| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Stroke. 2004;35:514.)
© 2004 American Heart Association, Inc.
Original Contributions |
From the Departments of Neuroradiology (J.F., K.K., T.K., B.G., O.W., H.Z.) and Neurology (G.T., C.W., J.R.), University Hospital Eppendorf, University of Hamburg, Hamburg, Germany, and Departments of Neurology (C.S.K.) and Radiology (J.R.A.) and UCLA Stroke Center (C.S.K., J.R.A.), University of California, Los Angeles, Calif.
Correspondence to Jens Fiehler, MD, Department of Neuroradiology, Universitätsklinikum Eppendorf, Martinistr 52, 20246 Hamburg, Germany. E-mail fiehler{at}uke.uni-hamburg.de
| Abstract |
|---|
|
|
|---|
Methods Sixty-eight acute ischemic stroke patients underwent MRI examination (including diffusion/perfusion imaging and MR angiography) within 6 hours (mean, 2.8 hours) after symptom onset, after 24 hours, and again 4 to 7 days later. Lesion volumes with decreased ADC and delayed time to peak in perfusion imaging were determined. In patients showing ADC normalization, volumes with ADC decrease graded as <50%, 50% to 60%, 60% to 70%, and 70% to 80% of the contralateral value were determined by thresholding. Patients were categorized as normalizers (demonstrating ADC normalization in >5 mL tissue with initially decreased ADC) or nonnormalizers (demonstrating ADC normalization in <5 mL tissue).
Results Fourteen patients (19.7%) were classified as normalizers. Eleven of 31 patients (35.5%) initially imaged <3 hours after stroke onset and 3 of 37 (7.5%) of those imaged 3 to 6 hours after onset were normalizers. ADC normalization occurred predominantly in the basal ganglia and white matter after thrombolytic therapy in patients with more distal vessel occlusions. All normalizers demonstrated at least partial tissue reperfusion. Tissue with more severe initial decrease in ADC was less likely to demonstrate normalization.
Conclusions ADC normalization is not a rare event in acute stroke after tissue reperfusion. Brain tissue with initially decreased ADC, especially within 3 hours after stroke onset, may include "tissue at risk."
Key Words: diffusion magnetic resonance imaging stroke, acute thrombolysis
| Introduction |
|---|
|
|
|---|
It is important to fully define the frequency and characteristics of DWI normalization. Recently, a 0.2% to 0.4% probability of apparent diffusion coefficient (ADC) reversal was reported in a patient group with heterogeneous diagnoses.9 There are limited data about the frequency and extent of persistent ADC normalizations in stroke patients,6,10 especially in patients with intravenous thrombolysis. This raises 2 questions: (1) How often does ADC normalization occur in acute stroke patients, and what are the clinical consequences? (2) Are there predictors of ADC normalization?
We studied the frequency, anatomical localization, MRI characteristics, and clinical data of patients who displayed ADC normalization in 68 stroke patients who were initially imaged at the hyperacute stage (mean, 2.8 hours). We sought to identify the clinical and imaging appearance of ADC normalization.
| Patients and Methods |
|---|
|
|
|---|
Multiparametric MRI was performed immediately after clinical evaluation and before possible intravenous thrombolysis with tissue plasminogen activator. The National Institutes of Health Stroke Scale (NIHSS) score was assessed by a stroke neurologist at each imaging time point. Informed consent was obtained from all patients. The study was approved by the local ethics committee. Thirty of the patients were included in a previous study with a different focus.11
Imaging Methods
MRI studies were performed on a 1.5-T clinical whole-body scanner (Magnetom Symphony/Sonata, Siemens) using a standard head coil. The measurements included an axial DWI sequence, a PWI sequence, an MRA, and in most cases a fluid-attenuated inversion recovery (FLAIR) sequence. Table time was <20 minutes in most cases. MRI sequence parameters were described recently.11
Postprocessing
Postprocessing of the DWI and PWI image data was performed offline with custom-written software, including MGH-Image and SCAN developed by the Stroke Center at the University of California, Los Angeles. The ADC was determined from diffusion trace images collected with b=0 and b=1000 on a pixel-by-pixel basis using the Stejskal-Tanner equation. According to the literature,12 minor ADC decreases of <10% cannot be measured reliably with a conventional spin-echo sequence. Therefore, regions with decreased ADC were manually delineated after standardized windowing to a 20% ADC decrease relative to the mean contralateral value (VolumeADCmanual). In patients with ADC normalization, all voxels within the manually delineated ADC volume were graded by thresholding into the following categories: ADC decrease <50%, 50% to 60%, 60% to 70%, and 70% to 80% (relative to the contralateral mean hemispherical value). Lesion volumes were calculated for each categorical range. Additionally, each of these volumes was subdivided into the volume showing ADC normalization and the volume evolving to infarction. For each level of ADC reduction, the proportion of tissue evolving to infarction versus normalization was determined.
