The ischemic penumbra: what it is and why trials select on it

When an artery supplying the brain is blocked, the tissue it feeds does not all die at once. At the center of the affected territory, blood flow falls so low that cells die quickly. Around that center is a zone receiving enough blood to stay alive but not enough to work. That surrounding zone is the ischemic penumbra, and it can recover if blood flow returns in time, or be lost if it does not.

The idea is more than forty years old, and it now sits underneath some of the most consequential decisions in acute stroke research. This article explains where the concept came from, how researchers estimate something that cannot be seen directly, why it became the basis for selecting patients in two landmark trials, and what the estimate cannot tell anyone. It is educational and not a substitute for personal medical advice.

What is the ischemic penumbra?

A few terms first. Ischemia means an inadequate blood supply to tissue. An infarct is tissue that has died because of that inadequate supply. Reperfusion is the restoration of blood flow, whether it happens on its own or through treatment.

With those terms in place, the picture has two regions. The infarct core is tissue that has already died or is committed to dying. The penumbra is tissue in which blood flow has dropped below the level needed for normal function but not yet below the level at which cells are irreversibly damaged. A review in Cerebrovascular Diseases by Heiss (2011) describes exactly this: two flow thresholds established in animal experiments, one for function and a lower one for infarction, with the perfusion range between them termed the penumbra.

Researchers also describe a third zone further out, sometimes called benign oligemia: tissue with mildly reduced flow that is expected to recover on its own. A systematic review in Stroke by Bandera and colleagues (2006) frames blood flow thresholds as the tool meant to discriminate among all three: infarct core, penumbra, and benign oligemia.

Where did the idea come from?

The concept was named in a short 1981 paper in Stroke, Astrup, Siesjö, and Symon, "Thresholds in cerebral ischemia: the ischemic penumbra". As later reviews summarize it, the insight was that the brain has more than one critical level of blood flow. Below one, electrical function fails. Below a lower one, tissue is irreversibly damaged. Tissue between the two is silent but alive.

That distinction matters because silent tissue still produces deficits. In the animal experiments behind the concept, Heiss notes, restoring flow above the functional threshold reversed the deficits without permanent damage.

Later experimental work added a second dimension: time. Heiss's review summarizes the finding that irreversible damage depends on the interaction of how severely and how long flow is reduced, so that the lower the flow, the shorter the time available for effective reperfusion. As a result, infarction spreads outward from the core into areas of less severe hypoperfusion.

This is why the penumbra is better thought of as a window than as a fixed place. Because damage depends on both how deep and how long the drop in flow is, no single clock applies to every patient.

How do researchers estimate the penumbra in a living patient?

No one can measure the penumbra directly at the bedside. It has to be inferred from imaging, and the inference is built from two separate measurements.

• Perfusion imaging, using CT or MRI, maps where blood flow or the timing of blood delivery is abnormal. This estimates the full territory at risk.
• Core estimation, for example with diffusion-weighted MRI, identifies tissue that already appears irreversibly injured.

The penumbra is then estimated as the mismatch between the two, and Heiss notes that this diffusion and perfusion MRI mismatch is the indicator of the penumbra used as a widely applicable clinical tool.

Clinical trials turned that general idea into concrete, measurable criteria. The DEFUSE 3 trial, reported by Albers and colleagues in the New England Journal of Medicine (2018), enrolled patients with a blockage of the proximal middle cerebral artery or internal carotid artery who had "remaining ischemic brain tissue that was not yet infarcted." In practice that meant an initial infarct smaller than 70 ml and a ratio of ischemic tissue volume on perfusion imaging to infarct volume of 1.8 or more.

The DAWN trial, reported by Nogueira and colleagues in the same journal (2018), used a different kind of mismatch. It enrolled patients whose clinical deficit was disproportionately severe relative to their infarct volume, with the mismatch criteria defined separately for patients younger than 80 and those 80 or older. Instead of comparing two images, DAWN compared how impaired a patient was with how much tissue had already died.

Both definitions are best read as operational choices made for trials, not as laws of nature, and they are not the same as each other.

Why did the penumbra change who was studied for late thrombectomy?

