This Article Abstract Full Text (PDF) All Versions of this Article: 35/11_suppl_1/2671    most recent 01.STR.0000143329.81997.8av1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Heiss, W.-D. Articles by Hesselmann, V. Search for Related Content PubMed PubMed Citation Articles by Heiss, W.-D. Articles by Hesselmann, V. Pubmed/NCBI databases Medline Plus Health Information MRI Scans Nuclear Scans Stroke Related Collections Acute Cerebral Infarction Computerized tomography and Magnetic Resonance Imaging PET and SPECT

(Stroke. 2004;35:2671.)
© 2004 American Heart Association, Inc.

Articles

Identifying Thresholds for Penumbra and Irreversible Tissue Damage

Wolf-Dieter Heiss, MD; Jan Sobesky, MD Volker Hesselmann, MD

From the Max Planck Institute for Neurological Research (W.-D.H.), Department of Neurology (J.S.), and Department of Diagnostic Radiology (V.H.), University of Cologne, Cologne, Germany.

Correspondence to Dr Wolf-Dieter Heiss, Max Planck Institute for Neurological Research, Gleueler Strasse 50, 50931 Cologne, Germany. E-mail wdh{at}pet.mpin-koeln.mpg.de

	Abstract

Top Abstract Introduction Subjects and Methods Results Conclusion References

 
Diffusion-weighted MRI (DWI) in combination with perfusion-weighted MRI (PWI) has become a widely accepted modality for the selection of patients amenable for acute therapy, if a mismatch between these procedures suggests viable penumbral tissue. However, DWI as well as PWI yields semiquantitative measures limiting the definitions of irreversible damage and of potentially viable penumbral tissue. These limitations of PWI/DWI may be better understood if findings in individual patients are compared with the results from measurements of blood flow, oxygen metabolism, and benzodiazepine receptor binding obtained with positron emission tomography (PET). Comparative studies with PET and MRI were performed in 3 groups of patients: (1) In 12 acute stroke patients, results from DWI (median, 6.5 hours after symptom onset) and ¹¹C-flumazenil (FMZ) PET (median, 85 minutes between DWI and PET) were compared with infarct extension 24 to 48 hours later on T2-weighted MRI. (2) In 11 acute stroke patients, results from PWI (median, 8 hours after symptom onset) were compared with cerebral blood flow measurements obtained with [¹⁵O]H₂O PET (interval, 60 minutes between PWI and PET). (3) In 10 patients with acute (n=5) or chronic stroke (n=5), results from PWI/DWI were compared with PET of cerebral blood flow and oxygen consumption to detect mismatch or increased oxygen extraction fraction as surrogate markers of penumbra. Results were: (1) from regions with increased DWI intensity, decreased apparent diffusion coefficient (ADC) and decreased FMZ binding probability curves were computed for eventual infarction, and 95% prediction limits were determined. These limits predicted 83.5% (FMZ), 84.7% (DWI), and 70.9% (ADC) of the final infarct volume. However, the false-positive predictions were much higher for the DWI variables (5.1 and 3.6 cm³ for DWI and ADC versus a median of 0 for FMZ). (2) The comparison of volumes generated by different time to peak (TTP) thresholds (PWI) and hypoperfusion <20 mL/100 g per minute (PET) indicates that a TTP delay of 4 to 6 seconds yields a fair estimate of hypoperfusion. (3) The PWI/DWI mismatch with TTP >4 seconds did not reliably correspond to the penumbra as assessed by PET (oxygen extraction fraction >150%). Only 6 of 10 patients with a mismatch had areas of penumbra. In these cases, the penumbra volume was overestimated by MRI. DWI correlates with FMZ results and, with a few exceptions, yields a good estimate of acute tissue damage and final infarct volume. PWI measures seem to be less reliable; the TTP prolongation of >4 seconds assessed only 83% of the volume of hypoperfusion <20 mL/100 g per minute. The mismatch volume imprecisely depicts increased oxygen extraction fraction, and, despite its clinical role for selection of patients for eventual therapy, it does not to seem to be a reliable correlate of penumbra.