For PWI, the changes in T2* were expressed as change in relaxation rate (
R2*, where R2*=1/T2*) and calculated as
R2*(t)=-ln[Si(t)/S0]/TE, where Si(t) is the signal intensity at the time point t after injection of the contrast agent, S0 is the signal intensity without contrast agent, and TE is the echo time. Principles of the indicator dilution theory for nondiffusible tracers were applied to the determined concentration-time curves. Time-to-peak (TTP) maps (time to the maximum of SI changes) were calculated. The volume with a perfusion abnormality was manually delineated as the area with a perfusion abnormality >2 seconds compared with the contralateral hemisphere.
The b=0 DWI images obtained at day 5 to 8 were used as T2-weighted (T2w) images for determination of final lesion volumes. During manual lesion delineation, FLAIR images from days 0 and 5 to 8 were available for comparison to exclude cerebrospinal fluid spaces and preexisting lesions. Furthermore, regions with hemorrhagic transformation or parenchymal hemorrhage, which could be identified on the T2*w PWI data, were included in the final lesion volume, so that false normalizations resulting from hemorrhagic transformation were excluded. Volumes were calculated using voxel count, voxel size, and interslice gap.
Tissue reperfusion was assumed when the TTP delay in a region showing an initial perfusion delay was absent on day 1. Recanalization was assessed from the follow-up MRI study on day 1 on the basis of the PWI and MRA studies according to the modified Thrombolysis in Myocardial Infarction (TIMI) criteria13 for perfusion and vessel status (TIMI 0=no recanalization/reperfusion; TIMI 1=minimal recanalization/reperfusion [<20%]; TIMI 2=incomplete recanalization/reperfusion; TIMI 3=complete recanalization/reperfusion).
Vessel occlusion type was determined from the MRA study as follows: no vessel occlusion, internal carotid artery (ICA) occlusion at origin, carotid T occlusion, proximal middle cerebral artery (MCA) trunk occlusion, distal MCA trunk occlusion, and M2 occlusion.
Statistical analysis was performed with SPSS 10.0. After confirmation of equal variances by Levenes test, Students t test was performed between perfusion parameters and ADC threshold areas. If variances were unequal, the nonparametric Mann-Whitney U test (2 tailed) was used to evaluate significant differences (P<0.05). Correction for multiple testing was applied.
| Results |
|---|
|
|
|---|
Occlusion types were carotid T (n=12), proximal ICA and MCA (n=11), combined anterior cerebral artery and MCA (n=2), proximal MCA trunk (n=19), distal MCA trunk (n=7), MCA branch (n=16), and no vessel occlusion (n=1). Mean initial NIHSS was 13.3±5.3. Patients were treated conservatively (aspirin and low-molecular-weight heparin; n=19), with early hemicraniectomy (n=2), and by intravenous (n=24 within <3 hours, n=20 within 3 to 6 hours) or intra-arterial thrombolysis (n=3 within 3 to 6 hours).
