Posterior Reversible Encephalopathy Syndrome - AJR

McKinney et al. Posterior Reversible Encephalopathy Syndrome

Neuroradiolog y • Original Research

Posterior Reversible Encephalopathy Syndrome: Incidence of Atypical Regions of Involvement and Imaging Findings Alexander M. McKinney1,2 James Short1 Charles L. Truwit1 Zeke J. McKinney1 Osman S. Kozak1 Karen S. SantaCruz1 Mehmet Teksam1 McKinney AM, Short J, Truwit CL, et al.

Keywords: CNS, contrast enhancement, diffusion, hemorrhage, MRI, posterior reversible encephalopathy syndrome DOI:10.2214/AJR.07.2024 Received February 7, 2007; accepted after revision May 18, 2007. 1Department

of Radiology and Neuroradiology, University of Minnesota Medical Center, Minneapolis, MN.

2Department

of Radiology, Hennepin County Medical Center, 701 Park Ave. S, Minneapolis, MN 55415. Address correspondence to A. M. McKinney ([email protected]). CME This article is available for CME credit. See www.arrs.org for more information.

AJR 2007; 189:904–912 0361–803X/07/1894–904 © American Roentgen Ray Society

904

OBJECTIVE. Posterior reversible encephalopathy syndrome (PRES) is classically characterized as symmetric parietooccipital edema but may occur in other distributions with varying imaging appearances. This study determines the incidence of atypical and typical regions of involvement and unusual imaging manifestations. MATERIALS AND METHODS. Seventy-six patients were eventually included as having confirmed PRES from 111 initially suspected cases, per imaging and clinical follow-up. Two neuroradiologists retrospectively reviewed each MR image. Standard sequences were unenhanced FLAIR and T1- and T2-weighted images in all patients, with diffusion-weighted imaging (n = 75) and contrast-enhanced T1-weighted imaging (n = 69) in most. The regions involved were recorded on the basis of FLAIR findings, and the presence of atypical imaging findings (contrast enhancement, restricted diffusion, hemorrhage) was correlated with the severity (extent) of hyperintensity or mass effect on FLAIR. RESULTS. The incidence of regions of involvement was parietooccipital, 98.7%; posterior frontal, 78.9%; temporal, 68.4%; thalamus, 30.3%; cerebellum, 34.2%; brainstem, 18.4%; and basal ganglia, 11.8%. The incidence of less common manifestations was enhancement, 37.7%; restricted diffusion, 17.3%; hemorrhage, 17.1%; and a newly described unilateral variant, 2.6%. Poor correlation was found between edema severity and enhancement (r = 0.072), restricted diffusion (r = 0.271), hemorrhage (r = 0.267), blood pressure (systolic, r = 0.13; diastolic, r = 0.02). Potentially new PRES causes included contrast-related anaphylaxis and alcohol withdrawal. CONCLUSION. This large series of PRES cases shows that atypical distributions and imaging manifestations of PRES have a higher incidence than commonly perceived, and atypical manifestations do not correlate well with the edema severity. osterior reversible encephalopathy syndrome (PRES) describes a usually reversible neurologic syndrome with a variety of presenting symptoms ranging from headache, altered mental status, seizures, and vision loss to loss of consciousness. The term describes a potentially reversible imaging appearance and symptomatology that is shared by a diverse array of causes, including hypertension, eclampsia and preeclampsia, immunosuppressive medications such as cyclosporine, various antineoplastic agents, severe hypercalcemia, thrombocytopenic syndromes, Henoch-Schönlein purpura, hemolytic uremic syndrome, amyloid angiopathy, systemic lupus erythematosus, and various causes of renal failure [1–9]. Given the multitude of potential offending conditions, some authors suggest that rather than concentrating on new causes of PRES, the focus should be on atypical and potentially

P

misleading imaging findings and common pathophysiology [10]. The mechanism is not entirely understood but is thought to be related to a hyperperfusion state, with blood–brain barrier breakthrough, extravasation of fluid potentially containing blood or macromolecules, and resulting cortical or subcortical edema [11–13]. Alternatively, others have proposed that vasospasm may precipitate the reversible edema, leading to cytotoxic edema if left untreated [14, 15]. The typical imaging findings of PRES are most apparent as hyperintensity on FLAIR images in the parietooccipital and posterior frontal cortical and subcortical white matter; less commonly, the brainstem, basal ganglia, and cerebellum are involved [7, 8, 16–18]. Atypical imaging appearances include contrast enhancement, hemorrhage, and restricted diffusion on MRI [1–6, 9, 19–27].

AJR:189, October 2007

Posterior Reversible Encephalopathy Syndrome

A

B

C

Fig. 1—Mild posterior reversible encephalopathy syndrome (PRES) in 12-year-old girl with seizures who was undergoing immunosuppressive therapy for lung transplantation. A and B, 3-T MR image shows bilateral parietooccipital cortical and subcortical edema on fat-suppressed FLAIR images (A) with avid leptomeningeal or cortical contrast enhancement on T1-weighted image (B) but no diffusion abnormalities (not shown). Tacrolimus was considered causative and was discontinued, and symptoms subsided. C, Edema has resolved on FLAIR image 1 month later.

