| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 From the Department of Radiology, Massachusetts General Hospital, Gray 2, Rm B285, 55 Fruit St, Boston, MA 02114-2696. Received September 12, 2000; revision requested November 1; revision received March 30, 2001; accepted May 14. Supported by grants NS3462, RR13213, CA83159, and Public Health Service grant R01NS8477-01. Address correspondence to P.E.G. (e-mail: ellen@nmr.mgh.harvard.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Three hundred magnetic resonance (MR) imaging studies were selected at random from a database of 7,147 examinations to determine the probability of a pathologically decreased ADC. In cases with decreased ADC, the clinical history was recorded and, if available, follow-up MR imaging findings were evaluated. Five cases of ADC reversal became known during the same period and were evaluated to determine the initial ADC decrease, clinical outcome, and findings at follow-up imaging.
RESULTS: Findings in 116 of 300 MR imaging studies revealed regions of decreased ADC. In 49 of 116 studies, follow-up MR imaging examinations were performed at least 4 weeks after the onset of symptoms; ADC did not reverse. Five cases of ADC reversal were identified in the same period, giving an estimated 0.2%–0.4% probability of ADC reversal. Clinical settings were venous sinus thrombosis and seizure (n = 3), hemiplegic migraine (n = 1), and hyperacute arterial infarction (n = 1). Both white matter (n = 3) and gray matter (n = 3) regions were involved.
CONCLUSION: Reversal of ADC lesions is rare, occurs in complicated clinical settings, and can involve white or gray matter.
Index terms: Brain, infarction, 10.78 • Magnetic resonance (MR), diffusion study, 10.12144
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
There have been a few isolated cases of ADC reversal (10–12), but these often are in the context of aggressive treatment, and the frequency of these events is unclear. In other cases, the areas of ADC abnormality were small, and therefore subtle volume loss on follow-up T2-weighted MR images could be missed (12). In some cases, aggressive intervention with intraarterial thrombolysis within 6 hours may result in reversal of the periphery of the ADC abnormality, but in more central portions that initially reverse, infarction may occur later (11).
This more common irreversibility of ADC changes in most clinical studies that follow the natural history of ADC changes in time contrasts with findings in animal studies in which areas of decreased ADC have reverted to normal without areas of residual T2 abnormality or volume loss. In rat models, the ability of cerebral tissue with decreased ADC to recover has been reported with vascular occlusions lasting for as long as 1 hour (13–15) and with changes in ADC less than 0.25 x 10-5 cm2/sec (16). Studies in rat models also have shown significant reduction in infarct size with pharmacologic intervention (17–19), including acute thrombolysis (20).
The data from these animal studies have motivated the push for rapid assessment and treatment of patients with cerebral infarction. The hope is that, with rapid intervention, human studies will reveal areas of decreased ADC that revert to normal without continuing to infarction, or that areas of mismatch between the diffusion and perfusion MR abnormalities will not proceed to infarction.
The purpose of this study was to perform a statistical review of our experience with diffusion-weighted imaging during the past 4 years in more than 7,000 patients, regardless of clinical history, to determine (a) the probability that a region of decreased ADC would return to normal without marked T2 change and with resolution of associated clinical symptoms and (b) the clinical syndromes associated with ADC reversals.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The diffusion data and the low-b-value echo-planar spin-echo T2-weighted MR image were processed off-line to provide trace diffusion-weighted images and ADC maps for clinical interpretation, the details of which have been described elsewhere (2).
Cases defined as "ADC reversals" were only those cases that met the following criteria: initial ADC at least 10% less than normal contralateral white matter; area of increased T2 signal intensity less than 10% of the size of the initial ADC abnormality at follow-up T2-weighted MR imaging; no volume loss at follow-up T2-weighted MR imaging; and clinical symptoms associated with the initial ADC abnormality resolved at follow-up imaging.
The institutional review board provided approval for this study, and patient informed consent was not required.
Initially, a retrospective random sampling of the diffusion archive was performed to estimate the number of images with regions of increased diffusion-weighted signal intensities and decreased ADC values. With a random number generator, our statistician (E.F.H.) selected 300 studies for review. The indication for the image, the official MR imaging interpretation, and the discharge diagnosis were recorded in all 300 studies (J.H.). Images in all 300 studies were retrieved, and the diffusion-weighted images, ADC maps, T2-weighted FSE, and proton density FSE or FLAIR MR images were reviewed by a neuroradiologist (P.E.G.), who was blinded to the clinical symptoms and official reading of the image. The presence of a region with decreased ADC and the neuroradiologist’s imaging diagnosis were recorded for all 300 studies. For studies with decreased ADC values, follow-up studies, if performed, were reviewed by the same neuroradiologist. The follow-up study findings were evaluated for the presence of persistent T2 change or volume loss, with the initial study findings available for comparison. For studies with decreased ADC values, the clinical charts were reviewed (J.H.) to determine the type and time course of the clinical symptoms as well as the discharge diagnosis and method for determining the discharge diagnosis.