Of the 68 patients, 14 (20.6%) showed a normalization of >5 mL (Figure 1). Mean±SD normalization volume was 16.1±9.8 mL. ADC normalization was time dependent. It was observed in 11 of 31 patients (35.5%) imaged within 3 hours after stroke onset but only in 3 of 37 patients (8.1%) imaged between 3 and 6 hours after stroke onset.
|
The initial total ADC lesion volumes (VolumeADCmanual) were not significantly different between patients with and without ADC normalization (Table 1). ADC normalization was observed in white matter (n=10) and in the basal ganglia (always incomplete; n=4) (Figure 2).
|
|
|
Lesions volumes that showed normalization had the following composition with respect to ADC categories: 6% (<50% ADC), 16% (50% to 60% ADC), 29% (60% to 70% ADC), and 49% (70% to 80% ADC). In contrast, the lesions evolving to infarction were composed by 13%, 25%, 28%, and 35% of the different ADC categories, respectively (Figure 3b). The ratio of tissue volume evolving to normalization compared with infarction increased continuously with increasing ADC for each of the ranges (Figure 3a). Complete reperfusion was present in most of the patients with ADC normalization on MRI follow-up at day 1 (TIMI 3; n=9 of 14). The remaining patients that showed ADC normalization had at least partial reperfusion (TIMI 2; n=4 of 14). In 1 patient with ADC normalization, perfusion MRI results from day 1 could not be obtained because of a technical problem. In the group showing ADC normalization, the rate of complete reperfusion after 24 hours was significantly higher compared with patients without normalization (9 of 14 versus 9 of 54; P<0.05) (Figure 4).
|
|
The initial PWI volume and T2w lesion volume on day 5 to 8 were smaller (each P<0.001) in patients showing ADC normalization. Occlusion of MCA branches was the predominant occlusion type in patients with ADC normalization (7 of 14 versus 9 of 54) (Figure 4). There were no significant differences in lateralization, sex, and age. For mean values and further comparisons, see Table 1 and Figure 4.
| Discussion |
|---|
|
|
|---|
ADC normalization occurred in
20% of the acute stroke patients in our sample and therefore was not a rare event, especially if patients were studied at a very early time point (mean, 2.8 hours) (Figure 1). Several recent articles report a lower frequency of normalization than we found.10,19,20 We believe that the lower frequency of normalization in these reports is due to the later initial imaging time point (<12 hours)19 and to the absence of treatment with thrombolytics.10,20 Our patients had comparatively proximal vessel occlusions and severe stroke symptoms and arrived in our center early with the option to receive thrombolytic therapy. Conclusions from our results cannot be easily transferred to all stroke patients.
All normalizers demonstrated at least partial tissue reperfusion in the scan performed on the first day after stroke onset, as presumptively demonstrated by normal or accelerated TTP in a region with previously decreased ADC (Figure 4). Thus, reperfusion was a prerequisite for ADC normalization in our study. However, information about time point and sufficiency of reperfusion cannot be derived from TTP maps on day 1. Patients with ADC normalization displayed more distal vessel occlusions, consistent with the known higher recanalization rates for distal vessel occlusions compared with proximal occlusions.21,22
In patients showing partial normalization, the proportion of tissue volume showing normalization compared with infarction decreased when ADC was initially more severely decreased (Figure 3). This finding supports previous reports suggesting that regions with ADC normalization have initially less severe ADC decreases in individual patients.8,23 On the other hand, we observed ADC normalization of tissue with an initial ADC decrease of <50% in individual patients. Thus, severe initial ADC decreases do not uniformly predict infarction.7,24
Two regions of ADC normalization were observed: white matter (Figure 2a) and parts of the basal ganglia (Figure 2b). It was somewhat surprising to detect a partial ADC normalization in the basal ganglia because the opposite has been shown in experimental work. The caudate putamen revealed a more pronounced ADC decrease (and tissue damage) than the substantia nigra and thalamus25 or the cortex.26 Although these reports argue against the recovery of deep brain structures, a recent study reported on ADC normalizations in subcortical gray matter ischemia.27 We believe that our observation can be explained by the complex flow patterns in the basal ganglia.28 We are not able to resolve this puzzle because of the finite spatial and temporal resolution of our methods. However, we occasionally found an ADC decrease without TTP delay, suggesting that reperfusion had occurred by the time of the initial study (Figure 2).
Patients with ADC normalization showed a tendency toward better clinical outcome in our study. However, the absolute improvement in NIHSS was not significantly different from that in patients without normalization in our sample. Although the mean tissue volume that displayed normalization was relatively small (mean, 16 mL; Figure 1), it is possible that the better clinical outcome in the group showing normalization was the result of salvaging functionally eloquent brain regions.