Given the varying reports of PRES-related imaging findings, we sought to document the frequency of edema in various regions of the brain and of various atypical imaging findings, and to discern new imaging findings or causes. The hypothesis was that atypical regions of involvement and atypical findings may be more common than generally perceived and may not correlate well with the edema severity. Materials and Methods This study was approved by the institutional review boards (via expedited review) of two hospitals: a tertiary care center and a nearby level 1 trauma center. Neuroradiology staff, fellows, and residents on the neuroradiology rotation had placed suspected PRES cases (based on the initial MRI) in a case file at the time the cases were interpreted, over a 9.5-year period between January 1, 1997, and June 1, 2006. To locate cases of PRES or suspected PRES that may not have been placed in this file, the researchers retrospectively reviewed MRI result logs from that period for any additional PRES cases. Eventually, 111 cases of suspected PRES were compiled on the basis of the initial interpretation results. Later, 76 of the initial 111 were confirmed as PRES on the basis of repeat imaging (n = 60) or, in the cases lacking repeat imaging, via clinical data and thorough chart review (n = 16). The presenting symptom and the reason for the MRI examination were recorded for each patient by correlating the radiology request with the acute symptom of presentation that was noted in the online or written chart. These 76 cases were then reviewed by two nonblinded neuroradiologists.


Technique The MRI examinations were performed on multiple units over a number of years; hence, the imaging sequence parameters varied over time. However, the standard protocol included unenhanced axial FLAIR, T1-weighted, and T2-weighted images in all 76 cases eventually confirmed as PRES, with diffusionweighted imaging (DWI) (n = 75) and gadoliniumenhanced T1-weighted imaging (n = 69) in most. Typically, between 10 and 15 mL of IV gadoliniumbased contrast material was administered for the contrast-enhanced examinations. When available, CT and gradient-echo MR images were also reviewed. Sixty of the 76 patients underwent repeat imaging.

Inclusion Criteria Two staff neuroradiologists jointly and retrospectively reviewed the 111 cases, determining via consensus which cases were truly consistent with PRES and eventually confirming 76 cases as PRES. Inclusion as a confirmed case of PRES was based primarily on regression of the findings of suspected PRES on subsequent imaging, when available (n = 60), or on clinical symptom resolution (when repeat imaging was unavailable) via extensive chart reviews (n = 16). Specifically, criteria for inclusion consisted of one of the following: First, initial MRI showed cortical or subcortical FLAIR and T2weighted hyperintensity with posterior predominance that resolved or significantly improved on follow-up MRI or CT. Second, initial MRI showed cortical or subcortical FLAIR or T2-weighted hyperintensity with posterior predominance in a parietooccipital distribution typical of PRES but lacking repeat imaging; these cases without repeat

imaging were considered to be confirmed PRES only if the patient had a complete return to baseline neurologic status. In addition, it was mandatory for inclusion that such patients had received a medication or experienced a condition known to cause PRES that was treated or removed before complete symptom resolution and that the clinician concurred that the symptoms were related to PRES (such agents and conditions are listed in the introduction) [1–9]. Third, initial MRI showed T2weighted or FLAIR hyperintensity in the brainstem, basal ganglia, or subcortical or cortical frontal regions without posterior predominance (atypical distribution), and the imaging findings resolved or significantly improved on follow-up MRI in the setting of a cause previously attributed to PRES. Cases lacking both clinical and imaging followup were excluded. Parietooccipital involvement was not an inclusion criterion because the intent was to detect atypical PRES.

Classification of Edema Severity Based on FLAIR Imaging In light of previous studies indicating that potentially severe cases involve the brainstem or basal ganglia, or have confluent vasogenic edema with mass effect [6, 7], the reviewers classified the eventually confirmed 76 PRES cases as either mild, moderate, or severe on the basis of the extent of hyperintensity on FLAIR imaging and the presence of mass effect. Mild—Mild PRES (Fig. 1) was defined as cortical or subcortical white matter edema without parenchymal hemorrhage, mass effect, herniation, or minimal involvement of only one of group of cerebellum, brainstem, or basal ganglia.

905

McKinney et al.

A

B

C

Fig. 2—Moderate posterior reversible encephalopathy syndrome (PRES) with restriction on diffusion-weighted imaging (DWI). A 51-year-old woman with acute alcohol withdrawal (after 8 days of binge drinking) presented with acute loss of vision. A, MR image shows edema in cortical, subcortical, and deep white matter and in cerebellum and thalami (not shown). B, DWI (upper image) and apparent diffusion coefficient map (lower image) confirmed moderately restricted diffusion focally (arrows) in right parietooccipital vasogenic edema (62.3 vs 85–100 × 10−3 mm2/s in other normal cortical and subcortical white matter locations). Symptoms improved markedly after a few days. C, Follow-up imaging at 2 months shows subtle focal right parietooccipital atrophy in location of previously restricted diffusion, even though follow-up imaging at 5 days (not shown) showed improvement.

A

B

C

Fig. 3—Two patients with severe posterior reversible encephalopathy syndrome (PRES), both related to hypertension. Both cases involved cerebellum. Repeat MRI after several weeks (not shown) showed marked improvement in both patients, with no ensuing encephalomalacia. A and B, Bilateral edema is present on FLAIR MR image in 54-year-old woman, involving brainstem and cerebellum (A) and diffusely involving basal ganglia (B). Diffuse sulcal hyperintensity is presumed to be hemorrhage. C, In another pateint, a 25-year-old woman, bilateral diffuse edema extends from cortex to ventricular margin on FLAIR MR image.

