The New England Journal of Medicine March 12, 2015

Endovascular Therapy for Ischemic Stroke with Perfusion-Imaging Selection

Bruce C.V. Campbell, Peter J. Mitchell, Timothy J. Kleinig, Helen M. Dewey, Leonid Churilov, Nawaf Yassi, Bernard Yan, Richard J. Dowling, Mark W. Parsons, Thomas J. Oxley, et al., for the EXTEND-IA Investigators

Bottom Line

In patients with acute ischemic stroke, proximal large vessel occlusion, and salvageable brain tissue on CT perfusion imaging, endovascular thrombectomy following intravenous alteplase dramatically improves reperfusion, early neurologic recovery, and functional independence compared to alteplase alone.

Key Findings

1. The median percentage of ischemic territory that underwent reperfusion at 24 hours was significantly greater in the endovascular therapy group compared to the alteplase-only group (100% vs. 37%; P<0.001) [4.1.9].
2. Early neurologic improvement at 3 days (defined as a reduction of ≥8 points on the NIHSS or a score of 0 or 1) occurred in 80% of patients in the endovascular group versus 37% in the alteplase-only group (P=0.002).
3. Functional independence (modified Rankin scale score of 0-2) at 90 days was significantly higher in the endovascular group compared to the control group (71% vs. 40%; P=0.01).
4. There were no significant differences in the rates of symptomatic intracerebral hemorrhage (0% in the endovascular group vs. 6% in the control group; P=0.49) or death at 90 days (9% vs. 20%; P=0.18).

Study Design

Design
Randomized Controlled Trial
Open-Label, Blinded-Endpoint
Sample
70
Patients
Duration
90 days
Median
Setting
Multicenter, Australia/NZ
Population Adult patients with acute ischemic stroke who were receiving IV alteplase within 4.5 hours of symptom onset, had an internal carotid or middle cerebral artery occlusion, and showed evidence of salvageable brain tissue with an ischemic core of <70 mL on CT perfusion imaging.
Intervention Endovascular thrombectomy using the Solitaire FR stent retriever after administration of standard intravenous alteplase (0.9 mg/kg).
Comparator Standard intravenous alteplase alone (0.9 mg/kg).
Outcome Co-primary outcomes: Percentage of the ischemic territory that had undergone reperfusion at 24 hours, and early neurologic improvement (reduction of ≥8 points on the NIHSS or a score of 0 or 1 at day 3).

Study Limitations

The trial was stopped early for efficacy after only 70 patients were randomized (triggered by the release of the MR CLEAN results), which limits subgroup analyses, introduces potential for overestimating the treatment effect size, and reduces the precision of safety endpoint estimates.
The initial treatment assignment could not be blinded (open-label design), although outcome assessors were blinded.
Stringent selection criteria requiring advanced CT perfusion imaging to identify a 'dual target' (large vessel occlusion plus a specific mismatch profile with ischemic core <70 mL) limits the generalizability of the findings to patients lacking favorable perfusion profiles or presenting to centers without rapid CT perfusion capabilities.

Clinical Significance

EXTEND-IA cemented the role of advanced neuroimaging—specifically CT perfusion—in selecting patients for endovascular therapy. By demonstrating an unprecedented 71% rate of functional independence, the trial validated that combining rapid, targeted patient selection (small ischemic core and viable penumbra) with newer-generation stent retrievers yields dramatic clinical benefits. This firmly established mechanical thrombectomy as the standard of care for properly selected patients with acute anterior circulation large vessel occlusions.

Historical Context

Prior to 2014, major stroke trials (such as IMS III, MR RESCUE, and Synthesis Expansion) failed to prove the efficacy of endovascular therapy over standard medical management. These failures were largely attributed to older, less effective devices, slow times to reperfusion, and unrefined patient selection that included patients with massive unsalvageable infarcts. In late 2014, the MR CLEAN trial revolutionized stroke care by demonstrating a clear benefit for endovascular therapy. EXTEND-IA, published concurrently with the ESCAPE trial in early 2015, built upon MR CLEAN by integrating strict CT perfusion imaging criteria and newer-generation stent retrievers, yielding even larger treatment effects and definitively shifting global acute stroke guidelines.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

What is the physiological concept of the ischemic penumbra, and how does CT perfusion imaging differentiate it from the ischemic core in acute stroke?

Key Response

The core represents irreversibly infarcted tissue marked by severely reduced cerebral blood flow and volume. The penumbra is hypoperfused but salvageable tissue, marked by increased mean transit time with relatively preserved blood volume due to autoregulation. Understanding this mismatch is foundational to grasping why reperfusion therapies like thrombectomy are time-critical and beneficial only when salvageable tissue remains.

Resident
Resident

A patient presents with an MCA stroke 3 hours after symptom onset. If CTA confirms an M1 occlusion, what are the immediate next steps according to EXTEND-IA protocols, and why is IV alteplase administered before thrombectomy rather than proceeding straight to the cath lab?