This perspective assumes that ADC normalization is an indicator of tissue salvage. However, it is not known whether the absence of a T2w lesion in follow-up MRI is an accurate indicator of functional tissue, given that selective neuronal loss has been observed in regions of normal-appearing T2w signal in postischemic MRI scans.29,30 Hence, it is not known whether the more favorable clinical outcome in the group showing ADC normalization is due to the salvage of small tissue areas, to some indirect benefit of vessel recanalization, or to the initially less extensive perfusion abnormalities found in this group.
One important question that arises from our study is whether thrombolysis should be attempted in patients who fail to show a PWI>DWI mismatch. The mismatch concept is supported by our data because patients with infarct growth had significantly larger mismatch volumes, indicating a larger tissue at risk (Table 1). On the other hand, one fifth of our patients showed a decrease in final lesion size compared with the initial DWI lesion size. This ADC normalization phenomenon seems to be time dependent: ADC normalizations occurred in 11 of 31 patients (35.5%) studied <3 hours after symptom onset but in only 3 of 37 patients (8.1%) studied within 3 to 6 hours after symptom onset. Based on our findings, the tissue at risk as target of thrombolysis therapy might be extended toward the DWI lesion at least within 3 hours after stroke onset. Consequently, at least within <3 hours, the absence of a PWI>DWI mismatch does not imply the absence of tissue at risk. In some patients, there can be some tissue that might profit from therapeutic reperfusion. The possible benefit of thrombolysis in these patients without mismatch needs further investigation.
| Conclusions |
|---|
|
|
|---|
| Acknowledgments |
|---|
Received March 6, 2003; revision received October 20, 2003; accepted October 29, 2003.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
J.M. Provenzale, K. Shah, U. Patel, and D.C. McCrory Systematic Review of CT and MR Perfusion Imaging for Assessment of Acute Cerebrovascular Disease AJNR Am. J. Neuroradiol., September 1, 2008; 29(8): 1476 - 1482. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Siemonsen, T. Fitting, G. Thomalla, P. Horn, J. Finsterbusch, P. Summers, C. Saager, T. Kucinski, and J. Fiehler T2' Imaging Predicts Infarct Growth beyond the Acute Diffusion-weighted Imaging Lesion in Acute Stroke Radiology, September 1, 2008; 248(3): 979 - 986. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Tourdias, V. Dousset, I. Sibon, E. Pele, P. Menegon, J. Asselineau, C. Pachai, F. Rouanet, P. Robinson, G. Chene, et al. Magnetization Transfer Imaging Shows Tissue Abnormalities in the Reversible Penumbra Stroke, December 1, 2007; 38(12): 3165 - 3171. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Delgado-Mederos, A. Rovira, J. Alvarez-Sabin, M. Ribo, J. Munuera, M. Rubiera, E. Santamarina, O. Maisterra, P. Delgado, J. Montaner, et al. Speed of tPA-Induced Clot Lysis Predicts DWI Lesion Evolution in Acute Stroke Stroke, March 1, 2007; 38(3): 955 - 960. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. N. Sylaja, S. B. Coutts, S. Subramaniam, M. D. Hill, M. Eliasziw, A. M. Demchuk, and for the VISION Study Group Acute ischemic lesions of varying ages predict risk of ischemic events in stroke/TIA patients Neurology, February 6, 2007; 68(6): 415 - 419. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Weller, H.-J. Wittsack, M. Siebler, V. Homberg, and R. J. Seitz Motor recovery as assessed with isometric finger movements and perfusion magnetic resonance imaging after acute ischemic stroke. Neurorehabil Neural Repair, September 1, 2006; 20(3): 390 - 397. [Abstract] [PDF] |
||||
![]() |