Because ADC reversal in humans is unusual, all interpretations of diffusion-weighted images at our institution have been performed in the context of searching for reversible diffusion-weighted imaging lesions. All cases of apparent ADC reversal in our database that have come to the attention of our neurologists, neurosurgeons, or neuroradiologists were collected. Therefore, the second part of this study is a collection of individual cases of ADC reversal. It is important to note that all of these individual cases came from the same database that we randomly sampled in the first part of this article. In all of these cases of ADC reversal, the same neuroradiologist as mentioned before reviewed the MR images, including the T2-weighted FSE and proton density FSE or FLAIR images as well as the diffusion-weighted MR images and ADC maps, to determine the location and T2 character of areas with decreased ADC. The follow-up study findings were evaluated for the presence of persistent T2 change or volume loss, with the initial study findings available for comparison. The ADC data sets for these patients were also analyzed on a computer workstation (Sun Ultra 10; Sun Microsystems, Palo Alto, Calif) by the neuroradiologist with custom-written software to calculate ADC ratios of abnormal regions compared with contralateral normal white matter on both the initial and follow-up images. For each image, a single section where the abnormality was greatest was chosen based on visual inspection. On this section, a region of interest (ROI) was drawn to outline the abnormality, and an ROI of similar size was also drawn on the contralateral white matter for comparison. The clinical charts were reviewed by a physician (J.H.) to determine the type and time course of the clinical symptoms as well as the discharge diagnosis and method for determining the discharge diagnosis.
With the results of random sampling of the database, the number of studies with decreased diffusion and the number of studies with decreased diffusion in which follow-up MR studies were performed were estimated. By using these estimations and the number of studies with ADC reversal found within the same data base, probabilities of ADC reversal in the patient population were estimated. Standard errors were calculated where appropriate.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
On 116 of 300 images, regions of increased diffusion-weighted signal intensity and decreased ADC values were present. The estimated percentage of images with decreased diffusion is therefore 39% with a standard error of 5%.
In 49 of 116 studies, follow-up MR imaging examinations were performed at least 4 weeks after the onset of symptoms. In all of these 49 cases, regions of decreased diffusion were associated with similar regions of increased T2 signal intensity at follow-up imaging. Discharge summaries in all 116 patients indicated that a neurologic deficit was still present at the time of discharge.
The imaging findings recorded in the clinical record at initial and follow-up studies were confirmed by findings at reevaluation of the images in all 116 cases with positive findings. There was agreement in the interpretation of the images between our review and the clinical report in all but one case. In that one study, a small right thalamic focus of increased diffusion-weighted signal intensity was reported as an artifact, but at blinded review of the diffusion-weighted images, ADC maps, and T2-weighted FSE MR images, the findings were thought to represent a small thalamic infarct. Chart review confirmed the presence of clinical symptoms consistent with an acute right thalamic infarct. This study was counted as one of the 116 abnormal diffusion-weighted imaging studies.
In 100 of 116 studies with positive findings, the diagnosis was an infarct on the basis of both imaging findings and clinical criteria. The remaining 16 of 116 studies with images showing increased diffusion-weighted signal intensity and decreased ADC values were associated with brain abscess (n = 2), tumor (n = 8), hemorrhage (n = 5), and traumatic brain injury (n = 1). These diagnoses were determined with cytologic or pathologic findings (abscess, tumors), MR imaging findings (hemorrhage), or history and MR imaging findings (traumatic brain injury). The number of studies in our archive from January 1, 1995, to May 1, 1999, was 7,147. If one assumes that a distribution of abnormalities in the random sample population is similar to that in the entire database, one may estimate that there are 2,764 studies with images showing increased diffusion-weighted signal intensity and decreased ADC.