Key Words: magnetic resonance imaging, diffusion-weighted • magnetic resonance imaging, perfusion-weighted • penumbra • stroke, ischemic • tomography, emission-computed

	Introduction

Top Abstract Introduction Subjects and Methods Results Conclusion References

 
The concept of penumbra is based on animal experiments in which potentially reversible functional disturbances can be observed when blood flow decreases beyond a critical threshold (review in Heiss¹). With reperfusion within a limited time window, these functional disturbances are reversible without leading to morphological damage and irreversible neurological deficits. In the last 20 years, surrogate markers of penumbra and irreversible tissue damage were studied, especially in regard to patients who could benefit from treatment in acute ischemic stroke (reviews in Keir and Wardlaw² and Donnan and Davis³). Positron emission tomography (PET) has been the gold standard but was replaced by diffusion- (DWI) and perfusion-weighted imaging (PWI) because of its wider distribution and the less complex logistics involved.^4,5 However, there are some limitations with conventional DWI/PWI: increased diffusion signals can be reversible,⁶ and the determination of the threshold of critical perfusion by PWI is still a matter of ongoing debate.⁷ Therefore, we compared these methods in a limited number of patients with acute ischemic stroke.

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Conclusion References

 
Comparative studies with PET and MRI were performed in 3 groups of patients: (1) In 12 acute stroke patients (5 men, 7 women, aged 37 to 74 years; mean age, 62.4 years), DWI was performed 2.7 to 19 hours after symptom onset (median, 6.5 hours) and was followed by PET of ¹¹C-flumazenil (FMZ) within 20 to 130 minutes (median, 85 minutes). The early imaging findings were compared with infarct extension on T2-weighted MRI 24 to 48 hours later. (2) In 11 acute stroke patients (7 men, 4 women, aged 37 to 78 years; mean age, 56 years), PWI was performed 1.8 to 20 hours after symptom onset (median, 8 hours) and was followed by PET of [¹⁵O]H₂O within 20 to 140 minutes (median, 60 minutes). Increases in time to peak (TTP) maps on PWI were compared with regional cerebral blood flow (CBF) values from PET. (3) In 10 patients with acute (n=5; median, 5 hours after symptom onset) and chronic stroke (n=5; median, 8 days after symptom onset), PWI and DWI were performed and were followed by PET of H₂O (for CBF) and of [¹⁵O]O₂ (for oxygen consumption) within 1.5 to 4 hours (median, 2 hours). In these cases, the extent of PWI/DWI mismatch was compared with that of increased oxygen extraction fraction (OEF) as a surrogate marker of penumbra.

All patients included in this study suffered from ischemic hemispheric stroke of different severity, with clinical deficits ranging from mild (and transient) paresis of an arm to severe contralateral sensorimotor deficit (initial National Institutes of Health Stroke Scale score between 1 and 22 points; median, 6 points). Initial assessment included general medical and standardized neurological examinations and CT scanning. The patients received standard medical therapy; 8 patients arriving within the 3-hour window received intravenous thrombolysis before the imaging protocol was started. Fully informed consent was obtained from all patients.

MRI was performed on a Philips Intera 1.5-T whole-body scanner with single-shot and multishot spin-echo-planar imaging sequences providing 20 slices with a thickness of 6 mm. For DWI, 2 b values (b=0 and b=1000 s/mm²) were used, and DWI intensity maps and maps of the apparent diffusion coefficient (ADC) were calculated. The perfusion studies consisted of multiple measurements (40 at 2.6-second intervals or 60 at 1.3-second intervals) after a standardized injection of 20 mL GdDTPA with a flow rate of 10 mL/s. TTP maps were calculated from the raw data with the use of an interactive data language–based interactive program (Research Systems Inc). DWI intensity and ADC values were expressed as a ratio, and TTP values were expressed as the difference from the contralateral homotopic region. Size and location of the final infarct were determined 24 to 48 hours later on T2-weighted MRI.