Key Response

Residents must know that IV alteplase should not be delayed to wait for endovascular therapy (EVT), nor should it be skipped if EVT is planned in eligible early-window patients. Alteplase treats distal emboli and begins dissolving the clot, while EVT mechanically removes the proximal large vessel occlusion that alteplase alone usually fails to clear.

Fellow
Fellow

EXTEND-IA required an ischemic core volume of less than 70 mL. How does this strict perfusion imaging criteria impact the number needed to treat (NNT) for functional independence, and what are the limitations of applying this strict cutoff in real-world clinical practice?

Key Response

By highly selecting patients with small cores and large penumbras, EXTEND-IA achieved a massive effect size and an exceptionally low NNT of around 3 for functional independence. However, fellows must recognize that strictly requiring advanced CTP may delay care or exclude patients with larger cores (e.g., ASPECTS 3-5) who might still derive some meaningful clinical benefit from EVT, a nuance later explored in trials like SELECT2 and RESCUE-Japan.

Attending
Attending

As an attending leading a stroke code, how do you balance the time delay of acquiring advanced CT perfusion imaging against the principle of time is brain when deciding on EVT for a patient presenting within 6 hours of onset?

Key Response

While EXTEND-IA highlights the utility of perfusion imaging, attendings must synthesize workflow efficiency. For patients within the early 0-6 hour window, standard non-contrast CT (evaluating ASPECTS) and CTA are often sufficient to proceed with EVT. Over-reliance on CTP in the early window can cause unnecessary delays, though it remains absolutely crucial for late-window (6-24 hr) selection.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The EXTEND-IA trial was terminated early due to overwhelming efficacy after only 71 patients were randomized. What are the statistical risks of truncating a randomized controlled trial early for benefit, particularly regarding the estimation of treatment effect size?

Key Response

Trials stopped early for benefit are known to systematically overestimate treatment effects due to random high bias. Researchers must critically evaluate whether the massive odds ratios seen in EXTEND-IA (such as the dramatically higher reperfusion rates) reflect the true population parameter or are inflated by early stopping rules, which impacts sample size and power calculations for subsequent trials.

Journal Editor
Journal Editor

From a peer-review perspective, how does the exclusive use of a single device (the Solitaire FR stent retriever) and specific automated perfusion software (RAPID) impact the external validity and generalizability of the trial?

Key Response

A rigorous reviewer would flag that standardizing the intervention and imaging to proprietary products minimizes confounding and maximizes internal validity, but inherently limits generalizability. It raises questions about whether the observed benefits are class-effects of all stent retrievers and aspiration catheters, and whether centers without access to automated RAPID software can achieve similar imaging selection accuracy.

Guideline Committee
Guideline Committee

Given the findings of EXTEND-IA alongside similar pivotal trials (like MR CLEAN and ESCAPE), how should AHA/ASA guidelines grade the recommendation for EVT in early-window (0-6 hours) acute ischemic stroke, and is perfusion imaging strictly mandated for this subgroup?

Key Response

EXTEND-IA contributed to a Class 1, Level of Evidence A recommendation for EVT in large vessel occlusion strokes within 6 hours. However, the committee determined that while EXTEND-IA used CTP to definitively prove the concept of salvageable tissue, advanced perfusion imaging is not mandatory for the 0-6 hour window (ASPECTS greater than or equal to 6 on non-contrast CT is sufficient) to avoid delaying reperfusion, whereas advanced imaging is required for the 6-24 hour extended window.

Clinical Landscape

Noteworthy Related Trials

2015

MR CLEAN

n = 500 · NEJM

Tested

Endovascular treatment plus usual care

Population

Patients with acute ischemic stroke and proximal arterial occlusion within 6 hours

Comparator

Usual care alone

Endpoint

Modified Rankin scale score at 90 days

Key result: Endovascular therapy significantly increased the rate of functional independence at 90 days compared to usual care alone.
2018

DAWN Trial

n = 206 · NEJM

Tested

Thrombectomy plus standard medical care

Population

Patients with acute ischemic stroke 6 to 24 hours after onset with clinical-core mismatch

Comparator

Standard medical care alone

Endpoint

Utility-weighted modified Rankin scale score at 90 days

Key result: Thrombectomy yielded significantly better functional outcomes at 90 days for patients with late-window stroke and a clinical-core mismatch.
2018

DEFUSE 3

n = 182 · NEJM

Tested

Endovascular thrombectomy plus standard medical therapy

Population

Patients with acute ischemic stroke 6 to 16 hours after onset with favorable perfusion imaging profile

Comparator

Standard medical therapy alone

Endpoint

Modified Rankin scale score at 90 days

Key result: Endovascular therapy led to a higher rate of functional independence and better overall functional outcomes compared to medical therapy.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis