Diffusion-Weighted Imaging

View larger version:In this page In a new window Download as PowerPoint Slide Case ReportA 30-year-old right-handed woman was transferred to our hospital at 29 weeks of gestation complaining of dyspnea and tachycardia. She presented in preterm labor complicated by gestational diabetes and borderline preeclampsia, with elevated transaminases (SGOT, 114 U/L; SGPT, 122 U/L; alkaline phosphatase, 138 U/L;lactate dehydrogenase, 416 U/L), hyperuricemia (uric acid, 11.9 mg/dL), pulmonary edema, proteinuria (24-hour urinary total protein, 297 mg/L; creatinine, 858 mg),and mild hypertension (blood pressure, 158/108 mm Hg). Her pulmonary edema was treated with furosemide. Tocolytic therapy with terbutaline and magnesium did not arrest labor, and she delivered premature twins by cesarean section within 48hours. That night she experienced severe neck and occipital pain, and the following day she had a generalized, tonic-clonic seizure. Intravenous magnesium therapy was initiated, and a noncontrast enhanced head CT demonstrated multiple 3- to 15-mm hypodensities in the parietal lobes involving gray and white matter and in the left frontal subcortical white matter. She was evaluated by the Massachusetts General Hospital Stroke Service 36 hours after her seizure and was found to be alert and oriented with fluent speech and no cognitive deficits. Her fundi were benign.She had a few beats of gaze-evoked horizontal nystagmus with gaze to either side and a slightly flattened left nasolabial fold. Strength, sensation, and cerebellar function were normal. Deep tendon reflexes were increased in the right upper limb,and a Hoffmann’s reflex was elicited in the right hand. An MRI was obtained on a 1.5-T GE Signa system with echo-planar capabilities provided by Advanced NMR Systems. Fast spin-echo, T2-weighted images (TR, 4200 ms; TE, 104 ms; ETL, 8;FOV, 20×20 cm; matrix, 256×192; slice thickness, 5 mm with 1 mm gap; 1 NEX)(Figure ⇓, panel A) demonstrated extensive T2 hyperintensity involving the gray and subcortical white matter in all lobes with predominance in the temporal and occipital lobes as well as in the frontal and parietal convexities; small T2hyperintense foci were also seen in the cerebellar hemispheres bilaterally. Isotropic diffusion-weighted images (single-shot EPI; diffusion gradients applied in six orthogonal directions with effective gradient amplitude of 14 mT/m; TR, 6000 ms;TE, 108 ms; FOV, 40×20 cm; matrix, 256×128; slice thickness 6 mm with 1 mm gap; 1 NEX; b value of 1221 s/mm ) (Figure ⇓, panel B) demonstrated increased diffusion in most of these regions with four sampled ADCs ranging from 1.2 to 1.8×10 mm /s. Four sampled ADCs in normal subcortical white matter ranged from 0.68 to 0.88×10 mm /s; no areas of decreased diffusion were identified.Repeated physical examination 3 days later revealed new onset of bilateral cerebellar dysmetria, overshoot with leftward saccades, mild truncal ataxia, and pronounced horizontal nystagmus on left lateral gaze. A repeated MRI demonstrated partial decrease in the extent of some of the T2 hyperintense regions with corresponding partial resolution of the abnormally increased diffusion. MR angiography of the neck showed no evidence of vertebral artery dissection.Transthoracic echocardiography was normal. The patient improved rapidly, and within 3 weeks all previously noted neurological abnormalities had resolved.Follow-up MRI at that time demonstrated complete resolution of the T2 signal abnormalities (Figure ⇓, panel C) and of the corresponding regions of increased diffusion (Figure ⇓, panel D); four ADCs sampled in locations similar to those of the initial elevated ADCs returned to normal values and ranged from 0.53 to 0.88×10 mm /s.Figure 1.A, Axial T2-weighted MR image (TR, 4200 ms; TE, 104 ms; ETL,8; FOV, 20×20 cm; matrix,256×192; slice thickness, 5 mm with 1 mm gap; 1 NEX) obtained shortly after onset of eclampsia demonstrates multiple T2hyperintense regions predominantly involving white matter with some involvement of the adjacent cortex. B, Axial isotropic diffusion-weighted image (single-shot EPI; diffusion gradients applied in six orthogonal directions with effective gradient amplitude of 14 mT/m; TR, 6000 ms; TE, 108 ms;FOV, 40×20 cm; matrix, 256×128; slice thickness, 6 mm with 1 mm gap; 1 NEX; b value of 1221 s/mm ) obtained on the same date as panel A at the same level demonstrates low signal (high diffusion) in the regions that are hyperintense on T2-weighted images. C, Axial T2-weighted image (TR, 4200 ms; TE, 104 ms; ETL, 8; FOV, 20×20 cm;matrix, 256×192; slice thickness, 5 mm with 1 mm gap; 1 NEX) at the2−32−32−322same level obtained 3 weeks after initial presentation demonstrates resolution of all of the previously seen T2 hyperintense signal abnormalities. D, Axial isotropic diffusion-weighted image (single-shot EPI; diffusion gradients applied in six orthogonal directions with effective gradient amplitude of 14 mT/m; TR, 6000 ms; TE, 108 ms;FOV, 40×20 cm; matrix, 256×128; slice thickness, 6 mm with 1 mm gap; 1 NEX; b value of 1221 s/mm ) obtained 3 weeks after initial presentation demonstrates normal diffusion in all of the regions where increased diffusion was previously seen.DiscussionThe clinical hallmarks of eclampsia are hypertension, peripheral edema,proteinuria, and seizures. Other neurological findings include headaches, loss of vision, mental status changes, and coma. Previously reported neuroimaging findings include reversible CT hypodense and MRI T2 hyperintense abnormalities seen predominantly in the subcortical white matter and basal ganglia, increased perfusion on SPECT imaging, occasional infarctions, diffuse vasoconstriction on angiography, and intracerebral hemorrhage. While the pathophysiology of eclampsia remains unclear, the clinical, pathological, and neuroimaging findings have led to two major hypotheses. Based on angiographic findings of vasoconstriction, permanent infarctions in the more severely affected patients, and pathological findings of arteriolar vasoconstriction, vessel thrombosis, and infarctions, one proposed mechanism is hypertension-induced vasoconstriction causing ischemia with the development of cytotoxic edema. An alternative view suggests that eclampsia is a form of hypertensive encephalopathy;hypertension induces a loss of autoregulation, which leads to passive arteriolar dilatation, opening of tight junctions, extravasation of macromolecules,and vasogenic edema. This theory is supported by multiple reports of reversible T2hyperintense abnormalities, increased perfusion with SPECT imaging, and animal models demonstrating focal areas of arteriolar dilatation separated by segments of normal vessel.One would predict that the appropriate management of acute hypertension in eclampsia would be different under the two competing hypotheses. If cerebral vasospasm were the underlying pathophysiological mechanism, then the vasospasm would have to be severe enough to cause distal ischemia and cytotoxic edema. Cerebral perfusion to these regions should be further reduced by the acute lowering of mean arterial pressure, and additional ischemic injury should ensue.Treatment then ought to resemble that which is initiated for vasospasm from other etiologies, such as for the vasospasm associated with subarachnoid hemorrhage. In this setting, mean arterial pressure is increased and intracranial pressure is reduced to ensure adequate cerebral perfusion pressure. Several observations do not support this hypothesis in the majority of patients with eclampsia. First, most patients respond well to reductions in mean arterial pressure, and ischemic infarction or permanent neurological injury is rarely seen. Second, some angiographic studies of eclamptic patients with neurological impairment have been unrevealing, indicating that vasospasm of large or medium-sized vessels as seen on angiography is not necessary to cause the clinical syndrome. Third, the degree of ischemia necessary to cause cytotoxic edema (failure of the ATP-dependent ion pumps) results in depression of neuronal cellular function. However, patients with eclampsia often have relatively minor neurological impairment compared with the rather extensive territory of T2 signal abnormality, making it unlikely that these regions are in fact ischemic. For these reasons, it is likely that hypertension-induced vasogenic edema is the underlying pathophysiological mechanism in most eclampsia cases. However, a small percentage of patients do develop infarctions and permanent neurological deficits; in some of these patients, vasospasm has been demonstrated on angiograms.To date, there has not been a reliable method of distinguishing between vasogenic and cytotoxic edema in the living patient at the onset of eclampsia at a time when this knowledge could greatly affect management. On conventional MRI, vasogenic edema is usually marked by T2 hyperintensity predominantly involving white matter, and cytotoxic edema is marked by T2 hyperintensity involving gray matter,white matter, or both. Conventional MRI cannot