Simulating patient selection for mechanical thrombectomy in acute large vessel occlusion stroke — from clinical suspicion to final procedural decision
Every minute an LVO stroke goes untreated, an estimated 1.9 million neurons are lost. Rapid recognition — in the field by EMS, and at the door by the stroke team — is the first and most time-critical gate in the thrombectomy pathway. Clinical severity scores act as a coarse but fast filter, flagging patients likely to have a proximal occlusion before any imaging is available.
Emergency medical services use abbreviated severity scales to triage stroke patients directly to comprehensive stroke centers capable of thrombectomy, bypassing primary centers when appropriate:
• RACE (Rapid Arterial oCclusion Evaluation): 0–9 scale assessing facial palsy, arm/leg motor function, gaze deviation, aphasia/agnosia. Score ≥5 has ~85% sensitivity for LVO. • CPSSS (Cincinnati Prehospital Stroke Severity Scale): simplified 0–4 scale; ≥2 suggests LVO. • LAMS (Los Angeles Motor Scale): facial droop, arm drift, grip strength; ≥4 flags LVO.
On hospital arrival, the NIHSS (0–42) is the formal severity instrument. A score ≥6 substantially raises pretest probability of a proximal occlusion, though it is neither necessary nor sufficient — some LVOs present with lower scores (especially non-dominant hemisphere or posterior circulation), and some non-LVO strokes score higher.
The overriding principle at this stage is speed: a stroke code triggers simultaneous mobilization of the CT scanner, stroke neurology, and the interventional team, targeting door-to-CT times under 20 minutes.
A positive prehospital severity scale should trigger direct transport to (or transfer toward) a thrombectomy-capable center — even before any imaging exists — because the downstream time saved by avoiding a secondary transfer often determines candidacy for the standard window.
Clinical suspicion alone cannot distinguish LVO from stroke mimics, hemorrhage, or distal occlusions unlikely to benefit from mechanical retrieval. CT angiography (CTA), obtained immediately after non-contrast CT, images the full cervical and intracranial arterial tree in seconds and is the definitive confirmatory test that gates every subsequent eligibility decision.
Mechanical thrombectomy trials established benefit specifically for proximal, large-caliber occlusions where a stent-retriever or aspiration catheter can be safely navigated and the clot volume is retrievable:
• ICA terminus ("T" or "L" occlusion): often the most disabling, involves both MCA and ACA territory; strongest thrombectomy benefit. • M1 segment of the MCA: the single most common target, proximal to the major lenticulostriate and cortical branch bifurcation. • Proximal M2: increasingly treated as catheter technology has improved, though evidence is less robust than for M1/ICA. • Basilar artery occlusion: a distinct posterior-circulation LVO now also supported by randomized trial evidence, treated on a separate but analogous pathway.
Distal occlusions (M3, A2, P2 and beyond) generally carry smaller vessel-supplied territories, harder catheter access, and less certain risk-benefit — most protocols exclude them from standard thrombectomy pathways.
A single-phase or multiphase CTA is acquired from the aortic arch through the vertex following a timed contrast bolus. Beyond simply localizing the occlusion, the study is read for:
• Occlusion site and clot length (longer clots are more resistant to first-pass recanalization) • Collateral circulation grade — robust leptomeningeal collaterals slow core growth and can extend the treatable window even when time is uncertain • Vessel tortuosity and access route, relevant to procedural planning • Concurrent stenosis or dissection that may complicate catheter navigation
Good collateral flow is one of the reasons two patients with identical occlusions and identical elapsed times can have very different core volumes — collaterals, not the clock alone, determine the rate of infarct growth.
Not all brain tissue downstream of an occlusion is equally salvageable. The infarct core — tissue with severely reduced blood flow that is already irreversibly injured — is distinguished from the penumbra, which remains at risk but viable if flow is restored promptly. Estimating core size, via ASPECTS on non-contrast CT or via CT-perfusion volumetrics, is the single strongest imaging predictor of whether thrombectomy will help or harm.
The Alberta Stroke Program Early CT Score (ASPECTS) divides the MCA territory on non-contrast CT into 10 standardized regions (caudate, lentiform nucleus, internal capsule, insula, and six cortical M1–M6 regions). One point is subtracted for each region showing early ischemic change (hypoattenuation or loss of grey-white differentiation); a fully normal scan scores 10.
ASPECTS is fast, requires no contrast, and correlates reasonably well with infarct volume — each point lost corresponds roughly to a meaningful increment of infarcted tissue. Scores ≥6 are traditionally considered favorable for thrombectomy in the standard window; scores ≤5 indicate a larger, more established infarct where benefit becomes uncertain (though recent trials have extended treatment to selected low-ASPECTS patients).
