This website is intended for healthcare professionals

Penumbral imaging in acute stroke: a triumph of hope over experience?

Posted in Stroke Series on 10th Sep 2014


Introduction to the ACNR Stroke Series

The ischaemic penumbra is a concept familiar to all clinicians and scientists interested in ischaemic stroke. The concept of “potentially salvageable tissue” has revolutionised the practice of stroke medicine: intravenous thrombolysis is now routine, and other thrombus removal techniques are under active investigation. However, despite many years of investigation, how imaging the penumbra may help to select patients most likely to benefit from reperfusion treatment remains to be fully settled. In this article in our Stroke Series, we are very lucky to have a clear and authoritative account of the development of penumbral imaging – from initial experiments to recent randomised trials incorporating advanced MR imaging – by Josef Alawneh and Jean-Claude Baron. Standardisation of clinical application of penumbral imaging is clearly needed, and the limitations of simply applying the “pretty blobby maps” available commercially are made clear. However, recent randomised data give optimism that penumbral selection at various time windows in acute stroke will be useful in the future.

David Werring, Reader in Clinical Neurology, UCL Institute of Neurology, National Hospital for Neurology and Neurosurgery, Queen Square, WC1N 3BG.


  • Penumbra imaging for the selection of candidates for reperfusion therapy is widely used in routine clinical practice
  • Randomised controlled trials using penumbra imaging, however, have been negative possibly due to inadequate methodology
  • Further trials with improved design are ongoing

Penumbral imaging in acute stroke: a triumph of hope over experience?

stroke-bio-boxIt’s been more than thirty years since Astrup, Siesjö and Symon1 described in a model of middle cerebral artery (MCA) occlusion in baboons the presence of two separate cerebral blood flow thresholds below which electrical and energy failure occur, respectively. They noted that between those two thresholds neuronal activity temporarily seizes but could potentially recover if perfusion is restored; this area resembled the half-shaded zone around the centre of a complete solar eclipse and thus was called ‘the ischaemic penumbra’. The potential clinical benefits of its salvage and the importance of developing tools to identify it were already noted since those early days.

The first studies that proved the existence of the penumbra in acute stroke in humans used Positron Emission Tomography (PET) which has remained the gold standard imaging technique until now. Furlan et al2 showed using 15 O-PET the presence of a critically hypoperfused tissue with high oxygen extraction fraction (OEF) that either recovered or progressed to infarction, whose volume correlated with acute neurological deficit and whose recovery was strongly associated with clinical improvement. Therefore, this tissue fulfilled all the prespecified operational criteria for the penumbra,3,4 namely a) ischaemic brain tissue (CBF <20mls/100g.min, i.e., ~40% of normal, and high OEF), b) that is functionally affected, c) is of uncertain fate, becoming part of the infarct or recovering depending on subsequent (early) reperfusion, and d) whose survival, if any, underlies clinical recovery.5

The importance of treatments that can salvage the penumbra therefore became very apparent6 and it wasn’t long before salvage of penumbra was documented after IV thrombolysis using PET imaging.7 However, PET imaging has several limitations precluding its widespread use in acute stroke; these include unavailability, technical difficulty and high cost. This in turn shifted the interest to other imaging modalities potentially able to depict the penumbra in the acute stage. One of these modalities is Magnetic Resonance Imaging (MRI) and in particular the development of the Diffusion Weighted and Perfusion Weighted Imaging (DWI and PWI respectively). The DWI sequence is very sensitive in detecting severely ischaemic tissue within minutes after stroke onset; furthermore, DWI lesions strongly predict final infarct within a few hours of an acute stroke and thus these lesions have been regarded as representing the infarct core. PWI, on the other hand, can depict all hypoperfused tissue and thus can identify not only the infarct core but also the tissue at-risk of infarction. It became, therefore, common to regard the mismatch between DWI and PWI as penumbral tissue due to its location outside the infarct core but still being at risk of infarction. Moreover, several early studies have shown that salvage of the DWI/PWI mismatch is associated with neurological improvement, thus fulfilling another of the cardinal operational criteria of penumbra.9-11

The DWI/PWI mismatch concept, however, has its limitations as well. One of these is that very early DWI lesions don’t always represent the infarct core and can permanently reverse in part or in whole following acute reperfusion, thus behaving like penumbral tissue;12 in addition, the precise penumbral perfusion thresholds on PWI imaging are still the matter of ongoing research. This has led researchers to search for other sequences that might depict the penumbra more precisely. One of these used the BOLD T2* MRI signal affected by deoxyhaemoglobin to map the OEF. This has been reported with varying success but hasn’t been able to take over from the DWI/PWI paradigm so far.13-15

As an alternative to MRI, Computer Tomography Perfusion (CTP) imaging16-18 has been extensively used to define the penumbra. Compared to MRI, it is more easily accessible and faster than MR; moreover, it is suitable for uncooperative patients and those with possible MR contraindications. Recently the methodology using CTP has improved to provide both full brain coverage and CT angiogram following a single contrast injection. CTP thresholds to identify core and penumbra have been validated by comparing to the final infarct19-23 and by clinical correlations.19, 22, 24, 25 However, the uncertainties of the still imprecise penumbral thresholds that exist with PWI apply to CTP as well; with in addition the issue of poor signal to noise ratio in severely hypoperfused areas.