clearly distinguish between these different types of edema. Echo-planar diffusion-weighted imaging, however, is a new technique that clearly distinguishes between the two. With very strong gradient pulses, images sensitive chiefly to molecular diffusion of water molecules can be generated. Regions with cytotoxic edema demonstrate diffusion coefficients that are decreased compared with those of white matter. In the setting of acute stroke, this is thought to result from decreased Na ,K -ATPase activity in glial cell membranes and consequent decrease in water molecule transport.Conversely, regions with vasogenic edema demonstrate diffusion coefficients that are markedly increased compared with those of normal white matter.28111213141511511131513256713891011261819202122++42324ADCs =apparent diffusion coefficients EPI =echo planar imaging ETL =echo train length FOV =field of view NEX =number of excitations SGOT =serum glutamic-oxaloacetic transaminase SGPT =serum glutamate pyruvate transaminase SPECT=single-photon emissioncomputed tomography TE =echo time TR=repetition time In our patient, routine fast spin-echo, T2-weighted MR images demonstrated multiple regions of extensive T2 signal abnormality involving both white and gray matter. These were interpreted as representing either infarctions or multiple areas of vasogenic edema; we could not readily differentiate between the two. On diffusion-weighted imaging, all T2 hyperintense regions demonstrated markedly increased diffusion, consistent with vasogenic edema; no regions of decreased diffusion to suggest cytotoxic edema were identified. These findings were consistent with the fact that our patient had relatively minor neurological impairment. These initial findings strongly support the pathophysiological mechanism of hypertension-induced vasogenic edema. We were able to confidently lower blood pressure, avoid further invasive evaluation, and defer anticoagulation therapy because we were confident that the patient had not developed thromboembolic infarctions. Follow-up imaging demonstrated complete resolution of the diffusion and T2 abnormalities and confirmed our hypothesis.Our data confirm only that we could accurately identify that the MRI T2 signal abnormality in our patient represents vasogenic edema. Occasionally patients, for unclear reasons, develop infarctions and permanent neurological deficits.Diffusion-weighted imaging is noninvasive, requires 126 seconds to perform, and in the near future will be readily available on most MR scanners. Since this technique accurately distinguishes between vasogenic and cytotoxic edema, we believe that this technique will be an invaluable tool in helping physicians to optimize the treatment of eclamptic patients from the onset; the majority who develop vasogenic edema will undergo blood pressure reduction and supportive measures, while the rare patient who develops infarction will undergo more aggressive therapy. Further prospective evaluation of these patients is needed to test our hypothesis.Selected Abbreviations and AcronymsReceived December 26, 1996.Revision received February 27, 1997.Accepted February 27, 1997.Copyright © 1997 by American Heart AssociationReferences1. Trommer BL, Homer D, Mikhael MA. Cerebral vasospasm and eclampsia.Stroke. 1988;19:326-329. Abstract/FREE Full Text2. Duncan R, Hadley D, Bone I, Symonds EM. Blindness in eclampsia: CT and MR imaging. J Neurol Neurosurg Psychiatry.. 1989;52:899-902.Abstract/FREE Full Text3. Coughlin W, McMurdo S, Reeves T. MR imaging of postpartum cortical blindness. J Comput Assist Tomogr.. 1989;13:572-576. Medline4. Brunberg J, Chenevert T, McKeever P, Ross D, Junck L, Muraszko K, Dauser R,Pipe K, Betley A. In vivo MR determination of water diffusion coefficients and diffusion anisotropy: correlation with structural alteration in gliomas of the cerebral hemispheres. AJNR Am J Neuroradiol.. 1995;16:361-371.Abstract/FREE Full Text5. Lewis L, Hinshaw D, Will A, Hasso A, Thompson J. CT and angiographic correlation of severe neurological disease in toxemia of pregnancy.Neuroradiology. 1988;30:59-64. CrossRef Medline6. Govan A. The pathogenesis of eclamptic lesions. J Pathol Microbiol..1961;24:561-575.7. Sheehan H, Lynch J. Cerebral lesions. In: Sheehan H, Lynch J, eds. Pathology of Toxemia of Pregnancy. Baltimore, Md: Williams & Wilkins; 1973:524-553.8. Hauser RA, Lacey DM, Knight MR. Hypertensive encephalopathy. Arch Neurol..1988;45:1078-1083. CrossRef Medline9. Strandgaard S, Paulson OB. Cerebral autoregulation. Stroke. 1984;15:413-。