Moderate—Moderate PRES (Fig. 2) was defined as confluent edema extending from the cortex to the deep white matter without extension to the ventricular margin, or mild involvement of two of

906

the group of cerebellum, brainstem, or basal ganglia. Mild mass effect but no herniation or midline shift, particularly if parenchymal hemorrhage was present, was also classified as moderate.

Severe—Severe PRES (Fig. 3) was defined as confluent edema extending from the cortex to the ventricle, or edema or hemorrhage causing midline shift or herniation. Alternatively, involvement of all


Posterior Reversible Encephalopathy Syndrome TABLE 1: Presenting Symptoms and Causes of Posterior Reversible Encephalopathy Syndrome (PRES) in 76 Patients No. of Patients

Presenting Symptoms

58 10 3 3 1 1

Seizures Mental status changesa Visual symptoms or loss Severe headache Aphasia Facial numbness

No. of Patients

Primary PRES Cause CSA for transplant Hypertension Eclampsia Tacrolimus Cocaine, methamphetamine use TTP/ITP Combined chemotherapy Systemic lupus erythematosus Chronic renal failure PEG L-asparaginase Hemolytic uremic syndrome NOS, anaphylaxis? NOS, alcohol withdrawal? NOS, steroids?

34 17 5 4 3 3 2 2 1 1 1 1 1 1

Note—CSA = cyclosporine A, TTP/ITP = thrombotic or idiopathic thrombocytopenic purpura, PEG = polyethylene glycol, NOS = not otherwise specified. a Include acute decrease in consciousness or in responsiveness or acute confusion.

80

No. of Patients

70 60

75 60

50

52

40 30 26

20 14

10

12

5

23

3

0 Parietooccipital Frontal

Temporal Brainstem: Brainstem: Brainstem: Brainstem: Cerebellum Thalamus overall pons midbrain medulla

9

3

Lentiform nucleus or caudate

Putamen

Fig. 4—Bar graph shows regional involvement by posterior reversible encephalopathy syndrome (PRES) in 76 patients. Number of patients in each region are listed in or above each bar.

three of the group of cerebellum, brainstem, and basal ganglia was considered severe.

Localization of Lesions and Description Cases were classified as “yes” or “no” for involvement in the following locations: frontal, temporal, parietooccipital, brainstem, basal ganglia, thalamus, and cerebellum. Basal ganglia involvement was further characterized as to the nuclei involved, and brainstem involvement as to the level. Lesions were also characterized as to whether involving the cortical or subcortical white matter versus deep white matter, and whether there was intracranial hemorrhage, restricted diffusion, contrast enhancement, or unilaterality. Hemorrhage was discerned by a hyperintense sulcal signal on FLAIR or T1-weighted images, CT hyperdensity, or dark signal on gradient-echo images, with description of the abnormalities as parenchymal hemorrhage, subdural, or subarachnoid. All cases were evaluated for signs of venous thrombosis. Cases of en-


hancement were further described as leptomeningeal or cortical, parenchymal, or pachymeningeal. DWI hyperintensity was recorded and apparent diffusion coefficient (ADC) maps were reviewed for cases after 2001 (when ADC map generation became standard); in cases before 2001, the reviewers noted whether the trace image DWI hyperintensity was greater than that on the b = 0 and T2-weighted image, using PACS localization and manual ADC formula calculation [28, 29] in questionable cases: ADC = ln (S0 / S1) / (b1 – b0), where S0 and S1 are the image intensities at b values of b0 and b1, respectively.

Clinical Correlation of Cause, Follow-Up, and Blood Pressure Measurements In all 76 patients, the presenting symptom, PRES cause, and other agents potentially causing PRES were recorded. However, 16 of these patients

lacked repeat imaging but were included on the basis of clinical and imaging findings typical of PRES, with symptom resolution after therapy. Patients without known causes were included only if their repeat imaging was consistent with PRES. Average and maximum systolic and diastolic blood pressures on the day of the MRI were recorded.

Results Of 111 patients initially suspected of having PRES, 35 were excluded by repeat imaging (n = 24), lack of clinical or imaging follow-up (n = 10), or death without autopsy (n = 1). Of the 24 cases not consistent with PRES on repeat imaging, the most common mimicker was subacute hypoxic–ischemic encephalopathy (HIE) with cortical edema lacking clearly restricted diffusion on the initial MRI; these lesions and the symptoms did not resolve on repeat imaging (n = 10). Other mimickers included bilateral subacute posterior infarcts (n = 4), central or extrapontine myelinolysis (n = 4), chronic white matter lesions (n = 4), and reversible chemotherapy-related deep white matter lesions without the cortical or subcortical edema usually seen in PRES (n = 2). The remaining, confirmed 76 PRES patients (40 female, 36 male) ranged from 5 to 80 years old (mean, 33.5 years). Sixty of the 76 were confirmed as having PRES on the basis of marked improvement or resolution on repeat imaging. The remaining 16 were confirmed clinically by an extensive chart review; in these, the mean time to symptom resolution based on clinical examination was 10.2 days after MRI. Seventy-three of the 76 were associated with known offending causes (Table 1), most commonly cyclosporine (n = 34), hypertension (n = 17), or eclampsia (n = 5). Regarding the other three, in one the only known medication was steroids, in another acute alcohol withdrawal occurred with no known medications, and in the third anaphylaxis occurred from iodinated contrast material; these three cases all had bilateral parietooccipital involvement on FLAIR and near-resolution of the hyperintensity on follow-up imaging without atrophy, which is typical of PRES. Regarding the edema extent and severity based on FLAIR, 32 (42.1%) were classified as mild, 27 (35.5%) as moderate, and 17 (22.3%) as severe (Figs. 1–3). The most common presenting symptoms are also listed in Table 1. The regions of brain most commonly involved (Fig. 4) were the parietooccipital (n = 75, 98.7%), frontal (n = 60, 78.9%), and temporal (n = 52, 68.4%). Less common areas of involvement included the cerebellum

907

McKinney et al.