Clinical trials

Despite the extensive research on imaging penumbral tissue, the first positive randomised controlled studies (RCTs) on IV thrombolysis in acute stroke used straightforward non contrast CT head to exclude haemorrhage and no penumbral imaging.26,27 This was a pragmatic and immediately widely applicable approach, but it deprived us from assessing whether selecting patients for thrombolysis using penumbral imaging is of benefit in the first 4.5 hours after stroke onset. However, several open-label cohort studies using DWI/PWI mismatch to select patients for thrombolysis beyond the proven time window were promising28-31 and opened the way for large-scale prospective studies applying the mismatch concept. One of those studies was DEFUSE,32 a prospective open-label study aiming to determine whether baseline MRI profiles can identify patients who would benefit from reperfusion between 3-6 hours post onset. However, the hypothesis of the study was reviewed in an interim analysis; a subgroup of patients who had unfavourable outcome was identified and labelled as the ‘malignant profile’; those patients were excluded and the remaining patients with mismatch were labelled as the ‘target mismatch’ profile. Target mismatch was defined as PWI/DWI ratio ≥1.2, with PWI lesion defined as Tmax >2s and ≥10ml (Table 1), and the ‘malignant profile’ as the presence of a DWI lesion and/or severe (Tmax≥8s) PWI lesion ≥100mls. The study showed that reperfusion in patients with the ‘target mismatch’ profile was associated with favourable outcome when compared to reperfusion in patients with no mismatch.32,33

The first randomised placebo controlled study that prospectively acquired MRI with DWI/PWI imaging before iv thrombolysis was the EPITHET study, which aimed to test the validity of the mismatch concept in the 3-6 hours time window.34 However, the study design was to assess in retrospect whether the presence of mismatch was a predictor of good response to iv therapy versus placebo, i.e., randomisation was not based on MRI criteria. Patients with mismatch were defined as those with PWI/DWI ratio >1.2 and a PWI-DWI volume ≥10ml, with the PWI lesion defined as tissue with Tmax ≥2s (Table 1).


101 patients were randomly assigned to receive alteplase or placebo. The primary endpoint was infarct growth between baseline DWI and day 90 T2 lesion in the mismatch patients only, and this was non-significantly lower in the alteplase group. The study found, however, that alteplase was significantly associated with reperfusion in the mismatch group, and reperfusion in turn was significantly associated with better neurological outcome. However, as will be discussed below, post-hoc reanalysis of this dataset using more appropriate perfusion thresholds and endpoints showed strongly positive results. The need for larger randomised studies using well-designed penumbra imaging in the inclusion criteria became ever more apparent following this study.

Around the same time two pilot studies (DIAS and DEDAS) of a new thrombolysis drug were published. DIAS was a dose-finding phase II randomised trial designed to evaluate the safety and efficacy or intravenous desmoteplase, a novel plasminogen activator, administered within 3-9 hours of ischaemic stroke onset in patients with perfusion/diffusion mismatch on MRI (Table 1).35 Part 1 of the trial was terminated prematurely due to the high rate of symptomatic intracranial haemorrhage (sICH). Following reduction in the administered dose, part 2 of the study showed lower rates of sICH using doses up to 125µg/kg. The conclusions were that desmoteplase in patients with mismatch was associated with a higher rate of reperfusion rates and better clinical outcome compared with placebo. This pilot study was followed by DEDAS36 which also used perfusion /diffusion mismatch ratio ≥1.2 to assess the safety and efficacy of desmoteplase 90µg/kg and 125µg/kg in patients with acute ischaemic stroke 3-9 hours after onset. This small phase II study (n=37) confirmed the findings of DIAS that IV desmoteplase at those doses is safe and may improve clinical outcome.

The findings of the above two phase II desmoteplase studies led to the first large phase III study that used penumbra imaging in its inclusion criteria. DIAS II37 was a randomised, placebo-controlled, double blinded, dose-ranging study comparing desmoteplase 90 µg/kg and 125 µg/kg versus placebo in acute ischaemic stroke given 3-9 hours after onset. The primary end point was clinical response rate at day 90, defined as a composite of improvement in NIHSS score of ≥ 8 points or an NIHSS score ≤ 1, a modified Rankin scale score 0-2, and a Barthel index of 75-100. 193 patients were randomised, and 186 patients received treatment: 57 received 90µg/kg desmoteplase; 66 received 125µg/kg desmoteplase; and 63 received placebo. The clinical response rates at day 90 were 47% for 90µg/kg desmoteplase, 36% for 125µg/kg desmoteplase and 46% for placebo, showing no statistical benefit of desmoteplase given 3-9 hours after onset of stroke. Patients were included if they had a distinct ‘penumbra’ with a perfusion lesion/core ratio ≥1.2 based on MRI (PWI/DWI) or CT (CTP/CT) (Table 1). Although the desmoteplase trials deliberately applied still unproven imaging markers to a time-window beyond those investigated in observational studies, a strategy that in retrospect was unlikely to meet with success, the negative results of DIAS II came as a surprise to the stroke community at large, and raised several questions on the imaging definition of penumbra as will be discussed below.