In 49 of 116 studies in the random sample, follow-up MR imaging study findings confirmed that the initial region with decreased ADC values corresponded with a similar region of tissue injury at the follow-up study. On that basis, we estimate that findings in 1,167 studies in our database would reveal an initial region with decreased ADC values that corresponded to a T2 abnormality at follow-up MR imaging, confirming irreversible injury. During the same period of the random sampling of our database, with routine monitoring, we identified five cases of decreased ADC that did not have tissue damage at follow-up study. Reevaluation of these cases, including ROI analysis, confirmed reversal of ADC abnormality without significant T2 change and resolution of clinical symptoms in four patients. In one additional patient, we confirmed resolution of clinical symptoms without significant T2 abnormality, but follow-up diffusion-weighted images or ADC values were not obtained. The results of the ROI analysis of ADC values in the affected region compared with those in normal contralateral white matter are listed in Table 1.
|
The clinical setting in which ADC reversals were observed included venous sinus thrombosis and seizure (n = 3) (Figs 1, 2), hemiplegic migraine (n = 1) (Fig 3), and acute stroke with intravenous thrombolysis (n = 1) (Fig 4). The locations of the ADC abnormalities that reversed were white matter (n = 3), deep gray nuclei (n = 1), and cortical gray matter (n = 2). One patient had reversible changes in both deep gray nuclei and white matter (patient 4). All white matter areas with reversibly decreased ADC had associated areas of increased T2 signal intensity on the initial image that were similar in size to the associated ADC abnormality. In one case of venous thrombosis and seizures, there was also associated increased T2 signal intensity in the cortex. The deep gray nucleus in one case of venous thrombosis and seizures and the cortex in the case of acute arterial ischemia did not have associated T2 changes. In all cases, the follow-up T2-weighted MR images showed at most minimal residual T2 abnormalities less than 10% of the initial size of the ADC abnormality. Summaries of the clinical histories in these patients are presented in Table 2.
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study was a randomized, blinded, retrospective study combined with five individual cases. When an event occurs rarely, a random sample of a database may not identify any instance of that rare event. In fact, assuming a binomial probability distribution, no instance of a rare event identified in the 300 cases we reviewed indicates that the probability of that rare event occurring is less than 1.2% with a 95% CI. Although it may be argued that the probability of ADC reversal would be better determined with a prospective study, retrospective studies do not appear to result in over- or underestimates (28).
In this study, all cases of ADC reversal included follow-up MR imaging studies. One could, therefore, argue that only the number of those cases with decreased ADC that included follow-up MR imaging studies should be used as the denominator when determining the probability. This is the method used to calculate the probability of 0.4%. However, by making this assumption, underestimation of the number of cases that did not reverse may occur, since patients with fixed deficits often do not undergo follow-up imaging. If patients who did not undergo follow-up MR imaging studies but who did undergo follow-up clinical evaluations to confirm a fixed deficit are included, the probability of reversal becomes 0.2%. One may expect that the actual probability lies somewhere between the two estimates of 0.2% and 0.4%.
As stated in Materials and Methods, the diffusion protocol was changed 1 year after the diffusion database was started, which resulted in a full tensor acquisition and improved signal-to-noise ratio. These improvements are unlikely to significantly alter the results, since regions with ADC decreases greater than 10% can be detected with both sequences.
Review of the literature revealed cases of 21 patients with reversible ADC changes dating back to 1996 (10–12,21–27). These cases were associated with the following conditions: acute stroke or transient ischemic attack, seven cases (11,12,22,26); transient global amnesia, seven cases (25); status epilepticus, four cases (21,27); hemiplegic migraine, two cases (23,26); and venous sinus thrombosis with a seizure immediately after the MR imaging, one case (24). Little history is provided in the case of an 11-year-old girl reported to have reversible ischemic changes in the white matter described by Marks et al (10). The imaging findings are similar to those in the case of hemiplegic migraine in this study and in that described by Chabriat et al (23).
The range of reversible ADC ratios in gray matter was 0.72–0.79 in this study, compared with 0.64–0.73 in the literature. In the case of transient ischemic attack described by Lecouvet et al (22), imaging was performed 4 hours after symptom onset, compared with 2.5 hours for the patient in this study, and the patient of Lecouvet et al had a slightly lower ADC ratio of 0.72 for the involved cortex, compared with 0.79 in the patient in this study. None of the reversible acute ischemic lesions treated with thrombolytic agents described by Kidwell et al (11) met the criteria for reversal in this study, since the follow-up T2 abnormalities were more than 10% of the initial ADC abnormalities. The mean ADC ratio for cortex in the cases of status epilepticus was 0.64 in the study of Lansberg et al (21) and 0.73 in that of Wieshmann et al (27). In the previously reported (24) case of venous thrombosis, gray matter involvement was not described. In the case of venous thrombosis and seizures with gray matter involvement in this study, ADC ratios of 0.77 in the thalamus and 0.72 in the cortex 1 day after the onset of symptoms were calculated.
