table), keeping in mind that sessions >60 minutes are unlikely to be easily tolerated by most stroke patients. Imaging biomarkers are time-dependent; some may be most useful to monitor recovery early after stroke (eg, CBF),25 whereas others may have limited value in the very early phase (eg, FA may be transiently elevated on DTI as a consequence of edema).96 The selection of regions of interest for quantitative MRS, perfusion, or diffusion measurements requires hypotheses about the areas likely to be involved in poststroke recovery, and focal abnormalities are often much easier to interpret if da
Referência(s)