A

B

Fig. 5—Tumefactive posterior reversible encephalopathy syndrome (PRES) in 57-year-old woman with lymphoma who is taking cyclosporine after bone marrow transplant. A, FLAIR MR image shows unilateral left temporal edema and mass effect. No contrast enhancement (not shown) or diffusion restriction was seen, suggesting primary brain tumor. B, One month after cyclosporine cessation, symptoms and imaging are vastly improved.

TABLE 2: Blood Pressure Values (mm Hg) in 76 Patients with Posterior Reversible Encephalopathy Syndrome (PRES) Patient Group Average Blood Pressures

All (n = 76)

With Hypertension or Eclampsia (n = 22)

Taking Cyclosporine A (n = 34)

Systolicaverage

140.8

149.0

131.7

Systolicmaximum

170.8

184.2

159.0

Diastolicaverage

83.2

87.3

80.2

Diastolicmaximum

103.3

112.6

98.6

Note—All values are from same day as presenting MRI.

(n = 26, 34.2%), thalamus (n = 23, 30.3%), brainstem (n = 14, 18.4%), and basal ganglia (lentiform or caudate, n = 9, 11.8%). No cases involved only the orbitofrontal region. Cortical or subcortical white matter edema sparing the deep white matter was present in 54 patients (71.0%), with both cortical and subcortical white matter and deep white matter edema in 22 (29.0%). Only one patient (cocaine usage with severe hypertension) lacked parietooccipital edema, but this patient had severe brainstem, thalamic, and deep white matter edema; this patient dramatically improved and symptoms nearly resolved after undergoing antihypertensive therapy. We considered this a central variant, not an ischemic manifestation of cocaine usage, because no DWI abnormalities were noted on

908

the initial MRI, and the findings nearly resolved on repeat MRI. Another patient with malignant hypertension and predominately brainstem and thalamic involvement occurred but had minimal parietooccipital involvement outside of the brainstem, and hence was counted with the other 75 patients with parietooccipital edema. Also, two tumefactive cases were completely unilateral and simulated neoplasm (Fig. 5), with no enhancement; in both, discontinuation of cyclosporine resulted in dramatic imaging and neurologic improvement. In the 75 patients with PRES on DWI, most (n = 41) had edema isointense to normal-appearing parenchyma on DWI. Twenty-one (28.0%) showed “T2 shine-through” hyperintensity on FLAIR and DWI. Thirteen (17.3%)

had restricted diffusion consisting of small, patchy, or punctate areas much less extensive than the surrounding vasogenic edema. Two of the 13 had mild cortical gyral restriction, with a 10–20% ADC decrease relative to normal areas (Fig. 2); on follow-up imaging, those regions lacked restricted diffusion, the edema resolved, and there was minimal focal atrophy much smaller in size than the initial area of decreased ADC. Only a weak correlation was present between the presence of restricted diffusion and the FLAIR severity (r = 0.271, χ2 = 5.443, p = 0.32). Of note, restricted diffusion from PRES was differentiated from HIE because the DWI abnormalities of HIE were more extensive than on FLAIR (the reverse was true with PRES), cortical contrast enhancement in HIE (when present) simulating enhancing PRES did not resolve on repeat MRI, and there were generally poor, irreversible neurologic outcomes with HIE. Also, the restricted diffusion in PRES was either punctate or focally gyriform, without the multifocality of HIE. Regarding hemorrhagic PRES, there were 13 cases (17.1%): five with parenchymal hemorrhage and 10 with subarachnoid hemorrhage; two had both parenchymal hemorrhage and subarachnoid hemorrhage. Of the eight cases of subarachnoid hemorrhage (without parenchymal hemorrhage) that were initially called subarachnoid hemorrhage on the basis of sulcal hyperintensity on FLAIR, three were confirmed as hemorrhage on CT, T1-weighted imaging, or gradient-echo MRI (Fig. 6). A weak correlation existed between the presence of hemorrhage and the edema extent or severity on FLAIR imaging (r = 0.267, χ2 = 5.415, p = 0.33). Twenty-six of the 69 PRES patients given IV contrast material had abnormal enhancement (37.7%). Most consisted of mild, gyriform, leptomeningeal or cortical enhancement (n = 25) (Fig. 1); deep white matter (n = 1) and dural (n = 2) enhancement also occurred in three cases, simultaneous to leptomeningeal or cortical enhancement. No correlation was noted between the presence of enhancement and the extent or severity of edema on FLAIR (r = 0.072, χ2 = 0.356, p = 0.99). Notably, four PRES patients had known leukodystrophies, two with metachromatic leukodystrophy and two with adrenoleukodystrophy. All four received cyclosporine after bone marrow transplantation. The cortical and subcortical white matter edema was readily discernible from the underlying deep white matter inflammation (Fig. 7).