To enrich the population studied by recruiting a more homogeneous group, a randomised trial of tenecteplase versus alteplase for acute ischaemic stroke used the presence of penumbra on perfusion imaging as an inclusion criterion, in order to enrich the population studied by recruiting a more homogeneous group.38 This novel design allowed them a considerable reduction of the required sample size. This was a Phase 2B 3-arms trial that assigned 75 patients to receive alteplase (0.9mg/kg) or tenecteplase (0.1mg/kg or 0.25mg/kg) less than six hours after stroke onset. Inclusion criteria included the presence of perfusion lesion at least 20% greater than core, with volume of at least 20 ml and an associated vessel occlusion on CTA. Perfusion lesions were based on CTP transit-time maps, and core defined using cerebral blood volume maps. The coprimary end points were the proportion of the perfusion lesion that was reperfused at 24 hours on MRI PWI and the extent of clinical improvement at 24 hours assessed on the NIHSS score. Together, the two tenecteplase groups had greater reperfusion (p=0.004) and clinical improvement (p<0.001) at 24 hours than the alteplase group with no significant differences in serious adverse events. In this study, time to treatment was 2.7 hours in the alteplase group and 3.1 hours in the tenecteplase group which is the earliest that a randomised study with penumbral imaging had achieved. However, there was no control arm in this study (which would have been unethical to have) and so no comparison is available to a control group that was not treated.

The above studies were exploring IV thrombolysis treatment; and while one would expect similar findings with endovascular reperfusion, the time came for that to be proven. DEFUSE 239 was set up to address the question: Can MRI identify patients who are most likely to benefit from endovascular reperfusion? This was a prospective study that enrolled consecutive patients scheduled to have endovascular treatment within 12 hours of onset of hemispheric stroke. Patients had acute MRI imaging at baseline before treatment to assess the presence of ‘target mismatch’ and had repeat MRI within 12 hours to assess for reperfusion and a later structural MRI to assess final infarct. The primary outcome was favourable clinical response, defined as an improvement of 8 points or more on the NIHSS between baseline and day 30 or a score of 0-1 at day 30. Ninety-nine patients were included in the analysis; 78 patients had target mismatch and 59% of them had reperfusion; 21 had no target mismatch and 57% of them had reperfusion. The adjusted odds ratio for favourable clinical response associated with reperfusion was 8.8 in the target mismatch group and 0.2 in the no target mismatch group (p=0.003). Interestingly, the clinical benefit from reperfusion was not significantly different between patients in the ≤6 hours and those in the 6-12 hours time window, suggesting that penumbral imaging is able to identify patients who would benefit from late intra-arterial therapy even beyond 6 hours and up to 12 hours, after clinical onset. It is important to note that the perfusion parameters to define target mismatch in this study had been modified from DEFUSE and made more specific for the penumbra as follows: Tmax >6s and mismatch ratio ≥1.8 (Table 1); furthermore, mismatch volume had to be ≥15ml, DWI lesion <70ml and severe (Tmax>10s) perfusion lesion <100ml. Around half of the patients received IV rt-PA before endovascular procedure and the median onset from stroke onset to baseline MRI was 4.5 hours.

The need for a randomised study of endovascular treatment using penumbra imaging was apparent and MR RESCUE 40 was designed to address that. One hundred and eighteen patients were randomly assigned within 8 hours after the onset of large-vessel anterior circulation stroke to undergo mechanical embolectomy or receive standard care. Patients underwent pretreatment computer tomography or magnetic resonance imaging of the brain and randomisation was stratified according to whether the patient had a favourable penumbral pattern (substantial salvageable tissue and small infarct core) or non-penumbral pattern (large core or small/absent penumbra). Favourable penumbral pattern was assigned on a CT or MRI brain if the predicted infarct core was ≤ 90ml and when the proportion of predicted infarct tissue within the at-risk region was ≤ 70% (i,e mismatch ratio around 1.4; table 1)41. The primary outcome was the 90 day mRS mean score. The mean age was 65.5 and the mean time to enrolment was 5.5 hours; 58% had a favourable penumbra pattern. Mean scores on the mRS did not differ between embolectomy and standard care (3.9 vs 3.9, P=0.99). Embolectomy was not superior to standard care in patients with either a favourable penumbral pattern (mean score 3.9 vs 3.4, P=0.23) or non-penumbral pattern (4.0 vs 4.4; P=0.32). As a conclusion the authors stated that a favourable penumbral pattern on neuroimaging did not identify patients who would differentially benefit from endovascular therapy for acute ischaemic stroke nor was embolectomy superior to standard care. However, as will be discussed below, this study has been criticised and its results called into question.


We reviewed above the major thrombolysis or acute interventional studies that included penumbral imaging in their inclusion criteria. A first observation is that while the results of the pilot and observational studies were promising, those of the main randomised studies were negative. This raised important questions regarding the methodology used, and even some went as far as questioning the penumbral concept itself. The possible reasons for these negative results will be explored here by attempting to answer the question: Why did EPITHET, DIAS II and MR RESCUE fail to support the notion that using penumbral imaging is clinically useful?

To start with, as already noted, when the first positive alteplase study26 was conducted, imaging of the penumbra was not widely available and thus wasn’t required for recruitment. Subsequently, once alteplase was licensed for use within 3 hours of stroke onset, the emphasis was to get plain CT imaging swiftly and very little room was left for any research on penumbra imaging within that 3 hour window. The first thrombolysis studies using penumbral imaging (EPITHET and DIAS II), therefore, recruited patients beyond the 3 hours, looking into extending that window. This deprived us, however, and perhaps for ever, of any evidence of the clinical benefit from penumbra imaging in the first 3 hours post stroke onset, at a time when penumbral tissue is expected to be the largest. This was despite the suggestion from observational studies that penumbral imaging even within the 3-hour window enhanced the clinical benefit from iv t-PA.29 The challenge faced by EPITHET and DIAS II, therefore, was to capture this shrinking penumbra beyond 3 hours and prove that its salvage was of clinical benefit. While this has proved to be very hard, the negative results from the primary analysis of those two studies have provided us with important lessons for the future.

One of the cardinal questions these two studies had to address was how to define penumbral tissue using MR; and both EPITHET and DIAS II relied on the PWI/DWI mismatch concept. EPITHET used Tmax ≥2s as the perfusion threshold to define penumbral tissue and DWI lesion for the core; DIAS II used any perfusion abnormality seen without specifying a threshold as the penumbral threshold and the DWI lesion again as the core. In retrospect, it is now clear that these definitions were too liberal and resulted in the inclusion of significant oligaemic tissue that was not dysfunctional nor at risk.42 Observational studies that directly compared MRI perfusion parameters to PET reported Tmax values around 5.5s as the closest to the penumbral threshold;43-45 implying that significant amounts of the tissue regarded as penumbra in those studies were mere oligaemia. In DIAS II the perfusion/core mismatch ratio in percentage was reported to be >500% more than 6 hours after stroke onset; this, indeed, is unlikely to represent true penumbral tissue at those late time points. Moreover, up to 67% of patients screened were thought to have significant mismatch volume at 6 hours post onset, again a very large proportion which might have included patients with oligaemia but not true mismatch. To address the above, a post hoc analysis was recently published that refined the definition of mismatch, first in DIAS II, and then in a pooled cohort from DIAS, DEDAS and DIAS II.46 Mismatch was re-defined as the difference between the calculated volume of hypoperfusion based on MTT maps (no defined threshold) and volume of DWI, but tested only in patients with mismatch volume >60ml. The difference in favourable response rate to desmoteplase between placebo and treatment group increased with the new definition but was statistically significant only in the pooled analysis. The authors concluded that desmoteplase appears beneficial in patients with large mismatch volume but that this will need to be tested in prospective studies. Along the same lines, in EPITHET significant oligaemic tissue was regarded as penumbra and it is suspected that a sizeable number of patients were misclassified. This is supported by the fact that the vast majority of patients (86%) screened were judged to have significant mismatch volume 3-6 hours post onset, again a clear overestimation. And when penumbra was redefined in a post hoc analysis using a threshold of Tmax>6s and proper image co-registration, the median mismatch volume fell from 126ml to 37ml and only 53% of patients were now found to have had true penumbra. This lead to several patients being reclassified, and a reanalysis with the new definition of penumbra including pooled patients from DEFUSE showed that favourable tissue and clinical outcome from reperfusion in the target mismatch group became statistically significant.33

The other important question the above two studies had to address was what mismatch ratio is to be considered significant. This is the ratio of total affected tissue divided by core. Increasing the ratio would decrease the absolute volume of mismatch and also the chances of observing clinical improvement should that mismatch tissue survive; however, the more you increase the ratio, the more patients will be excluded as having no mismatch. Both DIAS II and EPITHET used 1.2 as their cut-off ratio, meaning that the affected but not core tissue ‘penumbra’ had to be at least 20% larger than the DWI lesion ‘core’, which in retrospect seems small. As the results of both trials were negative, subsequent studies used larger ratios in order to maximise the chances of capturing and saving clinically relevant mismatch. MR RESCUE used a ratio around 1.4 but failed to demonstrate any positive effect. DEFUSE II went up to 1.8 and had more promising results; however, this was an observational study with no randomised control arm. In order to address the question of which should be the optimum ratio to use, a reanalysis of the DEFUSE data was carried out, which evaluated the odds ratio for a favourable clinical response in mismatch patients with reperfusion compared with no reperfusion for various mismatch ratio thresholds.47 A mismatch ratio of 2.6 provided the highest sensitivity and specificity for identifying patients in whom reperfusion was associated with favourable response. Such a high mismatch ratio has not been tested in any randomised study yet, but clearly would considerably reduce the number of recruitable patients.

