Large-Core Stroke Thrombectomy: What Changed

For years, thrombectomy evidence was strongest for people with a proximal large-vessel occlusion and a relatively small established infarct. A low Alberta Stroke Program Early Computed Tomography Score, or ASPECTS, and a large estimated core were often treated as reasons to withhold the procedure. The logic was understandable: if much tissue was already irreversibly injured, reperfusion might add hemorrhage and procedure risk without enough brain left to save.

A series of randomized trials changed that boundary. RESCUE-Japan LIMIT reported in 2022. SELECT2 and ANGEL-ASPECT reported together in 2023, followed by TENSION and LASTE. Across differing populations and imaging strategies, assignment to thrombectomy shifted 90-day functional outcomes in a favorable direction, and the 2026 AHA and ASA acute ischemic stroke guideline incorporates this evidence into expanded recommendations for selected patients with large cores.

Large core is defined by an imperfect image#

ASPECTS divides the middle cerebral artery territory into ten regions on noncontrast CT. One point is subtracted for each region with early ischemic change, so a lower score indicates a larger affected area. The score is fast and widely available but depends on reader, scan quality, timing, and the distribution of injury.

Perfusion CT and diffusion-weighted MRI estimate core volume using signal thresholds. These maps can overestimate or underestimate final irreversible injury, especially very early, after partial reperfusion, with motion, or at extreme flow values. Different software can classify the same scan differently.

“Large core” therefore described overlapping but nonidentical trial groups. Some trials used ASPECTS 3 to 5. Some required a volume threshold. Some allowed either route. This variation strengthens the broad conclusion that benefit is not limited to one software cutoff, while it complicates application at the edges.

What thrombectomy changes biologically#

An arterial clot can leave a central region with critically low flow and a surrounding region receiving enough collateral blood to remain viable temporarily, and core estimates attempt to identify tissue already severely injured. The penumbra concept identifies threatened tissue that may be saved.

These are dynamic states, not sharply bounded compartments. Core can grow while the artery remains occluded. Successful thrombectomy restores flow to downstream tissue and can limit further injury. It cannot revive neurons already destroyed or erase swelling and network damage. So the decision is not “dead tissue versus normal brain.” It is whether reperfusion is likely to preserve enough remaining tissue and function to outweigh procedure, hemorrhage, edema, and reperfusion risks.

RESCUE-Japan LIMIT opened the large-core question#

RESCUE-Japan LIMIT enrolled patients with internal carotid or proximal middle cerebral artery occlusion and ASPECTS 3 to 5, selected mainly with MRI in Japan, and thrombectomy increased the proportion achieving a modified Rankin Scale score of 0 to 3 at 90 days compared with medical care.

The trial also raised safety and transport questions. Any intracranial hemorrhage was more frequent after thrombectomy, while symptomatic hemorrhage estimates were imprecise. Imaging practice and alteplase dosing in Japan differed from some other systems. The result was promising rather than a universal mandate. Its importance was to challenge the assumption that a large infarct eliminated meaningful benefit. The next trials tested broader settings and imaging strategies.

SELECT2 broadened geography and imaging#

SELECT2 enrolled 352 patients with proximal anterior-circulation occlusion within 24 hours of last known well. Eligibility included ASPECTS 3 to 5 on noncontrast CT or a core volume of at least 50 mL on perfusion CT or diffusion MRI. Sites spanned several countries.

The trial stopped early after interim analysis for efficacy. The primary ordinal analysis showed better modified Rankin Scale outcomes with thrombectomy. Functional independence remained uncommon in both groups, but more patients moved toward less severe disability.

SELECT2 did not impose an upper core-volume limit in the main criterion, producing information at large volumes, though few patients occupied the most extreme ranges and confidence became wider, and subgroup analyses did not establish a sharp point at which benefit disappears. You cannot turn them into a promise for every extreme core. One-year follow-up later supported persistence of the functional advantage, with the usual caveats from longer follow-up and incomplete observations.

ANGEL-ASPECT confirmed benefit in China#

ANGEL-ASPECT randomized 456 patients in China within 24 hours. The design used combinations of ASPECTS and core volume to enrich for large infarction while handling very early scans and discordant imaging.

Endovascular therapy produced a favorable shift in 90-day modified Rankin scores. Intracranial hemorrhage was more common with thrombectomy, while symptomatic hemorrhage was not statistically different in the primary report. The exact definitions and confidence intervals matter more than the statement that one category was “significant.”

The trial showed that the result was not confined to the health systems in SELECT2. It also demonstrated how selection rules can vary. Do not reduce both trials in your head to a single core threshold.

