Acute ischemic stroke imaging is one of the leading causes of death and disability worldwide. Neuroimaging plays a crucial role in early diagnosis and yields essential information regarding tissue integrity, a factor that remains a key therapeutic determinant. Given the widespread public health implications of stroke and central role of neuroimaging in overall management, acute stroke imaging remains a heavily debated, extensively researched, and rapidly evolving subject. There has been recent debate in the scientific community due to divided opinions on the use of CT perfusion and access-related limitations of MRI. In this paper we review and summarize recent updates relevant to acute stroke imaging and propose an imaging paradigm based on the recently available evidence.
Acute ischemic stroke is one of the leading causes of mortality and morbidity worldwide. Statistics from the American Heart Association estimate an average of 1 stroke every 40 seconds in the United States amounting to approximately 795,000 people experiencing new or recurrent strokes, per year [
There is a critical need for reproducible and sensitive imaging biomarkers that allow accurate assessment of efficacy of rapidly evolving thrombolytic treatments. This underscores the primary need for standardization of imaging techniques across institutions so data from multicenter trials can be collectively analyzed. The glaring lack such consensus amongst imaging techniques was highlighted in a recent systematic review which found wide variability in the employed thresholds for CT and MR perfusion imaging and significant inconsistency in definitions of tissue states; factors which add to the widespread variability in perfusion-based assessment [
Despite the inherent challenges and past failures, stroke imaging is rapidly evolving with enormous ongoing research and global public health impact. In this paper we sought to review recent cumulative evidence including evolving expert opinions and recommendation to assess the adequacy of current state of clinical practices in acute stroke imaging. Based on our assessment we propose an optimal imaging paradigm for patients presenting with suspected acute ischemic stroke.
The initial step in approaching the imaging paradigm is to summarize the targeted clinical goals for patients with suspected acute ischemic stroke in the acute care setting. The fundamental objective of treatment is to enable rapid reperfusion for maximal tissue salvation. There is substantial evidence to suggest efficacy of intravenous thrombolytic therapy in the first 4.5 hours from onset of symptoms as well as increased risk of hemorrhagic complications and lower efficacy outside the therapeutic window. The European Cooperative Acute Stroke Study (ECASS) investigators demonstrated the efficacy of treatment instituted within the first 4.5 hours [
Based on the above considerations the following goals must be achieved to allow early initiation of treatment. The choice of imaging approach and interpretation protocol should be designed with the intent of addressing the primary clinical goals such as to allow safe and prompt initiation of thrombolytic strategies.
Initial evaluation focuses on exclusion of primary intracerebral hemorrhage (PICH), intracranial metastasis, tumor with herniation, or other alternate etiologies explaining the clinical picture.
Theoretically Goals 1 and 2 can be assessed on NECT/CTA combination, which is the most ideal single-step imaging solution. Guidelines published by American Heart Association and American Stroke Association Stroke Council in 2007 mandate universal and immediate availability of nonenhanced head CT (NECT) within 30 minutes of initial presentation to the ER [
There is no controversy regarding the utility of NECT with regards to accomplishing Goal 1. In particular it is widely accepted that NECT can reliably exclude intracranial hemorrhage, which is critical for therapeutic decision making.
In terms of infarct characterization to address Goal 2, NECT-based scoring system designed by the Alberta Stroke Program, commonly referred to as the ASPECTS scoring system, (Alberta Stroke Program Early CT Score) provides an effective tool for quantifying early ischemic changes in the MCA territory (Figure
Obtained from
There is conclusive evidence regarding the sensitivity of MRI with diffusion-weighted sequences for the detection of infarct core including those cases in which the infarct core remains occult on standard T2-weighted imaging [
Recent considerations for the same were reviewed by R. Gonzalez indicating a significant role of MRI with diffusion-weighted sequences in identifying infarct core and allowing assessment of core volume, which is a useful predictor of treatment efficacy [
A recent study demonstrated that posttreatment final infarct volume (FIV) also has significant influence on clinical outcome in patients undergoing IAT [
Additionally, there are accessibility issues due to individual contraindications to MR and limited availability of MRI leading to underutilization of MRI in emergent setting. A recent study evaluated the adherence to AAN guidelines of preferring MR over CT in the initial 12 hours of presentation and revealed that the target was met in less than 1/3 of patients in their study [
Perfusion imaging either CT or MR is most relevant in terms of ability to delineate the ischemic penumbra. The clinical utility of penumbra imaging has long remained an issue of debate. The hypothesis of penumbra identification is that identification of “at-risk” tissue may allow widening of the treatment window beyond 4.5 hours and allow detection of patients who will either benefit from treatment or those in whom treatment is not likely to cause improved outcome.
It is notable that the recent study by the MR Rescue investigators found no role of penumbra imaging in selecting patients likely to benefit from endovascular therapy within 8 hours from onset of symptoms. There was evidence of good functional outcome in patients with favorable penumbral pattern in the late time window regardless of recanalization. Interestingly, this study raises the possibility that patients who have a favorable penumbral pattern may be inherently more resilient to the effects of vascular occlusion and therefore harbor a favorable outcome regardless of treatment, thus explaining lack of differential effect of therapy when stratified on the basis of penumbral pattern [
Nevertheless there remains an interest in imaging penumbra due to its potential role as a prognostic biomarker. Previously, CT perfusion performed soon after the initial NECT has been supported as being a safe and efficacious strategy for imaging tissue at risk [
Reliance on postprocessing, restricted brain coverage, and vendor related-differences in processing algorithm are primary limitations which have not yet been completely addressed and remain as mitigating factors in enabling wider utility of CTP. Additionally the inherent low contrast to noise ratio increases susceptibility to artifacts and lowers overall sensitivity. These considerations were reviewed by R. Gonzales and for the same reasons the clinical utility of CTP was felt to be doubtful in the current state of practice [
On a contrary note, a recent expert commentary by M. Lev acknowledges the aforementioned CTP limitations but continues to endorse this method due to its relative cost efficacy, rapid availability, and potential for quantitative assessment relative to MRI [
MR perfusion parameters are equally sensitive in depicting tissue at risk although expense, lack of universal applicability and access issues in the ER setting remain limiting factors. Recently the applicability of MR Perfusion was reviewed by M. Fisher, highlighting its utility in delineating tissue at risk of infarction and thus holding promise in expansion of the therapeutic window. A key aspect underscored in this review relates to the identification of “benign oligemia,” which refers to hypo perfused tissue which will not proceed to infarction regardless of treatment. The MR perfusion parameter
CT angiography continues to be the superior method for characterization of vascular anatomy. Figure
(a) and (b) ADC map and DWI map with restricted diffusion in the setting of cytotoxic edema from acute ischemic infarct in right MCA territory. (c) NECT showing hyperdense right MCA compatible with acute thrombosis. (d) CTA image with thrombosis in the corresponding segment of right MCA. (e), (f), and (g) CTP with elevated MTT, reduced cerebral blood flow, and blood volume in the right MCA territory.
A recent study showed that CTA evidence of occlusion of distal internal carotid, proximal middle cerebral, or basilar arteries as a predictor of poor outcome and added incremental predictive value to NIHSS. This suggests the utility of CTA in early phase treatment decision making [
A potential confounding factor is acquisition protocol-dependent overestimation of infarct core volume using CTA source images for detecting of early ischemic changes [
Overall, CTA has a definite prognostic role in acute phase of stroke imaging. In particular, with relevance to our defined Goal 2, acute-phase CTA can enable assessment of site of occlusion, integrity of vessels in terms of atherosclerotic disease, and degree of collateral flow, all of which influence management decision making (Figure
Despite significant advances in our understanding of physiologic surrogates of imaging observations and the respective technical confounds, there are still considerable debate and lack of consensus particularly with relevance to penumbra imaging and role of perfusion imaging as it relates to core characterization and penumbra estimation.
If we take a minimalist approach, the expert opinion seems to converge most definitively on two standard queries prior to therapeutic decision making (1) is there primary intracranial hemorrhage? and (2) what is the volume of the infarct core? These two components combined with clinical neurological assessment seem to be most directly related to clinical outcome in postperfusion recovery phase and will help stratify patients appropriately for treatment decision-making.
The limitation of the minimalistic approach is that although it raises specificity, by helping us identify those that will have a favorable outcome after-reperfusion, it also at the same time lowers sensitivity, thereby potentially excluding patients who may have benefitted from more aggressive reperfusion therapy.
One of the ways to achieve efficient imaging selection for treatment triage is development of a unimodal imaging protocol. Simplistically stating, a one-stop, all-inclusive imaging protocol can accurately and reproducibly classify patients who will either (a) benefit from treatment or (b) have no impact or negative impact of treatment. NECT is the most obvious choice for Goal 1-related aspects of management. However, beyond that it would be ideal if the CT imaging and interpretation protocols can be optimized in such a way that Goal 2 can be consistently and reliably achieved in acute-phase urgent care setting on the same CT scanner without having to transfer the patient. We believe that in many cases this is feasible, particularly when the ASPECTS score is carefully interpreted, providing infarct core volume information. To assist in the latter, we encourage the use of online resource described by Modi et al. [
However, we acknowledge coexistent evidence that indicates limitation of NECT in core characterization due to multitude of factors including inherently low sensitivity to early ischemic changes. It is agreed upon that core volume is a key determinant of treatment efficacy, for which MR with DWI is the imaging gold standard. In an ideal world with no cost, availability or individual applicability issues, an all inclusive MR protocol for acute stroke imaging would be more preferred from the viewpoint of obtaining accurate and consistent tissue specific information with minimum susceptibility to postprocessing variability.
As stated above we emphasize that if MRI is not available and decision for endovascular therapy has to been taken based on the initial NECT, then immediate CTA (to assess vessel occlusion and collateral status) and CTP (infarct core assessment on rCBF map) should be considered (Figure
The utility of mismatch imaging is undoubtedly promising; however, the recent body of evidence does not provide compelling arguments to necessitate a paradigm shift particularly in routine clinical settings outside of major academic institutions. This is especially true in light of the results from the recent penumbra-based trial of imaging selection by the MR Rescue investigators demonstrating no utility of penumbra imaging in detecting patients who would benefit from endovascular therapy of acute ischemic stroke [
At this time with proven efficacy of IVT within the first 4.5 hours, it is notable that the frequency of thrombolytic therapy in patients with acute ischemic stroke remains remarkably low. A large multicenter study found only 3% utilization of thrombolysis for all acute ischemic stroke patients and only 10% for those who presented within the first 3 hours [