Posterior Reversible Encephalopathy Syndrome

A

B

C

D

E

F

Fig. 6—Two patients with hemorrhagic posterior reversible encephalopathy syndrome (PRES). A–C, 34-year-old man taking cyclosporine for seizures after bone marrow transplantation has mild edema with small subarachnoid hemorrhage. Initial brain MR image shows small amount of sulcal FLAIR hyperintensity (arrows, A) and mild cortical edema in high frontal and parietal lobes (B). Gradient-echo image confirms small amount of sulcal hemorrhage (arrow, C). D–F, In 44-year-old woman taking cyclosporine, large cerebellar parenchymal hematoma with bilateral edema is seen on CT (D) and gradient-echo MRI (E), with only mild edema on FLAIR image (F) of parietooccipital region.

Regarding blood pressure values, the maximum and average systolic and diastolic blood pressure values on the day of MRI were generally less elevated in cyclosporine-related PRES than in hypertensive or eclamptic PRES (Table 2). Two of the 34 cyclosporinerelated cases had a maximum systolic blood pressure of less than 140 mm Hg and a maximum diastolic blood pressure of less than 85 mm Hg on the day of the MRI. No significant


or weak correlation was noted between the edema severity and the maximum systolic (r = 0.13, p = 0.27) or diastolic (r = 0.02, p = 0.89) blood pressure on the day of MRI. In addition, considering together only the hypertensive or eclamptic patients (n = 22), only weak correlation was noted between the severity of edema and the systolic (r = 0.02, p = 0.94) or diastolic (r = 0.35, p = 0.14) blood pressure.

Discussion The intent of this study was to show both the typical and atypical distributions and manifestations of PRES using a variety of MRI sequences in one of the largest series of PRES cases. In our study, reversible vasogenic edema was almost always present in the cortical or subcortical white matter of the parietooccipital region, the exception being the uncommon central variant of brainstem and basal ganglia

909

McKinney et al.

A

B

Fig. 7—How to discern posterior reversible encephalopathy syndrome (PRES) in setting of underlying leukodystrophy. A and B, 16-year-old girl with metachromatic leukodystrophy underwent MRI with FLAIR imaging (A) just before bone marrow transplantation. Ten days later, mild PRES was noted in frontal and parietooccipital cortical and subcortical white matter related to cyclosporine (arrows, B). Cyclosporine was discontinued, and resolution of PRES was seen on MRI 1 month later (not shown).

involvement, which has been noted in previous reports [2, 7, 8, 16–18]. It is not entirely known why PRES favors the posterior circulation, but this may arise from a relative lack of sympathetic innervation at the level of the arterioles supplied by the vertebrobasilar system compared with the anterior circulation; this innervation presumably protects the brain from marked increases in intravascular pressure, such as with severe hypertension [13, 30, 31]. PRES is not an entirely posterior phenomenon, but rather appears in a gradient-like fashion from posterior to anterior, presumably reflecting the gradient of sympathetic innervation [27, 30–32]. Accordingly, frontal lobe involvement was present in most of our cases (79%), usually in the posterior portion of the superior frontal gyrus (anterior cerebral artery distribution) and the precentral gyrus (middle cerebral); the lentiform or caudate nuclei were involved in 11.8%, usually being supplied by anterior circulation lenticulostriate branches [33]. This distribution confirms that the “posterior” in “posterior reversible encephalopathy syndrome” is a misnomer because most cases involve anterior circulation structures. However, a posterior predominance is certainly seen in each lobe; for example, the orbitofrontal region was spared in all but the most severe cases. Hence, “multifocal,” “posterior dominant,” or simply “revers-

910

ible encephalopathy syndrome” may better apply to this reversible syndrome. We note that central PRES, with brainstem or basal ganglia involvement sparing the parietooccipital region, is rare but certainly occurs, consistent with the low incidence noted in previous case reports; correlation should be made with the blood pressure because these patients usually have extreme hypertension [16–18]. Five such cases were suspected at the outset of this study; one was confirmed as PRES by rapid improvement after antihypertensive therapy, whereas the other four were related to central or extrapontine myelinolysis. A similar-appearing case of PRES in this study had brainstem involvement but was not deemed a central variant because of the minimal presence of parietooccipital edema. Hence, radiologists should be aware that PRES may occasionally present with minimal or no detectable parietooccipital edema. In such cases, it is necessary to exclude other causes of acute brainstem or basal ganglia edema, such as myelinolysis or encephalomyelitis, using a combination of clinical history and follow-up imaging, when necessary. Two cases of completely unilateral involvement occurred (2.6%), which to our knowledge has not been previously described. The MRI differential diagnosis included neoplasm, encephalitis, and inflammatory or infectious leukoencephalopathy; these entities differ in

