19.2 Mechanical Thrombectomy Windows and Imaging Selection (03.C.3)

19.2 Mechanical Thrombectomy Windows and Imaging Selection

Quick Answer: 0–6 hours: ICA or proximal MCA (M1) large-vessel occlusion (LVO) with a disabling deficit — offer endovascular thrombectomy (EVT) on NCCT + CTA; perfusion is not required (MR CLEAN / HERMES). 6–16 hours: DEFUSE 3 perfusion mismatch — core <70 mL, ratio ≥1.8, penumbra ≥15 mL. 6–24 hours: DAWN clinical-core mismatch by age and NIHSS versus core volume. After mTICI 2b–3, do not drive SBP <140 mm Hg for 72 hours. Failed recanalization is a diagnosis (ICAD, dissection, clot) before it is a rescue stent.

A normal “tPA window closed” sentence is not a thrombectomy plan. On this examination, a 7-hour M1 occlusion with a 10 mL core and NIHSS 18 is a DAWN/DEFUSE patient, not a comfort-care patient. Independent OpenExamPrep teaching here covers mechanical thrombectomy windows and imaging selection for acute ischemic stroke in the ABPN Content Specifications disease list. This guide is not an ABIM or ABPN product.

0–6 hours: HERMES and MR CLEAN physiology

In 2015, five trials — MR CLEAN, ESCAPE, EXTEND-IA, SWIFT PRIME, and REVASCAT — established stent-retriever (and later aspiration) thrombectomy plus medical care over medical care alone for anterior-circulation LVO. The patient-level HERMES meta-analysis (Goyal and colleagues) showed a large shift toward better 90-day mRS, with benefit that remained statistically detectable out to about 7 hours 18 minutes from onset to expected puncture, and with each hour of delay after reperfusion costing disability. MR CLEAN was the pragmatic Dutch trial: NIHSS could be low, ASPECTS was not a rigid lockout, and selection was NCCT + CTA-based. Because MR CLEAN and THRACE did not require CTP, you must not refuse 0–6 hour EVT solely because perfusion software is down.

Classic 2018 teaching that still appears in stems: age ≥18, prestroke mRS 0–1, NIHSS ≥6, ASPECTS ≥6, causative occlusion of the intracranial ICA or M1, treatment within 6 hours. The 2026 AHA/ASA early-management guideline broadened early-window EVT: ASPECTS 3–10 within 6 hours is Class 1 for ICA/M1 LVO with NIHSS ≥6 and prestroke mRS 0–1; selected patients with ASPECTS 0–2 may still be reasonable within 6 hours if they are younger and lack massive mass effect (LASTE-style MRI-selected large core). Know both layers: the stem that quotes ASPECTS 8 and NIHSS 16 at 4 hours is an easy yes; the stem with ASPECTS 4 at 3 hours is no longer an automatic no.

Alberta Stroke Program Early CT Score (ASPECTS) is a 10-point NCCT score of the MCA territory. You start at 10 and subtract 1 for early ischemic change in each of: caudate, lentiform, internal capsule, insula, and M1–M6 cortical regions. Lower scores mean more visible early injury. ASPECTS is not a perfusion volume in milliliters. A reader can “fail” ASPECTS by calling chronic small-vessel change an acute M3 hypodensity. Frank, well-established hypodensity that looks subacute should make you re-check last known well — that tissue may be yesterday’s stroke.

Do not delay arterial puncture to watch whether IVT “opens the vessel.” REVASCAT’s observation pause is the historical warning. Do not skip eligible IVT to speed the lab; several skip-lytic trials failed to show that EVT-alone is better, and a meta-analysis found bridging helpful when the lytic can be given early.

DAWN: 6–24 hours, clinical-core mismatch

DAWN (Nogueira and colleagues, NEJM 2018) enrolled ICA or proximal MCA occlusions 6–24 hours after last known well with a mismatch between a severe clinical deficit and a small imaged core. Core was DWI or CTP relative CBF <30%, not a Tmax ratio. Mismatch bins:

Thrombectomy produced 90-day mRS 0–2 in 49% versus 13% (absolute difference about 33 points). sICH and mortality were not significantly different. The NNT for independence was about 2.8. If the stem gives an 82-year-old, NIHSS 12, and a 40 mL core at 18 hours, that patient fails DAWN (≥80 requires core <21 mL). If the stem gives a 65-year-old, NIHSS 22, core 40 mL at 18 hours, that patient fits the third bin.

DEFUSE 3: 6–16 hours, perfusion mismatch

DEFUSE 3 (Albers and colleagues, NEJM 2018) used a target mismatch profile on CTP or MR perfusion, typically processed with automated software (RAPID):

• Ischemic core <70 mL (DWI or rCBF <30%).
• Mismatch ratio ≥1.8 (Tmax >6 second hypoperfusion volume ÷ core).
• Absolute mismatch (penumbra) ≥15 mL.

Window: 6–16 hours. Functional independence mRS 0–2 was 45% versus 17%. About 40% of DEFUSE 3 patients would not have met DAWN, so a 60 mL core with a huge Tmax >6 s volume can still be a DEFUSE patient at 12 hours even if DAWN core caps are exceeded. Conversely, a tiny core and NIHSS 18 at 20 hours is a DAWN patient after the 16-hour DEFUSE clock has expired. Learn both names; do not merge them into “late window perfusion.”

AURORA pooled late-window randomized data and still favored EVT. Subsequent large-core trials (SELECT2, ANGEL-ASPECT, TESLA and related programs) pushed ASPECTS 3–5 into the 6–24 hour conversation for selected patients. The examination still expects you to quote DAWN and DEFUSE 3 as the mismatch trials that opened the late window.