In this study, ADC ratios in white matter had a larger range of values, from 0.45 to 0.85, and showed more marked decreases in ADC than gray matter ADC ratios. The largest (ADC ratio, 0.45) and smallest (ADC ratio, 0.85) ADC decrease was observed 11 hours and days after symptom onset, respectively. A similar range of values was reported in the three other studies reporting reversible changes in white matter (10,23,24), with ratios of 0.87 at 3 hours, 0.30 at 18 hours, 0.75 at 21 days, and 0.87 at 36 days after onset of symptoms. It is interesting to note that, as in this study, the literature also reports larger reversible reductions in ADC in the white matter than in gray matter.
All three reversible white matter lesions and one of three reversible gray matter lesions in this study were associated with regions of increased T2 signal intensity at presentation. At first, this may seem contrary to a rat arterial ischemia model that showed reversibility of ADC changes only in cortical regions that did not have associated T2 hyperintensity (14,29). However, in this study, the case of cortical ADC reversal associated with an acute ischemic event did not have associated increased T2. Thus, the presence of T2 hyperintensity in association with a low ADC may portent a poor outcome in cases of acute arterial ischemia, but it may not portent a poor outcome in other clinical settings.
Although these cases have been termed reversible, one cannot exclude a degree of subclinical, irreversible cellular injury. In this study, the minimal follow-up of 3 days should be sufficient to exclude a significant volume of delayed cell death, but the subtle areas of T2 prolongation that persisted in some cases suggest that an astroglial response occurred. In cases reported by Lansberg et al (21), subtle volume loss was occasionally observed at follow-up studies.
The causes of decreased ADC values remain to be elucidated, but a number of mechanisms have been proposed, the majority of which are associated with cortical ischemia. These include cellular necrosis (30), loss of adenosine triphosphate pump function (31,32), shifts of fluid from extra to intracellular spaces (33), increased restriction in the extracellular space caused by a reduction in size and increase in tortuosity of the extracelluar space (34,35), and reduction in intracellular ADC (36–38). A recent animal and in vitro study of cortical ischemia provided evidence that molecular motion decreases in both the intra- and extracellular space, but the intracellular compartment is the major contributing factor to the regional decrease in ADC (38). These authors postulated that this intracellular decrease is caused by a decrease in the adenosine triphosphate-fueled cytoplasmic streaming.
It is unclear whether the same mechanisms that cause ADC reduction in cortical ischemia cause ADC reduction in other pathophysiologic processes. For example, ischemic necrosis is associated with a low partial pressure of oxygen, or PO2, and a markedly reduced metabolism, whereas seizures are associated with a normal PO2 and hypermetabolism. Ischemia, status epilepticus, hypoglycemia, and spreading depression are all associated with shifts of water from the extra- to intracellular space. Ischemia, status epilepticus, and hypoglycemia can cause neuronal necrosis, and all three mechanisms appear to involve increased release of excitatory amino acids, calcium influx into cells, lipolysis, and possibly other calcium-triggered events (39). Thus, necrotic change is expected to play a role in the decreased ADC observed in the cortex in these three situations. However, in spreading depression, reductions in ADC are not associated with cell death, and perhaps glutamate release or transient depolarization may be involved in the transient ADC reductions (40). We have also observed reduced ADC values in the basal ganglia, limbic cortex, and white matter of patients with Creutzfeldt-Jakob disease (41). At pathologic examination, myelin vacuolization is seen in these regions and may be the cause of the decreased ADC values.
The cause of the decreased diffusion observed in the white matter is poorly understood. This includes clinical settings such as arterial ischemia, hemiplegic migraine, and venous thrombosis associated with seizures. Currently, we do not understand the mechanisms of ADC decrease, even in cortex, well enough to know whether changes in glial cells markedly contribute to the observed ADC changes. Chabriat et al (23) suggested that in at least some cases of hemiplegic migraine, the observed decreases in white matter ADC may be due to long neuronal depolarizations caused by a genetic abnormality of calcium channel.