Indeed, an important aspect of these studies is the number needed to recruit to be able to show significant difference in clinical response. The aim of penumbra imaging is to screen patients and select those who are most likely to benefit from thrombolysis and thus reduce the total number of participants in the studies. However, this hasn’t happened on the ground as shown above; and thus larger numbers of patients possibly were eventually needed to demonstrate any benefit as the screening process has not been optimum. For example the NINDS study,26 the first to demonstrate benefit from rtPA within the first 3 hours recruited 333 patients at a time-point where penumbra is expected to be large. ECASS III had to recruit 821 patients to demonstrate thrombolysis benefit between 3 and 4.5 hours post onset.27 EPITHET recruited only 101 and DIAS II 186 patients. While these numbers are relatively large compared to previous imaging studies, they are quite small compared to previous thrombolysis studies; particularly if you take into account that recruitment was between 3-6 hours at a time where penumbra volume has significantly declined. On the other hand, the above-mentioned tenecteplase trial requiring presence of penumbra as an inclusion criterion needed only 25 patients per arm to document efficacy. Calculating the sample size needed for those trials is challenging due to difficulty in estimating an accurate treatment effect. It is been estimated that about 330 patients will be needed in a study using penumbral imaging before thrombolysis treatment.48

We addressed above the difficulties related to penumbra imaging in the two large IV thrombolysis studies. It is also expected that trials using an endovascular approach would face the same issues. This was noted in the MR RESCUE40 trial which used penumbra imaging before intervention. Here Tmax >6s threshold was used to define mismatch; mismatch ratio of around 1.4 was used which is marginally but definitely better than the 1.2 previously used; and the mean time to enrolment was 5.5 hours. The total number of patients recruited was 118 but those with a favourable penumbra pattern were only 68 (34 embolectomy and 34 standard care group) thus significantly reducing the power of this study. An additional confounder in this study was that a percentage of the patients had IV thrombolysis before enrolment. This may be relevant to the fact that despite allocation to embolectomy, recanalisation and reperfusion rates were similar between the treatment and control group. Reperfusion is the most important factor leading to salvage of the penumbra, and without it, no matter the volume of penumbra, good recovery is unlikely. Taking all the above limiting factors together, plus the fact that this study started many years ago and therefore used largely outdated thrombectomy devices and that both MR and CT perfusion were used to determine the favourable penumbral pattern, maybe it is not surprising that it was negative.

There are a few additional difficulties that studies aiming at imaging the penumbra have to face. One is the post-processing used after acquiring the imaging. Various methods can be used, leading to different maps, each providing potentially different information.48,49 There has been no consensus yet on the best methodology, software, maps and thresholds to use for each particular method; however, there are recommendations regularly published addressing these topics (Stroke imaging research roadmap)50 which would be advisable to adhere to in order to improve consistency between studies. This remains a topic of active research and there is still some way to go until these questions are resolved. However, proprietary issues are involved here as well, as, at the moment each manufacturer offers commercial software which works as a ‘black box’ that delivers pretty blobby maps of penumbra and core of unclear physiological validity. In addition, in order for post-processing tools to be useable in clinical trials and also in future routine practice, they must consistently deliver reliable core/penumbra maps within a minute or two of data acquisition at most. Recently developed tools are being tested currently.33

While the future of penumbra imaging in acute stroke within or beyond treatment window is uncertain, there are several potential uses of penumbral imaging which would be useful to explore in future studies. Within the 4.5 hours window, future imaging-based studies could address the following issues: a) exclusion of patients who would not benefit from treatment such as those with no perfusion deficit or already extensive core; b) exclusion of patients who might be harmed by treatment (i.e high bleeding risk); c) inclusion of patients who might benefit but are otherwise often excluded like those with mild, improving or fluctuating deficits; and d) exploring additional treatment options in those who have not rapidly recanalised after initial thrombolysis, such as ‘bridging’ from iv to intra-arterial therapy. Outside the 4.5 hour window, future studies could explore the following: a) inclusion of selected patients with significant mismatch and small core after taking in to consideration the issues addressed in this review; and b) inclusion of patients with unknown time of onset (e.g., awakening stroke). A few ongoing studies are currently attempting to address some of the above questions. One of them is EXTEND, which is exploring IV thrombolysis beyond thrombolysis window in selected patients based on penumbral imaging ( Identifier: NCT00887328; accessed 24/08/2013). This trial recruits patients between 3-9 hours post stroke onset. Penumbral imaging inclusion criteria include: a mismatch ratio >1.2, a DWI volume ≤70mL and a PWI-DWI difference >10 ml. Perfusion lesion is defined as tissue with Tmax > 6 second. While the perfusion threshold chosen here is appropriate to identify the penumbra, the mismatch threshold used is similar to previous negative studies. Another, similarly designed trial to be run in Europe is ECASS 4, currently being set up. A third study, from China, entitled “Imaging-based Thrombolysis trial in Acute Ischemic Stroke-II (ITAIS-II)”51 also recruits patients 3-9 hours post stroke onset; this study is using CT perfusion in the inclusion criteria and eligible patients are those with significant perfusion deficit defined by a perfusion lesion >2cm diameter.