TENSION showed that noncontrast CT can select patients#

TENSION focused on established large infarct defined mainly by ASPECTS 3 to 5 and treated patients within 12 hours; the trial enrolled across European and Canadian centers and was stopped early for efficacy.

Thrombectomy improved functional outcome distribution and reduced mortality in the primary report. Selection could be made using noncontrast CT in a substantial real-world stroke-system context, which matters if perfusion imaging is unavailable where you are, or would delay transfer.

Early stopping can overestimate effect size, particularly in smaller trials. TENSION adds to, rather than replaces, the larger evidence pattern.

LASTE tested very low ASPECTS#

LASTE enrolled patients with very large infarcts, including ASPECTS 0 to 5, within roughly seven hours of last known well, and the trial included older adults and did not use a fixed upper core-volume limit.

Thrombectomy improved the disability distribution and reduced mortality, while symptomatic hemorrhage occurred in both groups. The result extended evidence into imaging severity previously regarded as largely futile.

Very low ASPECTS still signals poor prognosis. An average treatment benefit can coexist with a high probability of major disability or death. When you talk this through with a family, distinguish “better than without thrombectomy” from “likely to return to baseline.”

The ordinal outcome captures shifts that independence misses#

The modified Rankin Scale ranges from no symptoms to death. Trials often analyze the full ordered distribution rather than only a cutoff such as 0 to 2.

In large-core stroke, complete independence is difficult to achieve. Moving from death to survival with severe disability, or from severe dependence to the ability to walk with help, may matter greatly. An ordinal analysis can capture these shifts.

The scale is coarse and inter-rater variability exists. The value of a shift depends on the person's goals and the specific transition; trial reports should pair the common odds ratio with absolute outcome distributions, so that you and the family can see what actually changed.

Benefit does not erase procedural risk#

Thrombectomy requires arterial access, catheter navigation, clot retrieval or aspiration, contrast, and an experienced team. Risks include vessel injury, distal embolization, groin complications, contrast reactions, failed reperfusion, and intracranial hemorrhage.

Reperfusion of severely injured tissue can contribute to hemorrhagic transformation and edema. Large infarcts already carry substantial swelling risk regardless of treatment. Stroke-unit care, blood-pressure management, repeat imaging, neurosurgical assessment, and rehabilitation remain important after the artery is opened.

Safety comparisons depend on definitions. “Any hemorrhage” includes asymptomatic imaging findings. “Symptomatic intracranial hemorrhage” requires neurological worsening under a specified definition. Both should be reported.

The 2026 guideline changes the default question#

The 2026 AHA and ASA guideline treats endovascular thrombectomy as beneficial for selected adults with anterior-circulation large-vessel occlusion and large established infarction under defined imaging, time, and clinical criteria. Exact recommendation class varies across combinations of ASPECTS, time, and other features.

The practical shift is from categorical exclusion toward rapid eligibility assessment, and a low ASPECTS alone should not automatically end transfer or thrombectomy consideration when a patient otherwise resembles evidence-supported groups.

The guideline also emphasizes systems of care. Imaging and transfer workflows should identify candidates without unnecessary delay. “Large core” patients can be early rapid progressors; a 24-hour trial window does not mean waiting is harmless.

Where uncertainty remains#

Evidence is thinner for some extreme core volumes, extensive edema, mass effect, severe pre-stroke disability, distal or unusual occlusions, and very late presentations. Trial participants had defined baseline characteristics and treatment at capable centers. Procedural quality affects benefit.

ASPECTS regions are not equal in functional importance. Dominant hemisphere, eloquent cortex, age, collateral status, and baseline goals can change the meaning of an image. Post hoc subgroup estimates cannot reliably identify an individual treatment effect. Research continues on simplified imaging, transfer, anesthesia, blood pressure, adjunct medicines, and outcome prediction. These questions refine delivery after the major boundary moved.

Stroke recognition still comes first#

Face droop, arm weakness, speech difficulty, sudden vision loss, severe imbalance, or another abrupt focal neurological change requires emergency services. Record the last time you knew the person was normal. Do not drive, give food or drink, or wait for symptoms to resolve.

Thrombolysis and thrombectomy have different criteria and can be complementary. Only emergency imaging can identify hemorrhage, occlusion, infarct burden, and a safe path. The article on time and stroke evidence explains why every workflow minute matters.

The 2023 trials did not make a large infarct small. They showed that preserving the remaining brain can still improve the distribution of life after stroke. That is a meaningful change, provided hope remains calibrated to the evidence.

References#

1. 2026 AHA and ASA acute ischemic stroke guideline
2. SELECT2 trial
3. ANGEL-ASPECT trial
4. RESCUE-Japan LIMIT trial
5. TENSION trial
6. LASTE trial

For your own health, talk with your clinician.*