treatment, may have mass effect, and can undergo biopsy. Hence, we term this variant “tumefacient,” because the severe edema, mass effect, and symptoms dramatically improved with cyclosporine cessation. Another goal was to evaluate how often hemorrhage occurred in PRES, having previously been described in PRES from various causes [3, 4, 20, 26]. The mechanism of hemorrhage is presumed to arise from the phenomenon of breakthrough perfusion, in which the maximally constricted end arterioles cannot further respond to hyperperfusion; this failed autoregulation leads to macromolecule extravasation and possibly hemorrhage into the cortex or subarachnoid space [13]. This occurred in 17.1% of the patients (n = 13), with a weak correlation noted between the presence of hemorrhage and the extent or severity of edema on FLAIR imaging. The most common type was subarachnoid hemorrhage (13.2%); less commonly, parenchymal hemorrhage was noted (6.6%). However, one could criticize the use of FLAIR to detect subarachnoid hemorrhage because nonhemorrhagic exudates may create sulcal hyperintensity [34]. This methodology was used because FLAIR was routinely available in this retrospective study. Regarding this potential criticism, eight (10.5%) of the 13 cases had confirmatory findings of either subarachnoid hemorrhage or parenchymal hemorrhage on CT, T1-weighted imaging, or gradient-echo MRI, regardless of the FLAIR findings. Hence, the hemorrhage frequency is likely between 10.5% and 17.1%, and FLAIR may be suitable to screen for subarachnoid hemorrhage in PRES because it is a routine MRI sequence in PRES evaluation. Also, in patients with subarachnoid hemorrhage who do not have parenchymal hemorrhage, the apparent subarachnoid hemorrhage on FLAIR resolved on repeat imaging as the edema waned. Hence, foci of high sulcal signal on FLAIR (when parenchymal hemorrhage is not present) may ultimately be of little consequence because this finding correlated poorly with extent or severity of edema on FLAIR imaging and usually resolved after removal or treatment of the offending agent. On DWI, the most common appearances in this study were isointensity (54.7%) and DWI bright T2 shine-through (28%), as shown previously [19–22]. Diffusion restriction occurred in a minority (17.3%) of patients; it was usually punctate and surrounded by much larger areas of edema with no ensuing atrophy. The presence of this finding correlated poorly with the FLAIR extent or severity.


Posterior Reversible Encephalopathy Syndrome However, two patients (2.7%) did have a focal gyral configuration of restriction, with mild ensuing atrophy and residual neurologic deficit. In this regard, significant debate has developed in recent literature regarding the various DWI phenomena of PRES. Although lesions are typically isointense related to T2 washout (a balance of T2 effects and increased water diffusibility), T2 shine-through can arise from T2 prolongation being the dominant contributor to signal intensity, noted as ADC and b = 0 DWI hyperintensity greater than the trace DWI. Hypointense lesions may also occur from further increased diffusibility [19, 22]. A conundrum arises in the setting of DWI hyperintensity and “pseudonormal” ADC from intravoxel averaging of focal cytotoxic edema in larger areas of vasogenic edema; these lesions can progress to infarction [6, 23–25]. Several theories address this cytotoxic effect. First, hyperperfusion may cause severe mass effect from vasogenic edema compressing the local microcirculation, with pseudonormal or slightly elevated ADC values surrounded by larger areas of vasogenic edema [6, 12, 23–25]. Second, vasoconstriction may be a response to the edema, eventually causing cytotoxicity if not reversed; accordingly, narrowed intracranial arteries have been noted in hypertensive human and animal models [14, 25, 35–37]. Third, spasm could occur in response to the subarachnoid hemorrhage that uncommonly occurs in PRES. Also, aggressive correction of hypertension may induce ischemia [38]. On the basis of this study, tiny or punctate restricted diffusion foci appear to be unlikely to lead to atrophy, but a larger (> 3 mm), gyral pattern may lead to irreversible insult. Notably, even gyral restricted diffusion partially reversed; repeat MRI in such cases excludes alternative insults such as HIE. The incidence of contrast enhancement in PRES has not been previously well described in a large series; contrast use is only briefly mentioned in large studies of PRES [1, 4, 7, 19–20, 24]. Regarding enhancement, a gyriform pattern has been described [2, 20, 26, 27]. We found a higher frequency (37.7%) than expected from the literature [4, 20, 39]; the exception is a study noting enhancement in 33% of patients [6]. These discrepancies may relate to various factors, including contrast bolus timing, amount, and type, and may also be caused by lack of contrast use in many studies, possibly because contrast-enhanced imaging is not necessary to diagnose PRES. In addition, the degree of enhancement could relate to the


cause; transplantation drugs such as cyclosporine can cause direct endothelial injury and subsequent blood–brain barrier breakdown, potentially differing in pathophysiology from classic hyperperfusion syndromes such as hypertension or eclampsia [40–43]. The lack of elevated blood pressure in several patients in our series supports the possibility that enhancement in the nonhypertensive cases may result from direct endothelial injury causing blood–brain barrier breakdown. In hypertensive cases, the elevated hydrostatic pressure could cause capillary endothelial injury, hyperpermeability, and ultimately enhancement or hemorrhage from blood–brain barrier breakdown, particularly if combined with venous congestion or constriction [4, 44–47]. We found no correlation between the extent on FLAIR and the presence of enhancement, and we conclude that although contrast material is not required to evaluate PRES, it may aid in excluding other causes. Regarding blood pressure the day of the MRI, only weak correlations were found between blood pressure elevation and severity of edema on FLAIR. Blood pressure may even be normal in some cases of PRES, particularly in the setting of chemotherapy, immunosuppressive therapy, or sepsis [48, 49]. A potential criticism regarding our measurements is that the measurement at the time of MRI is not necessarily representative of the initial insult because the findings on MRI could persist for days after the symptom onset. Also, blood pressures can be labile, and our determination of MRI severity could relate to the timing after the symptom onset rather than the intrinsic severity. We note that our institutional standard has been to perform MRI of suspected PRES within 24 hours of admission, but given the retrospective nature of this study and the complexity of the cases, it was difficult to note the exact time of MRI after symptom onset. Two potentially new causes of PRES were identified: anaphylaxis as a reaction to iodinated IV contrast material and alcohol withdrawal. Both lacked a hypertensive history and improved on repeat imaging. In the case of anaphylaxis, PRES is possibly related to the release of endotoxins with resultant endothelial injury; we could not formulate a plausible reason for PRES in the setting of alcohol withdrawal without hypertension. Other medications or illicit drug use could cause PRES but were not noted on laboratory testing or extensive social histories. In conclusion, atypical distributions and imaging manifestations of PRES have a higher