PET studies were performed with the patients in a resting state on an ECAT EXACT HR scanner (Siemens/CTI) in 3-dimensional data acquisition mode, providing 47 contiguous 3-mm slices at 5-mm full width at half maximum reconstructed resolution. As a tracer of neuronal integrity, ¹¹C-FMZ (740 MBq) was injected intravenously, and ratios of cortical FMZ binding in the affected hemisphere relative to the contralateral white matter activity were assessed. Regional CBF was measured after intravenous bolus injection of [¹⁵O]H₂O (2.2 GBq) with the use of arterial blood sampling.⁸ Oxygen consumption was measured after a single-breath inhalation of [¹⁵O]O₂ (1.85 GBq), and the cerebral metabolic rate of oxygen (CMRO₂) and OEF were calculated with the use of the arterial input function.⁹

The early PET, DWI, and PWI as well as the late T2-weighted MRI results were individually coregistered by a multimodal coregistration program.^10,11 For DWI/FMZ comparison, spherical volumes of interest were placed in coregistered images of different variables. Regional values from different modalities were compared among each other or related to the final state, ie, infarcted or noninfarcted on late MRI. In a second step, the tissue compartments created by various thresholds were compared among the different modalities with the use of voxel-based atlases (Figure 1).

View larger version (20K): [in this window] [in a new window]   Figure 1. Data postprocessing with multimodal imaging: an individual atlas was generated based on morphological imaging (left). Within each individual atlas, the data volumes of the coregistered PET and MR images (eg, FMZ and DWI) were thresholded to create modality specific volumes (middle). After 3-dimensional fusion of these volumes (right), the subcompartments for further analysis were labeled. In the case of FMZ/DWI, the volumes were (a) infarcted tissue only identified by FMZ imaging, (b) infarcted tissue correctly identified by FMZ imaging and DWI, (c) infarcted tissue only identified by DWI, and (d) noninfarcted tissue correctly identified by FMZ imaging and DWI.

	Results

Top Abstract Introduction Subjects and Methods Results Conclusion References

 
Assessment of Cortical Tissue Damage
In the patients of group 1, cortical areas were categorized as infarction or normal according to their appearance on follow-up MRI, and volumes of interest of 6-mm diameter were fitted into the cortical rim of the coregistered DWI, ADC, and FMZ images. Across all patients’ volumes of interest, the threshold probability integrals of final infarction or noninfarction were interactively computed, and positive prediction curves were obtained on which 95% prediction limits could be defined.¹² These values—FMZ binding 3.2 times the mean of the contralateral white matter, DWI intensity 1.18 times the contralateral area, ADC 0.83 times the contralateral region—represent the 95% probability threshold of final infarction.

When the volumes of tissue beyond these thresholds were compared, close correlations between volumes with FMZ and DWI beyond threshold as well as between predicted and final infarct volumes were obtained (Figure 2), but the volumes did not completely overlap. Overall, 83.5% of the final infarct (median, 14.9 cm³) was predicted by decreased FMZ binding, 84.7% by increased DWI signal, and 70.9% by reduced ADC value. However, because of the incongruities, only a small part of the final infarct was not predicted by FMZ or DWI value beyond the critical limit (median, 1.1 cm³). The false-positive rates showed significant differences: only a small part (median, 0; mean, 0.9 cm³) of the finally noninfarcted tissue had initially decreased FMZ binding, whereas 5.1 cm³ of finally normal tissue showed an increased DWI signal (25.9% of the total volume of DWI increase) and 3.6 cm³ showed a decreased ADC value (22.3% of total volume). These differences were significant (P<0.01, Wilcoxon test). The volumes of infarcted tissue not predicted by decreased FMZ or changed DWI signal were comparable. In single cases, areas with markedly increased DWI signal did not show either impaired FMZ binding or a lesion on late MRI, as reported previously,⁶ but in most cases the differences with respect to FMZ binding and DWI signals were at the borderline of the ischemic territory.

View larger version (49K): [in this window] [in a new window]   Figure 2. Coregistered FMZ PET and DWI in a patient with left-sided hemiparesis 8 hours after stroke. The similar volumes for FMZ binding decrease and DWI hyperintensity (53.8 vs 61.2 cm3) indicate that the decreased diffusion of protons corresponds to irreversible neuronal loss.