ASPECTS in practice, tandem lesions, and distal vessels

Score ASPECTS on the first adequate NCCT, not after you have already decided. Early loss of insular ribbon plus caudate and lentiform in a dense left MCA syndrome is a typical 7–8, not a 2. A true ASPECTS 2 with swollen hypodensity and midline shift is a malignant infarct conversation (next section), not a “one more pass” conversation unless a trial-like large-core protocol applies and the family understands the disability trade.

Tandem lesions combine cervical ICA stenosis/occlusion/dissection with an intracranial LVO. HERMES included about 10% tandem occlusions; EVT still beat medical care. The intracranial clot is what is killing the hemisphere now. Extracranial management is heterogeneous: angioplasty alone, stenting plus antiplatelets, or leaving a chronic string-sign. Observational syntheses (including TITAN and later meta-analyses) associate emergent carotid stenting with higher reperfusion and better mRS and more ICH — especially after IVT, when dual antiplatelets for a new stent collide with a fibrinolytic. There is no single Class 1 recipe. Exam-safe moves: treat the intracranial occlusion, do not let a 90% proximal ICA stop you from attempting EVT, and if you stent, have an antiplatelet plan that acknowledges bleed risk.

Medium-vessel occlusions (distal M2/M3, A2, P2) looked attractive until ESCAPE-MeVO and DISTAL failed to show better 90-day mRS versus medical care and raised hemorrhage/mortality concern with stent retrievers in milder NIHSS cohorts. Do not export HERMES enthusiasm to every P2 occlusion with NIHSS 4.

Aspiration-first versus stent-retriever-first (COMPASS, ASTER) is a technical tie for adequate reperfusion. mTICI 2b/2c/3 is the radiographic goal (2b = ≥50% of the downstream territory filling). mTICI 0–2a is unsuccessful or inadequate reperfusion.

Post-EVT blood pressure

If the patient received IVT, the ≤180/105 mm Hg ceiling during and for 24 hours still applies and is reasonable after EVT in the 2026 AHA/ASA blood-pressure section. The new high-impact harm statement: in anterior-circulation LVO successfully recanalized (mTICI 2b, 2c, or 3) without another indication for a lower target, intensive SBP <140 mm Hg for the first 72 hours is harmful and not recommended. OPTIMAL-BP and related intensive-lowering trials showed worse disability and more deterioration. Residual tandem stenosis, a threatened perfusion-dependent hemisphere, or unruptured aneurysm management may justify a different number — that is a reason, not a default “keep them 110 systolic because we opened it.” Avoid MAP <70 mm Hg during the case when observational work has linked profound hypotension to worse outcome. Variability is also unfriendly: swinging 80 to 180 is not “autoregulation training.”

Failed recanalization: next steps, not nihilism

When the last run is still mTICI 0–2a, name the why:

• Intracranial atherosclerotic disease (ICAD) with in situ stenosis that recloses as soon as the stent retriever is out.
• Dissection (cervical or intracranial) you can worsen with more pulling.
• Hard clot / fibrin cap that will not engage.
• Access failure (arch anatomy, balloon-guide issues).
• Distal embolization into a previously unaffected territory.

ANGEL-REBOOT (rescue balloon angioplasty/stenting after TICI 0–2a) did not improve 90-day mRS and had more complications in a randomized Chinese cohort, albeit with off-label tirofiban baggage. Observational RESCUE ICAS and matched cohorts look more favorable for rescue stenting. Translation for the examination: rescue intracranial angioplasty/stenting may be considered, especially when the angiogram screams ICAD, but it is not a proven automatic step and dual antiplatelets after a fresh infarct plus lytic is a hemorrhage wager. Adjunctive intra-arterial alteplase after successful EVT looked promising in the small CHOICE trial and then failed to help in POST-TNK, POST-UK, and ATTENTION-IA — so “give IA tenecteplase because TICI is 2c” is not a 2026 mandate.

Medical next steps after a closed vessel: neuro ICU, treat BP without watersheding the remaining collaterals, swallow screen, glucose 140–180 mg/dL rather than 80–130, normothermia, and early hemicraniectomy planning if this is a young complete MCA that will swell (next section). Repeat NCCT for hemorrhage or malignant edema. Do not start full-dose heparin to “keep the stump liquid” as routine therapy for typical arterial occlusion.

Worked stems

NIHSS 17, left M1, ASPECTS 8, last known well 4 hours: EVT now, plus IVT if eligible. No CTP required.

NIHSS 16, right M1, last known well 14 hours, CTP core 12 mL, Tmax >6 s volume 110 mL (ratio >1.8): DEFUSE 3 — offer EVT.

Age 84, NIHSS 11, core 28 mL, hour 20: fails DAWN (core must be <21 mL if age ≥80). Do not quote DEFUSE 3; the 16-hour perfusion window has also passed. This is a goals-of-care plus medical management stem unless another protocol applies.

Age 62, NIHSS 21, core 40 mL, hour 18: DAWN third bin — offer EVT.

mTICI 3 at 22:00, night-shift SBP goal 110 mm Hg “to protect from ICH”: wrong. Intensive <140 mm Hg for 72 hours after successful recanalization is a harm recommendation in 2026 AHA/ASA text.

Exam traps

Requiring CTP at 3 hours. Mixing DAWN bins with DEFUSE ratios. Using 24 hours for DEFUSE 3. Declaring tandem ICA occlusion untreatable. Forcing SBP 100 after TICI 3. Skipping IVT in an eligible 0–4.5 hour patient to “save time.” Treating every M3 occlusion like HERMES. Independent practice items at /practice/abim-neurocritical-care exercise these windows.