incidence than commonly perceived. Given the common involvement of the regions supplied by the anterior circulation, the term “multifocal,” or simply removing “posterior” from the term “posterior reversible encephalopathy syndrome,” may be more appropriate terminology. The atypical manifestations of contrast enhancement, restricted diffusion, and hemorrhage all correlate poorly with the edema severity on FLAIR MRI and can resolve on repeat imaging with appropriate therapy.

References 1. Hinchey J, Chaves C, Appignani B, et al. A reversible posterior leukoencephalopathy syndrome. N Engl J Med 1996; 334:494–500 2. Schwartz RB, Jones KM, Kalina P, et al. Hypertensive encephalopathy: findings on CT, MR imaging and SPECT imaging in 14 cases. AJR 1992; 159:379–383 3. Schwartz RB, Bravo SM, Klufas RA, et al. Cyclosporine neurotoxicity and its relationship to hypertensive encephalopathy: CT and MR findings in 16 cases. AJR 1995; 165:627–631 4. Schwartz RB, Feske SK, Polak JF, et al. Preeclampsia–eclampsia: clinical and neuroradiographic correlates and insights into the pathogenesis of hypertensive encephalopathy. Radiology 2000; 217:371–376 5. Truwit CL, Denaro CP, Lake JR, et al. MR imaging of reversible cyclosporin A-induced neurotoxicity. Am J Neuroradiol 1991; 12:651–659 6. Covarrubias DJ, Luetmer PH, Campeau NG. Posterior reversible encephalopathy syndrome: prognostic utility of quantitative diffusion-weighted MR images. Am J Neuroradiol 2002; 23:1038–1048 7. Casey SO, Sampaio RC, Michel E, et al. Posterior reversible encephalopathy syndrome: utility of fluid-attenuated inversion recovery MR imaging in the detection of cortical and subcortical lesions. Am J Neuroradiol 2000; 21:1199–1206 8. Jarosz JM, Howlett DC, Cox TC, et al. Cyclosporine-related reversible posterior leukoencephalopathy: MRI. Neuroradiology 1997; 39:711–715 9. Schwartz RB, Mulkern RV, Gudbjartsson H, et al. Diffusion-weighted MR imaging in hypertensive encephalopathy: clues to pathogenesis. Am J Neuroradiol 1998; 19:859–862 10. Dillon WP, Rowley H. The reversible posterior cerebral edema syndrome. Am J Neuroradiol 1998; 19:591 11. Johansson B. The blood–brain barrier and cerebral blood flow in acute hypertension. Acta Med Scand Suppl 1983; 678:107–112 12. Tamaki K, Sadoshima S, Baumbach GL, et al. Evidence that disruption of the blood–brain barrier precedes reduction in cerebral blood flow in hyper-

911

McKinney et al.

13.

14. 15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

tensive encephalopathy. Hypertension 1984; 6[2 Pt 2]:I-75–I-81 MacKenzie ET, Strandgaard S, Graham DI, et al. Effects of acutely induced hypertension in cats on pial arteriolar caliber, local cerebral blood flow, and the blood–brain barrier. Circ Res 1976; 39:33–41 Trommer BL, Homer D, Mikhael MA. Cerebral vasospasm and eclampsia. Stroke 1988; 19:326–329 Ito T, Sakai T, Inagawa S, et al. MR angiography of cerebral vasospasm in preeclampsia. Am J Neuroradiol 1995; 16:1344–1346 Casey SO, Truwit CL. Pontine reversible edema: a newly recognized imaging variant of hypertensive encephalopathy? Am J Neuroradiol 2000; 21:243–245 de Seze J, Mastain B, Stojkovic T, et al. Unusual MR findings of the brain stem in arterial hypertension. Am J Neuroradiol 2000; 21:391–394 Chang GY, Keane JR. Hypertensive brainstem encephalopathy: three cases presenting with severe brainstem edema. Neurology 1999; 53:652–654 Provenzale JM, Petrella JR, Cruz LC Jr, et al. Quantitative assessment of diffusion abnormalities in posterior reversible encephalopathy syndrome. Am J Neuroradiol 2001; 22:1455–1461 Schwartz RB. Hyperperfusion encephalopathies: hypertensive encephalopathy and related conditions. Neurologist 2002; 8:22–34 Schaefer PW, Buonanno FS, Gonzalez RG, et al. Diffusion-weighted imaging discriminates between cytotoxic and vasogenic edema in a patient with eclampsia. Stroke 1997; 28:1082–1085 Casey S. “T2 washout”: an explanation for normal diffusion-weighted images despite abnormal apparent diffusion coefficient maps. Am J Neuroradiol 2001; 22:1450–1455 Koch S, Rabinstein A, Falcone S, et al. Diffusionweighted imaging shows cytotoxic and vasogenic edema in eclampsia. Am J Neuroradiol 2001; 22:1068–1070 Crasto SG, Rizzo L, Sardo P, et al. Reversible encephalopathy syndrome: report of 12 cases with follow-up. Neuroradiology 2004; 46:795–804 Ay H, Buonanno FS, Schaefer PW, et al. Posterior leukoencephalopathy without severe hypertension: utility of diffusion-weighted MRI. Neurology 1998;