Assessment of Perfusion
PWI-derived TTP maps were compared with quantified PET CBF images to test different TTP thresholds for their ability to identify hypoperfusion. After coregistration of the MR and PET images, an individual brain atlas was created for each patient. Then the volume of hypoperfusion of <20 mL/100 g per minute (PET CBF) was created with the use of a voxel-based threshold function. Within the same brain atlas, the TTP images were analyzed with stepwise increasing thresholds, ie, with increasing relative TTP delays (2, 4, 6, 8, 10 seconds with respect to the unaffected hemisphere). The volume of CBF hypoperfusion ranged from 1.2 to 362 cm³ (median, 34.5 cm³). The voxel-based 3-dimensional fusion of each patient’s hypoperfusion volume (CBF) and the respective set of TTP volumes were used to create subcompartments to calculate sensitivity and specificity values for each TTP threshold. The TTP threshold of 4 seconds reliably identified hypoperfused tissue (sensitivity, 0.827) and excluded normoperfused tissue (specificity, 0.768). Increasing the TTP threshold to 6 seconds impaired the ability to detect hypoperfusion (sensitivity, 0.765) but improved the rate of correctly identified normoperfused tissue (specificity, 0.875). From this small sample size, it can be concluded that a TTP delay between 4 and 6 seconds is useful to differentiate cerebral hypoperfusion <20 mL/100 g per minute.

Assessment of Penumbra
In 10 patients (5 with acute ischemic attacks 5 hours after symptom onset and 5 presenting in the subacute state after ischemic attack due to vascular stenosis), PWI/DWI revealed a mismatch between the volumes of TTP prolonged beyond 4 seconds and the volume of increased DWI signal. In these patients, CBF and CMRO₂ were measured by PET, and the volumes of increased OEF (>150%) was calculated. A comparison of the volumes of increased TTP and of increased OEF (Figure 3) demonstrated a high variability between the volumes identified by these modalities: all 10 patients showed areas of TTP prolongation on PW images (median volume, 162 cm³; range, 8 to 450 cm³). However, in only 6 of 10 patients was an elevated OEF identified on PET images (median volume, 65 cm³; range, 12 to 240 cm³). The areas of OEF elevation were always located within the areas of TTP prolongation but were significantly smaller and covered only 8% to 58% (median, 33%) of the TTP area. These preliminary data demonstrate a high sensitivity but a low specificity of the chosen threshold to identify penumbral tissue as defined by PET: in 40% of the patients with TTP >4, no OEF elevation was found. In the remaining 60% with OEF elevation and TTP >4, only a third of the TTP volume corresponded to elevated OEF. These findings may explain the poor relation between increased OEF and TTP values in a correlation analysis based on volumes of interest. The results of our small patient sample indicate that TTP threshold >4 seconds does not sufficiently discriminate between normal and increased OEF in ischemic tissue.

View larger version (30K): [in this window] [in a new window]   Figure 3. Multimodal imaging in a patient presenting with a right-sided hemiparesis and expressive aphasia 5 hours after symptom onset. The penumbra is characterized by hypoperfusion (CBF), preserved CMRO2, and elevated OEF. These findings correspond grossly to the mismatch between DWI hyperintensity and a TTP delay of 4 seconds compared with the unaffected hemisphere.