CT perfusion (CTP) generates quantitative maps of cerebral blood flow (CBF) and blood volume (CBV) using automated software (e.g., RAPID, Viz.ai). Two volumes are extracted:
• Ischemic core: tissue with relative CBF <30% of normal — treated as non-salvageable • Critically hypoperfused tissue (Tmax >6s): core plus penumbra — tissue at risk but potentially salvageable
The "mismatch volume" (hypoperfused minus core) and "mismatch ratio" quantify how much salvageable tissue remains. Trials used core-volume ceilings — generally <70 mL, with the most favorable outcomes clustering under 50 mL — because beyond that threshold, reperfusing already-dead tissue risks hemorrhagic transformation without a corresponding functional gain.
Core volume, not the clock alone, is the biological boundary of benefit — this is precisely why perfusion imaging can extend eligibility well past 6 hours in patients whose collaterals have kept the core small, while excluding fast progressors who are imaged early but already have a large core.
Time since the patient was last known to be neurologically normal remains a foundational eligibility axis, but it no longer acts alone. Landmark trials redefined the window from a rigid clock-based cutoff into a tissue-based one: within 6 hours, simple imaging suffices; beyond 6 hours, advanced perfusion or clinical-core mismatch imaging is required to identify patients who still have salvageable tissue.
Within 6 hours of last known well, the HERMES meta-analysis of five landmark trials (MR CLEAN, ESCAPE, REVASCAT, SWIFT PRIME, EXTEND-IA) established an overwhelming, consistent benefit of mechanical thrombectomy plus medical therapy over medical therapy alone for anterior-circulation LVO. In this window, eligibility rests mainly on NIHSS, confirmed occlusion site, and a favorable ASPECTS/core — advanced perfusion imaging is not required, though many centers obtain it anyway to inform prognosis.
Two pivotal trials pushed the window well beyond 6 hours by selecting patients on tissue physiology rather than the clock:
• DEFUSE 3 (6–16h): enrolled patients with a large perfusion mismatch — core <70 mL, mismatch ratio ≥1.8, mismatch volume ≥15 mL. Thrombectomy nearly tripled the rate of functional independence at 90 days.
• DAWN (6–24h): used a "clinical-core mismatch" — comparing NIHSS severity (a proxy for how much tissue should be at risk) against the actual measured core volume on perfusion imaging. A patient with a severe deficit but a small core implies a large territory still salvageable, i.e., good collaterals holding the line.
Both trials required the same class of advanced imaging (CTP or diffusion-perfusion MRI) precisely because the risk of treating already-dead tissue rises the longer the clock has run, making core-volume confirmation mandatory rather than optional past 6 hours.
Beyond roughly 24 hours from last known well, no randomized trial has demonstrated benefit, and thrombectomy is generally not offered regardless of imaging findings — the extended window has an outer boundary as well as a lower one.
The final eligibility decision synthesizes every prior stage: is there a proximal, catheter-accessible occlusion; is the infarct core small enough that meaningful tissue remains salvageable; and does the time elapsed fall within a window supported by the corresponding imaging standard. When all three align, mechanical thrombectomy offers one of the largest treatment effects in modern medicine; when they do not, the risks of an invasive procedure can outweigh any plausible benefit.
A patient proceeds to the angiography suite when: (1) a proximal LVO is confirmed by CTA/MRA at ICA, M1, or proximal M2; (2) infarct core remains below the applicable threshold (<70 mL generally, <50 mL for the most favorable extended-window outcomes); and (3) time since onset falls within the standard window with basic imaging, or the extended window with a qualifying perfusion or clinical-core mismatch.
Under conscious sedation or general anesthesia, an interventionalist navigates a catheter from femoral or radial access to the occlusion site and deploys a stent-retriever or aspiration catheter to extract the clot, aiming for a modified Thrombolysis in Cerebral Infarction (mTICI) score of 2b–3 (near-complete to complete reperfusion). Faster, more complete recanalization correlates directly with better 90-day outcomes.
Patients are deemed not candidates for thrombectomy for several distinct reasons, each reflecting an unfavorable risk-benefit calculation rather than a single rule:
• Core too large: once the infarct core exceeds the trial-validated ceiling, reperfusion of dead tissue offers little functional benefit and raises hemorrhagic transformation risk. • Occlusion too distal: M3/M4 or A2/P2-and-beyond occlusions supply smaller territories and are often not safely reachable with current catheter technology; the potential benefit rarely justifies procedural risk. • Beyond the time window: without trial evidence past ~24 hours (or a qualifying mismatch within it), the safety and efficacy of intervention are unproven. • Contraindications: severe coagulopathy, inability to obtain access, or a pre-stroke disability that makes functional recovery unlikely regardless of recanalization.
These patients continue on optimal medical therapy — antiplatelet or anticoagulation per etiology, blood pressure and glycemic management, and secondary prevention — rather than proceeding to an invasive procedure unlikely to help.
The eligibility decision is never a single variable in isolation — a small core past the "standard" 6-hour mark with strong collaterals can still be an excellent thrombectomy candidate, while a large core well within 6 hours may not be. Core, site, and time act together as a single physiological gate.