This study also opted for a mismatch ratio of 1.2 and requires the presence of large vessel occlusion on CTA as well. It will be interesting to see the results of this study that uses a completely new method of defining patients with significant perfusion deficit.

We have reviewed above how the ischaemic penumbra has evolved over the last three decades from a mere thought to a lab experiment, then to a proven concept, and nowadays it is a daily topic in almost every research or clinical stroke meeting. However, despite the intensive research done so far, and the extensive use of penumbral imaging worldwide,48 the evidence supporting its use in the treatment of acute stroke has significantly lagged behind.52 We explored the potential reasons for this situation and provided some guidance for future research taking into account the lessons learned. It is nevertheless a fact of life that acute stroke MR- or CT-based protocols including perfusion imaging on top of DWI, MR angiography and Gradient Echo for the former, and plain CT and CT angiography for the latter, are increasingly used worldwide in daily routine, not just beyond the licensed time-window for iv t-PA but also within it. This is supported by numerous scientific articles pointing to the overall clinical benefits from the array of diagnostic, aetiologic and pathophysiological information provided even beyond penumbra imaging, and used to guide overall patient management – although not based on definitive scientific evidence as yet.


The concept and the presence of penumbra have been documented by extensive experimental and clinical investigations over the last few decades. However, although already widely used in clinical routine, penumbra imaging for the selection of candidates for reperfusion therapy both within and beyond the current licensed time-window for iv thrombolysis has not been validated yet, perhaps due to inadequate methodology, and randomised controlled trials with improved design are ongoing.