51:1369–1376 26. Jones BV, Egelhoff JC, Patterson RJ. Hypertensive encephalopathy in children. Am J Neuroradiol 1997; 18:101–106 27. Lamy C, Oppenheim C, Meder JF, et al. Neuroimaging in posterior reversible encephalopathy syndrome. J Neuroimaging 2004; 14:89–96 28. Stejskal EO, Tanner JE. Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient. J Chem Phys 1965; 42:288–292 29. Schaefer PW, Grant PE, Gonzalez RG. Diffusionweighted MR imaging of the brain. Radiology 2000; 217:331–345 30. Edvinsson L, Owman C, Sjoberg NO. Autonomic nerves, mast cells, and amine receptors in human brain vessels: histochemical and pharmacologic study. Brain Res 1976; 115:377–393 31. Beausang-Linder M, Bill A. Cerebral circulation in acute arterial hypertension: protective effects of sympathetic nervous activity. Acta Physiol Scand 1981; 111:193–199 32. Sundt TM Jr. The cerebral autonomic nervous system: a proposed physiologic function and pathophysiologic response in subarachnoid hemorrhage and in focal cerebral ischemia. Mayo Clin Proc 1973; 48:127–137 33. Osborn AG. Diagnostic cerebral angiography, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 1999:117–151 34. Singer MB, Atlas SW, Drayer BP. Subarachnoid space disease: diagnosis with fluid-attenuated inversion-recovery MR imaging and comparison with gadolinium-enhanced spin-echo MR imaging—blinded reader study. Radiology 1998; 208:417–422 35. Lin JT, Wang SJ, Fuh JL, et al. Prolonged reversible vasospasm in cyclosporin A-induced encephalopathy. Am J Neuroradiol 2003; 24:102–104 36. Byrom FB. The pathogenesis of hypertensive encephalopathy and its relation to the malignant phase of hypertension: experimental evidence from the hypertensive rat. Lancet 1954; 267:201–211 37. Geraghty JJ, Hoch DB, Robert ME, et al. Fatal puerperal cerebral vasospasm and stroke in a young woman. Neurology 1991; 41:1145–1147 38. Casey SO, McKinney A, Teksam M, et al. CT per-

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

fusion imaging in the management of posterior reversible encephalopathy. Neuroradiology 2004; 46:2722–2776 Ugurel MS, Hayakawa M. Implications of postgadolinium MRI results in 13 cases with posterior reversible encephalopathy syndrome. Eur J Radiol 2005; 53:441–449 Wilasrusmee C, Da Silva M, Singh B, et al. Morphological and biochemical effects of immunosuppressive drugs in a capillary tube assay for endothelial dysfunction. Clin Transplant 2003; 17[suppl 9]:6–12 Wilasrusmee C, Da Silva M, Singh B, et al. A new in vitro model to study endothelial injury. J Surg Res 2002; 104:131–136 Benigni A, Morigi M, Perico N, et al. The acute effect of FK506 and cyclosporine on endothelial cell function and renal vascular resistance. Transplantation 1992; 54:775–780 Zoja C, Furci L, Ghilardi F, et al. Cyclosporin-induced endothelial cell injury. Lab Invest 1986; 55:455–462 Friedman SA, Schiff E, Emeis JJ, et al. Biochemical corroboration of endothelial involvement in severe preeclampsia. Am J Obstet Gynecol 1995; 172:202–203 Mushambi MC, Halligan AW, Williamson K. Recent developments in the pathophysiology and management of pre-eclampsia. Br J Anaesth 1996; 76:133–148 McCarthy AL, Woolfson RG, Raju SK, et al. Abnormal endothelial cell function of resistance arteries from women with preeclampsia. Am J Obstet Gynecol 1993; 168:1323–1330 Brubaker LM, Smith JK, Lee YZ, et al. Hemodynamic and permeability changes in posterior reversible encephalopathy syndrome measured by dynamic susceptibility perfusion-weighted MR imaging. Am J Neuroradiol 2005; 26:825–830 Bartynski WS, Zeigler Z, Spearman MP, et al. Etiology of cortical and white matter lesions in cyclosporin-A and FK-506 neurotoxicity. Am J Neuroradiol 2001; 22:1901–1914 Bartynski WS, Boardman JF, Zeigler ZR, et al. Posterior reversible encephalopathy syndrome in infection, sepsis, and shock. Am J Neuroradiol 2006; 27:2179–2190

F O R YO U R I N F O R M AT I O N

This article is available for CME credit. See www.arrs.org for more information.

912