O. Wu, S. Christensen, N. Hjort, R. M. Dijkhuizen, T. Kucinski, J. Fiehler, G. Thomalla, J. Rother, and L. Ostergaard Characterizing physiological heterogeneity of infarction risk in acute human ischaemic stroke using MRI Brain, September 1, 2006; 129(9): 2384 - 2393. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Hagemann, H.-J. Mentzel, H. Weisser, A. Kunze, and C. Terborg Multiple Reversible MR Signal Changes Caused by Epstein-Barr Virus Encephalitis AJNR Am. J. Neuroradiol., August 1, 2006; 27(7): 1447 - 1449. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. S. Geisler, F. Brandhoff, J. Fiehler, C. Saager, O. Speck, J. Rother, H. Zeumer, and T. Kucinski Blood Oxygen Level-Dependent MRI Allows Metabolic Description of Tissue at Risk in Acute Stroke Patients Stroke, July 1, 2006; 37(7): 1778 - 1784. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Cronqvist, R. Wirestam, B. Ramgren, L. Brandt, B. Romner, O. Nilsson, H. Saveland, S. Holtas, and E.-M. Larsson Endovascular Treatment of Intracerebral Arteriovenous Malformations: Procedural Safety, Complications, and Results Evaluated by MR Imaging, Including Diffusion and Perfusion Imaging AJNR Am. J. Neuroradiol., January 1, 2006; 27(1): 162 - 176. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. J. Seitz, S. Meisel, P. Weller, U. Junghans, H.-J. Wittsack, and M. Siebler Initial Ischemic Event: Perfusion-weighted MR Imaging and Apparent Diffusion Coefficient for Stroke Evolution Radiology, December 1, 2005; 237(3): 1020 - 1028. [Abstract] [Full Text] [PDF] |
||||
![]() |
P.-S. Loh, K. S. Butcher, M. W. Parsons, L. MacGregor, P. M. Desmond, B. M. Tress, and S. M. Davis Apparent Diffusion Coefficient Thresholds Do Not Predict the Response to Acute Stroke Thrombolysis Stroke, December 1, 2005; 36(12): 2626 - 2631. [Abstract] [Full Text] [PDF] |
||||
![]() |
K.S. Butcher, M. Parsons, L. MacGregor, P.A. Barber, J. Chalk, C. Bladin, C. Levi, T. Kimber, D. Schultz, J. Fink, et al. Refining the Perfusion-Diffusion Mismatch Hypothesis Stroke, June 1, 2005; 36(6): 1153 - 1159. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Sobesky, O. Z. Weber, F.-G. Lehnhardt, V. Hesselmann, M. Neveling, A. Jacobs, and W.-D. Heiss Does the Mismatch Match the Penumbra?: Magnetic Resonance Imaging and Positron Emission Tomography in Early Ischemic Stroke Stroke, May 1, 2005; 36(5): 980 - 985. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Kucinski, D. Naumann, R. Knab, V. Schoder, S. Wegener, J. Fiehler, A. Majumder, J. Rother, and H. Zeumer Tissue at Risk Is Overestimated in Perfusion-Weighted Imaging: MR Imaging in Acute Stroke Patients without Vessel Recanalization AJNR Am. J. Neuroradiol., April 1, 2005; 26(4): 815 - 819. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Manka, F. Traber, J. Gieseke, H. H. Schild, and C. K. Kuhl Three-dimensional Dynamic Susceptibility-weighted Perfusion MR Imaging at 3.0 T: Feasibility and Contrast Agent Dose Radiology, March 1, 2005; 234(3): 869 - 877. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Hjort, K. Butcher, S.M. Davis, C.S. Kidwell, on behalf of the UCLA Thrombolysis Investigators, W.J. Koroshetz, J. Rother, P.D. Schellinger, S. Warach, and L. Ostergaard Magnetic Resonance Imaging Criteria for Thrombolysis in Acute Cerebral Infarct Stroke, February 1, 2005; 36(2): 388 - 397. [Abstract] [Full Text] [PDF] |
||||
![]() |
W.-D. Heiss, J. Sobesky, and V. Hesselmann Identifying Thresholds for Penumbra and Irreversible Tissue Damage Stroke, November 1, 2004; 35(11_suppl_1): 2671 - 2674. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Fiehler, T. Kucinski, K. Knudsen, M. Rosenkranz, G. Thomalla, C. Weiller, J. Rother, and H. Zeumer Are There Time-Dependent Differences in Diffusion and Perfusion Within the First 6 Hours After Stroke Onset? Stroke, September 1, 2004; 35(9): 2099 - 2104. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||