Endovascular thrombectomy is the mechanical removal of a clot through a catheter threaded to the blocked artery. When DEFUSE 3 was published, its authors noted that thrombectomy was recommended for eligible patients treated within 6 hours of symptom onset. Both trials asked what happens later, and both answered by selecting patients on evidence of salvageable tissue rather than on the clock alone.

The results in the selected patients were large.

• DEFUSE 3 treated patients 6 to 16 hours after they were last known to be well. It was stopped early for efficacy after 182 patients were randomized across 38 U.S. centers. At 90 days, 45 percent of the thrombectomy group were functionally independent, compared with 17 percent of the group receiving medical therapy alone. Mortality at 90 days was 14 percent versus 26 percent (P=0.05). Symptomatic intracranial hemorrhage, bleeding into the brain that causes symptoms, occurred in 7 percent versus 4 percent, a difference that was not statistically significant.
• DAWN treated patients 6 to 24 hours after they were last known to be well. Enrollment of 206 patients stopped after a prespecified interim analysis. At 90 days, 49 percent of the thrombectomy group were functionally independent, compared with 13 percent of the control group. Symptomatic hemorrhage (6 percent versus 3 percent) and 90-day mortality (19 percent versus 18 percent) did not differ significantly.

In both trials, functional independence meant a score of 0 to 2 on the modified Rankin scale, a disability measure running from 0 to 6, with higher scores indicating greater disability.

A careful reader should take a precise lesson from these results. The trials did not show that time stopped mattering. They showed that among patients selected for salvageable tissue by these criteria, removing the clot improved outcomes even many hours after onset.

The field has continued to build on this. The 2026 American Heart Association and American Stroke Association guideline for the early management of acute ischemic stroke (Prabhakaran and colleagues, Stroke) replaces the 2018 guideline and its 2019 update, and its authors list new evidence on the determination of eligibility for endovascular thrombectomy among the key updates. A 2026 review for radiologists in the American Journal of Neuroradiology (Rai and colleagues) notes that the guideline retains many foundational recommendations while introducing updates spanning imaging and reperfusion strategies.

What can the penumbra estimate not tell us?

It is a model-based estimate. The boundaries drawn on a perfusion map depend on the thresholds chosen and the software used. In their 2006 systematic review, Bandera and colleagues found that the optimal reported blood flow thresholds varied widely across studies, from 14.1 to 35.0 mL per 100 g per minute for penumbra and from 4.8 to 8.4 for infarct core. They noted that the thresholds were based on animal studies and that their diagnostic accuracy in humans had never been established. That review predates the 2018 trials, but the underlying point about thresholds has not disappeared.

The boundary moves and varies by tissue. A 2018 review in the Journal of Cerebral Blood Flow and Metabolism by Leigh and colleagues reported that defining the outer boundary of the penumbra based solely on perfusion thresholds may not be sufficiently accurate, and that thresholds differ between white and gray matter and change over time.

MRI and the reference standard do not always agree. Heiss's 2011 review notes that studies comparing MRI mismatch with positron emission tomography, a more direct but less practical way of measuring blood flow and oxygen use, pointed to overestimation of both core and penumbra by MRI.

Trial results describe the people who were selected. DEFUSE 3 and DAWN say little about patients who did not meet their criteria. Whether people with a larger established core can benefit is a different question that needs its own trials.

Tissue is not destiny. Two people with similar imaging can recover differently. The numbers above are group results under trial conditions, not predictions for any individual.

Questions readers ask

Is the penumbra the part of the brain that will recover? Not necessarily. It is tissue that could recover if blood flow is restored in time. Whether it does depends on how severe and how prolonged the reduced flow is, and on treatment.

Can the penumbra be seen on a routine CT scan? Not directly. Estimating it requires perfusion imaging or specific MRI sequences, and the result is a model of tissue status rather than a photograph of it.

Does having a penumbra mean someone is eligible for treatment? Eligibility is a clinical decision made by the treating team under current guidelines, weighing many factors beyond imaging. This article explains the research concept only.

Readers interested in how outcomes after stroke are measured and studied can explore the stroke outcomes research program on this site, along with the peer-reviewed publications behind it.