  1. Astrup J, Siesjo BK, Symon L. Thresholds in cerebral ischemia – the ischemic penumbra. Stroke. 1981;12:723-725
  2. Furlan M, Marchal G, Viader F, Derlon JM, Baron JC. Spontaneous neurological recovery after stroke and the fate of the ischemic penumbra. Ann Neurol. 1996;40:216-226
  3. Baron JC. Mapping the ischaemic penumbra with pet: Implications for acute stroke treatment. Cerebrovasc Dis. 1999;9:193-201
  4. Muir KW, Buchan A, von Kummer R, Rother J, Baron JC. Imaging of acute stroke. Lancet Neurol. 2006;5:755-768
  5. Heiss WD. Experimental evidence of ischemic thresholds and functional recovery. Stroke. 1992;23:1668-1672
  6. Baron JC, von Kummer R, del Zoppo GJ. Treatment of acute ischemic stroke. Challenging the concept of a rigid and universal time window. Stroke. 1995;26:2219-2221
  7. Heiss WD, Grond M, Thiel A, von Stockhausen HM, Rudolf J, Ghaemi M, Lottgen J, Stenzel C, Pawlik G. Tissue at risk of infarction rescued by early reperfusion: A positron emission tomography study in systemic recombinant tissue plasminogen activator thrombolysis of acute stroke. J Cereb Blood Flow Metab. 1998;18:1298-1307
  8. Barber PA, Darby DG, Desmond PM, Yang Q, Gerraty RP, Jolley D, Donnan GA, Tress BM, Davis SM. Prediction of stroke outcome with echoplanar perfusion- and diffusion-weighted mri. Neurology. 1998;51:418-426
  9. Baird AE, Lovblad KO, Dashe JF, Connor A, Burzynski C, Schlaug G, Straroselskaya I, Edelman RR, Warach S. Clinical correlations of diffusion and perfusion lesion volumes in acute ischemic stroke. Cerebrovasc Dis. 2000;10:441-448
  10. Singer OC, Du Mesnil De Rochemont R, Foerch C, Stengel A, Sitzer M, Lanfermann H, Neumann-Haefelin T. Early functional recovery and the fate of the diffusion/perfusion mismatch in patients with proximal middle cerebral artery occlusion. Cerebrovasc Dis. 2004;17:13-20
  11. Hillis AE, Wityk RJ, Barker PB, Ulatowski JA, Jacobs MA. Change in perfusion in acute nondominant hemisphere stroke may be better estimated by tests of hemispatial neglect than by the national institutes of health stroke scale. Stroke. 2003;34:2392-2396
  12. Labeyrie MA, Turc G, Hess A, Hervo P, Mas JL, Meder JF, Baron JC, Touze E, Oppenheim C. Diffusion lesion reversal after thrombolysis: A mr correlate of early neurological improvement. Stroke. 2012;43:2986-2991
  13. Lee JM, Vo KD, An H, Celik A, Lee Y, Hsu CY, Lin W. Magnetic resonance cerebral metabolic rate of oxygen utilization in hyperacute stroke patients. Ann Neurol. 2003;53:227-232
  14. Geisler BS, Brandhoff F, Fiehler J, Saager C, Speck O, Rother J, Zeumer H, Kucinski T. Blood-oxygen-level-dependent mri allows metabolic description of tissue at risk in acute stroke patients. Stroke. 2006;37:1778-1784
  15. Jensen-Kondering U, Baron JC. Oxygen imaging by mri: Can blood oxygen level-dependent imaging depict the ischemic penumbra? Stroke. 2012;43:2264-2269
  16. Zhu G, Michel P, Aghaebrahim A, Patrie JT, Xin W, Eskandari A, Zhang W, Wintermark M. Computed tomography workup of patients suspected of acute ischemic stroke: Perfusion computed tomography adds value compared with clinical evaluation, noncontrast computed tomography, and computed tomography angiogram in terms of predicting outcome. Stroke. 2013;44:1049-1055
  17. Biesbroek JM, Niesten JM, Dankbaar JW, Biessels GJ, Velthuis BK, Reitsma JB, van der Schaaf IC. Diagnostic accuracy of ct perfusion imaging for detecting acute ischemic stroke: A systematic review and meta-analysis. Cerebrovasc Dis. 2013;35:493-501
  18. Agarwal S, Jones PS, Alawneh JA, Antoun NM, Barry PJ, Carrera E, Cotter PE, O’Brien EW, Salih I, Scoffings DJ, Baron JC, Warburton EA. Does perfusion computed tomography facilitate clinical decision making for thrombolysis in unselected acute patients with suspected ischaemic stroke? Cerebrovasc Dis. 2011;32:227-233
  19. Wintermark M, Reichhart M, Thiran JP, Maeder P, Chalaron M, Schnyder P, Bogousslavsky J, Meuli R. Prognostic accuracy of cerebral blood flow measurement by perfusion computed tomography, at the time of emergency room admission, in acute stroke patients. Ann Neurol. 2002;51:417-432
  20. Wintermark M, Flanders AE, Velthuis B, Meuli R, van Leeuwen M, Goldsher D, et al. Perfusion-ct assessment of infarct core and penumbra: Receiver operating characteristic curve analysis in 130 patients suspected of acute hemispheric stroke. Stroke. 2006;37:979-985
  21. Bivard A, McElduff P, Spratt N, Levi C, Parsons M. Defining the extent of irreversible brain ischemia using perfusion computed tomography. Cerebrovasc Dis. 2011;31:238-245
  22. Bivard A, Spratt N, Levi C, Parsons M. Perfusion computer tomography: Imaging and clinical validation in acute ischaemic stroke. Brain. 2011;134:3408-3416
  23. Campbell BC, Christensen S, Levi CR, Desmond PM, Donnan GA, Davis SM, Parsons MW. Cerebral blood flow is the optimal ct perfusion parameter for assessing infarct core. Stroke. 2011;42:3435-3440
  24. Alawneh JA, Jones PS, Mikkelsen IK, Cho TH, Siemonsen S, Mouridsen K, et al. Infarction of ‘non-core-non-penumbral’ tissue after stroke: Multivariate modelling of clinical impact. Brain. 2011;134:1765-1776
  25. Muir KW, Halbert HM, Baird TA, McCormick M, Teasdale E. Visual evaluation of perfusion computed tomography in acute stroke accurately estimates infarct volume and tissue viability. J Neurol Neurosurg Psychiatry. 2006;77:334-339
  26. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med. 1995;333:1581-1587
  27. Hacke W, Kaste M, Bluhmki E, Brozman M, Davalos A, Guidetti D, Larrue V, Lees KR, Medeghri Z, Machnig T, Schneider D, von Kummer R, Wahlgren N, Toni D. Thrombolysis with alteplase 3 to 4.5 hours after acute ischemic stroke. N Engl J Med. 2008;359:1317-1329
  28. Thomalla G, Schwark C, Sobesky J, Bluhmki E, Fiebach JB, Fiehler J, et al. Outcome and symptomatic bleeding complications of intravenous thrombolysis within 6 hours in mri-selected stroke patients: Comparison of a german multicenter study with the pooled data of atlantis, ecass, and ninds tpa trials. Stroke. 2006;37:852-858
  29. Kohrmann M, Juttler E, Fiebach JB, Huttner HB, Siebert S, Schwark C, Ringleb PA, Schellinger PD, Hacke W. Mri versus ct-based thrombolysis treatment within and beyond the 3 h time window after stroke onset: A cohort study. Lancet Neurol. 2006;5:661-667
  30. Schellinger PD, Thomalla G, Fiehler J, Kohrmann M, Molina CA, Neumann-Haefelin T, Ribo M, Singer OC, Zaro-Weber O, Sobesky J. Mri-based and ct-based thrombolytic therapy in acute stroke within and beyond established time windows: An analysis of 1210 patients. Stroke. 2007;38:2640-2645
  31. Ribo M, Molina CA, Rovira A, Quintana M, Delgado P, Montaner J, Grive E, Arenillas JF, Alvarez-Sabin J. Safety and efficacy of intravenous tissue plasminogen activator stroke treatment in the 3- to 6-hour window using multimodal transcranial doppler/mri selection protocol. Stroke. 2005;36:602-606
  32. Albers GW, Thijs VN, Wechsler L, Kemp S, Schlaug G, Skalabrin E, Bammer R, et al. Magnetic resonance imaging profiles predict clinical response to early reperfusion: The diffusion and perfusion imaging evaluation for understanding stroke evolution (defuse) study. Ann Neurol. 2006;60:508-517
  33. Lansberg MG, Lee J, Christensen S, Straka M, et al. Rapid automated patient selection for reperfusion therapy: A pooled analysis of the echoplanar imaging thrombolytic evaluation trial (epithet) and the diffusion and perfusion imaging evaluation for understanding stroke evolution (defuse) study. Stroke. 2011;42:1608-1614
  34. Davis SM, Donnan GA, Parsons MW, Levi C, et al. Effects of alteplase beyond 3 h after stroke in the echoplanar imaging thrombolytic evaluation trial (epithet): A placebo-controlled randomised trial. Lancet Neurol. 2008;7:299-309
  35. Hacke W, Albers G, Al-Rawi Y, Bogousslavsky J, et al. The desmoteplase in acute ischemic stroke trial (dias): A phase ii mri-based 9-hour window acute stroke thrombolysis trial with intravenous desmoteplase. Stroke. 2005;36:66-73
  36. Furlan AJ, Eyding D, Albers GW, Al-Rawi Y, et al. Dose escalation of desmoteplase for acute ischemic stroke (dedas): Evidence of safety and efficacy 3 to 9 hours after stroke onset. Stroke. 2006;37:1227-1231
  37. Hacke W, Furlan AJ, Al-Rawi Y, Davalos A, et al. Intravenous desmoteplase in patients with acute ischaemic stroke selected by mri perfusion-diffusion weighted imaging or perfusion ct (dias-2): A prospective, randomised, double-blind, placebo-controlled study. Lancet Neurol. 2009;8:141-150
  38. Parsons M, Spratt N, Bivard A, Campbell B, et al. A randomized trial of tenecteplase versus alteplase for acute ischemic stroke. N Engl J Med. 2012;366:1099-1107
  39. Lansberg MG, Straka M, Kemp S, Mlynash M, et al. Mri profile and response to endovascular reperfusion after stroke (defuse 2): A prospective cohort study. Lancet Neurol. 2012;11:860-867
  40. Kidwell CS, Jahan R, Gornbein J, Alger JR, et al. A trial of imaging selection and endovascular treatment for ischemic stroke. N Engl J Med. 2013;368:914-923
  41. Kidwell CS, Jahan R, Alger JR, Schaewe TJ, et al. Design and rationale of the mechanical retrieval and recanalization of stroke clots using embolectomy (mr rescue) trial. Int J Stroke. 2014;9:110-116
  42. Baron JC, Macrae IM, Adams HP, Jr., Dirnagl U. Esc-brain: Experimental and clinical stroke research–do they connect?. Meeting report of the esc-brain joint symposium held in london and shanghai in may 2013. Cerebrovasc Dis. 2013;36:306-321
  43. Takasawa M, Jones PS, Guadagno JV, Christensen S, et al. How reliable is perfusion mr in acute stroke? Validation and determination of the penumbra threshold against quantitative pet. Stroke. 2008;39:870-877
  44. Zaro-Weber O, Moeller-Hartmann W, Heiss WD, Sobesky J. Mri perfusion maps in acute stroke validated with 15o-water positron emission tomography. Stroke. 2010;41:443-449
  45. Zaro-Weber O, Moeller-Hartmann W, Heiss WD, Sobesky J. Maps of time to maximum and time to peak for mismatch definition in clinical stroke studies validated with positron emission tomography. Stroke. 2010;41:2817-2821
  46. Warach S, Al-Rawi Y, Furlan AJ, Fiebach JB, et al. Refinement of the magnetic resonance diffusion-perfusion mismatch concept for thrombolytic patient selection: Insights from the desmoteplase in acute stroke trials. Stroke. 2012;43:2313-2318
  47. Kakuda W, Lansberg MG, Thijs VN, Kemp SM, et al. Optimal definition for pwi/dwi mismatch in acute ischemic stroke patients. J Cereb Blood Flow Metab. 2008;28:887-891
  48. Donnan GA, Baron JC, Ma H, Davis SM. Penumbral selection of patients for trials of acute stroke therapy. Lancet Neurol. 2009;8:261-269
  49. Donnan GA, Davis SM. Neuroimaging, the ischaemic penumbra, and selection of patients for acute stroke therapy. Lancet Neurol. 2002;1:417-425
  50. Wintermark M, Albers GW, Broderick JP, Demchuk AM, et al. Acute stroke imaging research roadmap ii. Stroke. 2013;44:2628-2639
  51. Wang Y, Liao X, Zhao X, Wang C, et al. Imaging-based thrombolysis trial in acute ischemic stroke-ii (itais-ii). Int J Stroke. 2009;4:49-53; discussion 49
  52. Mishra NK, Albers GW, Davis SM, Donnan GA, et al. Mismatch-based delayed thrombolysis: A meta-analysis. Stroke. 2010;41:e25-33

To cite: Alawneh J, Baron, J-C.
ACNR 2014;14(3):18-23

Download this Article