	Conclusion

Top Abstract Introduction Subjects and Methods Results Conclusion References

 
These preliminary data on the comparison of PET and DWI/PWI for assessment of perfusion, identification of irreversible tissue damage, and distinction of penumbra (the final analysis of 1 part has been published,¹³ and the other parts are in preparation) indicate that imaging with these different modalities yields complementary information on the dynamics of pathophysiological events in ischemic brain tissue. The findings of DWI/ADC imaging correlate well with those of FMZ PET and predict the final infarct extension.^14,15 However, the increased DWI signal carries a considerable false-positive rate, a clinically important restriction of DWI that was demonstrated in previous studies of patients with transient ischemic attack and those treated with thrombolysis.^16–18 PWI-derived TTP maps are only indirect surrogates of CBF because they are based on a different pathophysiological approach.^2,19 Hence, a TTP prolongation of >4 seconds allows only a fair estimate of hypoperfusion <20 mL/100 g per minute.²⁰ Because of this indirect assessment of perfusion from TTP prolongation, which therefore inconsistently corresponds to elevated OEF,²¹ the mismatch volume in PWI/DWI does not reliably reflect misery perfusion as defined by PET. Despite the importance of MRI in clinical practice for the selection of patients who might benefit from revascularization procedures,^4,5 the differences between the imaging methods and their specific methodological restrictions should be taken into account when results are compared and definitions of various tissue conditions are transferred between the modalities.

Received June 15, 2004; revision received July 30, 2004; accepted August 5, 2004.

	References

Top Abstract Introduction Subjects and Methods Results Conclusion References

 

1. Heiss W-D. Ischemic penumbra: evidence from functional imaging in man. J Cereb Blood Flow Metab. 2000; 20: 1276–1293.[CrossRef][Medline] [Order article via Infotrieve]
2. Keir SL, Wardlaw JM. Systematic review of diffusion and perfusion imaging in acute ischemic stroke. Stroke. 2000; 31: 2723–2731.[Abstract/Free Full Text]
3. Donnan GA, Davis SM. Neuroimaging, the ischaemic penumbra, and selection of patients for acute stroke therapy. Lancet Neurology. 2002; 1: 417–425.
4. Röther J, Schellinger PD, Gass A, Siebler M, Villringer A, Fiebach JB, Fiehler J, Jansen O, Kucinski T, Schoder V, Szabo K, Junge-Hulsing GJ, Hennerici M, Zeumer H, Sartor K, Weiller C, Hacke W. Effect of intravenous thrombolysis on MRI parameters and functional outcome in acute stroke <6 hours. Stroke. 2002; 33: 2438–2445.[Abstract/Free Full Text]
5. Schellinger PD, Fiebach JB, Hacke W. Imaging-based decision making in thrombolytic therapy for ischemic stroke: present status. Stroke. 2003; 34: 575–583.[Abstract/Free Full Text]
6. Fiehler J, Knudsen K, Kucinski T, Kidwell CS, Alger JR, Thomalla G, Eckert B, Wittkugel O, Weiller C, Zeumer H, Rother J. Predictors of apparent diffusion coefficient normalization in stroke patients. Stroke. 2004; 35: 514–519.[Abstract/Free Full Text]
7. Thijs VN, Somford DM, Bammer R, Robberecht W, Moseley ME, Albers GW. Influence of arterial input function on hypoperfusion volumes measured with perfusion-weighted imaging. Stroke. 2004; 35: 94–98.[Abstract/Free Full Text]
8. Herscovitch P, Markham J, Raichle ME. Brain blood flow measured with intravenous H₂¹⁵O, I: theory and error analysis. J Nucl Med. 1983; 24: 782–789.[Abstract/Free Full Text]
9. Baron JC, Frackowiak RS, Herholz K, Jones T, Lammertsma AA, Mazoyer B, Wienhard K. Use of PET methods for measurement of cerebral energy metabolism and hemodynamics in cerebrovascular disease. J Cereb Blood Flow Metab. 1989; 9: 723–742.[Medline] [Order article via Infotrieve]
10. Pietrzyk U, Herholz K, Fink G, Jacobs A, Mielke R, Slansky I, Würker M, Heiss W-D. An interactive technique for three-dimensional image registration: validation for PET, SPECT, MRI and CT brain studies. J Nucl Med. 1994; 35: 2011–2018.[Abstract/Free Full Text]
11. Cizek J, Herholz K, Vollmar S, Schrader R, Klein J, Heiss WD. Fast and robust registration of PET and MR images of human brain. Neuroimage. 2004; 22: 434–442.[CrossRef][Medline] [Order article via Infotrieve]
12. Heiss W-D, Kracht LW, Thiel A, Grond M, Pawlik G. Penumbral probability thresholds of cortical flumazenil binding and blood flow predicting tissue outcome in patients with cerebral ischaemia. Brain. 2001; 124: 20–29.[Abstract/Free Full Text]
13. Heiss WD, Sobesky J, Smekal UV, Kracht L, Lehnhardt FG, Thiel A, Jacobs A, Lackner K. Probability of cortical infarction predicted by flumazenil binding and diffusion-weighted imaging signal intensity: a comparative positron emission tomography/magnetic resonance imaging study in early ischemic stroke. Stroke. 2004; 35: 1892–1898.[Abstract/Free Full Text]
14. Lövblad KO, Baird AE, Schlaug G, Benfield A, Siewert B, Voetsch B, Connor A, Burzynski C, Edelman RR, Warach S. Ischemic lesion volumes in acute stroke by diffusion-weighted magnetic resonance imaging correlate with clinical outcome. Ann Neurol. 1997; 42: 164–170.[CrossRef][Medline] [Order article via Infotrieve]
15. Oppenheim C, Grandin C, Samson Y, Smith A, Duprez T, Marsault C, Cosnard G. Is there an apparent diffusion coefficient threshold in predicting tissue viability in hyperacute stroke? Stroke. 2001; 32: 2486–2491.[Abstract/Free Full Text]
16. Kidwell CS, Saver JL, Mattiello J, Starkman S, Vinuela F, Duckwiler G, Gobin YP, Jahan R, Vespa P, Kalafut M, Alger JR. Thrombolytic reversal of acute human cerebral ischemic injury shown by diffusion/perfusion magnetic resonance imaging. Ann Neurol. 2000; 47: 462–469.[CrossRef][Medline] [Order article via Infotrieve]
17. Fiehler J, Foth M, Kucinski T, Knab R, von Bezold M, Weiller C, Zeumer H, Rother J. Severe ADC decreases do not predict irreversible tissue damage in humans. Stroke. 2002; 33: 79–86.[Abstract/Free Full Text]
18. Parsons MW, Barber PA, Chalk J, Darby DG, Rose S, Desmond PM, Gerraty RP, Tress BM, Wright PM, Donnan GA, Davis SM. Diffusion- and perfusion-weighted MRI response to thrombolysis in stroke. Ann Neurol. 2002; 51: 28–37.[CrossRef][Medline] [Order article via Infotrieve]
19. Grandin CB. Assessment of brain perfusion with MRI: methodology and application to acute stroke. Neuroradiology. 2003; 45: 755–766.[CrossRef][Medline] [Order article via Infotrieve]
20. Neumann-Haefelin T, Wittsack HJ, Wenserski F, Siebler M, Seitz RJ, Mödder U, Freund HJ. Diffusion- and perfusion weighted-MRI: the DWI/PWI mismatch region in acute stroke. Stroke. 1999; 30: 1591–1597.[Abstract/Free Full Text]
21. Kajimoto K, Moriwaki H, Yamada N, Hayashida K, Kobayashi J, Miyashita K, Naritomi H. Cerebral hemodynamic evaluation using perfusion-weighted magnetic resonance imaging: comparison with positron emission tomography values in chronic occlusive carotid disease. Stroke. 2003; 34: 1662–1666.[Abstract/Free Full Text]

This article has been cited by other articles:

				B. D. Murphy, A. J. Fox, D. H. Lee, D. J. Sahlas, S. E. Black, M. J. Hogan, S. B. Coutts, A. M. Demchuk, M. Goyal, R. I. Aviv, et al. White Matter Thresholds for Ischemic Penumbra and Infarct Core in Patients with Acute Stroke: CT Perfusion Study Radiology, June 1, 2008; 247(3): 818 - 825. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 35/11_suppl_1/2671    most recent 01.STR.0000143329.81997.8av1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Heiss, W.-D. Articles by Hesselmann, V. Search for Related Content PubMed PubMed Citation Articles by Heiss, W.-D. Articles by Hesselmann, V. Pubmed/NCBI databases Medline Plus Health Information MRI Scans Nuclear Scans Stroke Related Collections Acute Cerebral Infarction Computerized tomography and Magnetic Resonance Imaging PET and SPECT