In summary, reversible ADC changes in humans are rare but have been found in cases of transient ischemic attack in which imaging was performed within 4 hours, status epilepticus, venous infarction associated with seizures, hemiplegic migraine, and transient global amnesia. In these rare clinical settings, regions of ADC reduction do not portend a poor clinical outcome and do not progress to complete necrosis, but a degree of subclinical irreversible cellular injury cannot be excluded. Radiologists should be aware of these rare instances when a region of decreased ADC may reverse in order to provide the appropriate interpretation of clinical importance when regions of decreased ADC are identified.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, P.E.G.; study concepts, all authors; study design, P.E.G., E.F.H., R.G.G.; literature research, P.E.G.; clinical studies, P.W.S., L.H.S., R.F.B.; data acquisition, P.E.G., J.H., P.W.S., L.H.S., R.F.B.; data analysis/interpretation, P.E.G., J.H., E.F.H., O.W.; statistical analysis, P.E.G., E.F.H.; manuscript preparation, P.E.G., J.H., E.F.H.; manuscript definition of intellectual content, P.E.G., E.F.H., O.W., P.W.S., L.H.S., A.G.S., W.J.K., R.G.G.; manuscript editing, P.E.G., J.H., E.F.H., O.W., P.W.S., L.H.S., A.G.S., W.J.K., R.G.G.; manuscript revision/review and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Prabhakaran, J. Y. Chong, and R. L. Sacco Impact of Abnormal Diffusion-Weighted Imaging Results on Short-term Outcome Following Transient Ischemic Attack Arch Neurol, August 1, 2007; 64(8): 1105 - 1109. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. McKinney, C.L. Truwit, S. Kieffer, P.W. Schaefer, R.G. Gonzalez, M.H. Lev, and W.J. Koroshetz Reversibility of an "Apparent" Infarct on Dynamic Perfusion CT after Lytic Therapy: Comment Regarding Cerebral Blood Flow and Blood Volume Thresholds AJNR Am. J. Neuroradiol., August 1, 2006; 27(7): 1391 - 1393. [Full Text] [PDF]
|
|
|
|
|
|
R.G. Gonzalez Imaging-guided acute ischemic stroke therapy: From "time is brain" to "physiology is brain". AJNR Am. J. Neuroradiol., April 1, 2006; 27(4): 728 - 735. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H Axer, A Ragoschke-Schumm, J Bottcher, C Fitzek, O W Witte, and S Isenmann Initial DWI and ADC imaging may predict outcome in acute disseminated encephalomyelitis: report of two cases of brain stem encephalitis J. Neurol. Neurosurg. Psychiatry, July 1, 2005; 76(7): 996 - 998. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Doherty, S. Jayadev, N. F. Watson, R. S. Konchada, and D. K. Hallam Clinical Implications of Splenium Magnetic Resonance Imaging Signal Changes Arch Neurol, March 1, 2005; 62(3): 433 - 437. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Inatomi, K. Kimura, T. Yonehara, S. Fujioka, and M. Uchino DWI abnormalities and clinical characteristics in TIA patients Neurology, February 10, 2004; 62(3): 376 - 380. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Fiehler, K. Knudsen, T. Kucinski, C. S. Kidwell, J. R. Alger, G. Thomalla, B. Eckert, O. Wittkugel, C. Weiller, H. Zeumer, et al. Predictors of Apparent Diffusion Coefficient Normalization in Stroke Patients Stroke, February 1, 2004; 35(2): 514 - 519. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Bilgili and B. Unal Effect of Region of Interest on Interobserver Variance in Apparent Diffusion Coefficient Measures AJNR Am. J. Neuroradiol., January 1, 2004; 25(1): 108 - 111. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Calistri, F. Caramia, F. Bianco, F. Fattapposta, F. Pauri, and L. Bozzao Visualization of Evolving Status Epilepticus with Diffusion and Perfusion MR Imaging AJNR Am. J. Neuroradiol., April 1, 2003; 24(4): 671 - 673. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. W. Schaefer, G. J. Hunter, J. He, L. M. Hamberg, A. G. Sorensen, L. H. Schwamm, W. J. Koroshetz, and R. G. Gonzalez Predicting Cerebral Ischemic Infarct Volume with Diffusion and Perfusion MR Imaging AJNR Am. J. Neuroradiol., November 1, 2002; 23(10): 1785 - 1794. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Mullins, P. W. Schaefer, A. G. Sorensen, E. F. Halpern, H. Ay, J. He, W. J. Koroshetz, and R. G. Gonzalez CT and Conventional and Diffusion-weighted MR Imaging in Acute Stroke: Study in 691 Patients at Presentation to the Emergency Department Radiology, August 1, 2002; 224(2): 353 - 360. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Wintermark, M. Reichhart, O. Cuisenaire, P. Maeder, J.-P. Thiran, P. Schnyder, J. Bogousslavsky, and R. Meuli Comparison of Admission Perfusion Computed Tomography and Qualitative Diffusion- and Perfusion-Weighted Magnetic Resonance Imaging in Acute Stroke Patients Stroke, August 1, 2002; 33(8): 2025 - 2031. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK |
