The study drew on the German Stroke Registry–Endovascular Treatment (GSR-ET), a prospective, multicenter registry spanning 25 comprehensive stroke centers across Germany. Between June 2015 and December 2023, more than 18,000 patients were screened, and after excluding those with extracranial carotid occlusions, pre-stroke disability, or missing key clinical data, 5,448 patients treated with mechanical thrombectomy for anterior circulation strokes formed the analytical cohort. The mean age was 71.2 years, 48 percent were women, and the median stroke severity on admission, measured by the National Institutes of Health Stroke Scale (NIHSS), was 13. The researchers stratified patients into nine subgroups based on admission NIHSS (0–10, 11–15, or ≥16) and the Alberta Stroke Program Early CT Score (ASPECTS), a ten-point measure of early infarct signs on CT imaging (10, 8–9, or ≤7).

This dual stratification was designed to capture what neurointerventionalists call clinical-core mismatch. The concept is elegant: a patient whose neurological deficit is dramatically worse than the visible extent of infarction likely has a large penumbra, a zone of struggling but still salvageable brain tissue, whereas a patient whose symptoms are mild relative to a large established infarct has less to gain from reopening the vessel. The DAWN and DEFUSE-3 trials famously demonstrated that patients with pronounced mismatch benefit from thrombectomy even in extended time windows of up to 24 and 16 hours, respectively. But those trials used sophisticated perfusion imaging and volumetric core measurements. The new study instead used the readily available NIHSS and ASPECTS scores as a clinical surrogate, testing whether the same pattern holds in the messy reality of routine practice.

The results were striking. Reperfusion success was graded using the modified Thrombolysis in Cerebral Infarction (mTICI) scale: 650 patients achieved minimal or no reperfusion (mTICI 0–2a), 1,873 achieved partial reperfusion (mTICI 2b), and 2,925 achieved complete recanalization (mTICI 3). Unadjusted outcomes followed a steep gradient: only 20 percent of patients with failed reperfusion achieved functional independence, defined as a modified Rankin Scale (mRS) score of 0 to 2 at 90 days, compared with 46 percent after partial reperfusion and 54 percent after complete recanalization. Mortality at 90 days told the same story in reverse, at 39 percent, 20 percent, and 17 percent respectively. Symptomatic intracranial hemorrhage, a feared procedural complication, occurred in 5 percent of failed reperfusion cases but only 2 percent of complete recanalizations.

Because patients who achieve full recanalization may differ systematically from those who do not, the researchers applied inverse-probability-weighted regression adjustment (IPWRA), a double-robust statistical technique that controls for confounders such as age, sex, comorbidities, baseline NIHSS and ASPECTS, intravenous thrombolysis, and the number of retrieval attempts. The adjusted estimates confirmed the pattern: 25 percent good outcomes after mTICI 0–2a, 46 percent after mTICI 2b, and 52 percent after mTICI 3, with adjusted mean mRS scores of 4.0, 3.0, and 2.7 respectively. Crucially, significant benefit from successful reperfusion was observed in every one of the nine NIHSS-ASPECTS subgroups, including patients with large established infarcts and those with only mild symptoms.

The size of that benefit, however, varied dramatically. The clearest beneficiaries were patients with high clinical-core mismatch, particularly those with a perfect ASPECTS of 10 and severe deficits (NIHSS ≥16). In this group, only 9 percent of patients with failed reperfusion achieved functional independence, compared with 47 percent after complete recanalization, an absolute difference of 38 percentage points. Patients with moderate deficits (NIHSS 11–15) and ASPECTS 10 showed comparable gains. At the opposite extreme, patients with severe deficits and extensive infarct signs (NIHSS ≥16, ASPECTS ≤7) saw a much smaller but still significant benefit, with good outcomes rising from 12 percent to 23 percent, an absolute difference of 11 percentage points. The exploratory mRS shift analysis reinforced the pattern, showing the largest disability reduction, nearly two full mRS points, in the ASPECTS 10 and NIHSS 11–15 subgroup.

One of the most clinically consequential findings concerned the difference between near-complete and complete reperfusion. Pushing from mTICI 2b to mTICI 3 often requires additional retrieval attempts, each carrying a risk of vessel perforation, dissection, or distal clot migration. The data showed that the extra benefit of full recanalization was concentrated in high-mismatch patients: in the NIHSS ≥16, ASPECTS 10 subgroup, complete recanalization lifted good outcomes from 32 percent to 47 percent, a 15-point gain. But in patients with large infarcts (NIHSS ≥16, ASPECTS ≤7), outcomes were essentially identical after mTICI 2b and mTICI 3, both around 20 percent. The authors suggest that in selected low-mismatch cases, accepting mTICI 2b as an adequate procedural endpoint may be a pragmatic choice that avoids unnecessary procedural risk.

The findings also complicate the emerging picture from randomized trials. Recent studies such as SELECT2, ANGEL-ASPECT, RESCUE-Japan LIMIT, and TENSION demonstrated that even patients with large established infarcts benefit from thrombectomy, expanding treatment beyond the classic mismatch paradigm. Meanwhile, a post-hoc analysis of ANGEL-ASPECT failed to show benefit in patients stratified by clinical-radiological mismatch, and recent trials in distal medium vessel occlusions have questioned whether mismatch concepts translate uniformly across vascular territories. The new registry data reconcile some of this tension: mismatch profiles appear to modify the size of the treatment effect rather than determine whether treatment helps at all. In other words, clinical-core mismatch should inform expectations and procedural goals, not serve as a rigid gatekeeper for eligibility.

The authors are careful to note important limitations. Roughly 70 percent of screened patients were excluded in the complete-case analysis because of missing variables, a substantial selection process that may bias effect estimates and limit generalizability to unselected populations. Patients with pre-stroke disability were excluded, local investigators rather than a central core lab graded ASPECTS and mTICI, and unmeasured factors such as collateral circulation, infarct location, and time from onset to imaging could not be fully accounted for. The use of linearized mRS scores in exploratory analyses also assumes equal spacing between scale points, an assumption that may not fully hold. These caveats mean the findings in mild stroke and low-ASPECTS groups should be considered exploratory, even as the high-mismatch results align closely with prior randomized evidence.

Looking forward, the study points toward a more individualized era of stroke treatment. Rather than fixed subgroup thresholds, the authors argue that outcome effects across ASPECTS likely follow a continuous gradient, and future selection tools may integrate advanced imaging biomarkers such as net water uptake, a quantitative measure of early brain swelling, to estimate each patient’s likely benefit from reperfusion. For now, the message for clinicians is nuanced but actionable: successful reperfusion helps virtually every patient profile, but the magnitude of benefit is greatest when a severe deficit signals a large salvageable penumbra, and the pursuit of complete recanalization matters most precisely in those patients. In resource-limited settings where every thrombectomy counts, that distinction could help ensure the right patients receive the most aggressive endovascular effort.

Subject of Research: Reperfusion-dependent outcomes after endovascular thrombectomy for acute ischemic stroke, stratified by NIHSS-ASPECTS clinical-core mismatch

Article Title: Reperfusion‐Dependent Outcomes After Endovascular Thrombectomy Stratified by NIHSS‐ASPECTS Clinical‐Core Mismatch

Article References: Schlicht, F., Meyer, L., Broocks, G., Bechstein, M., Thaler, C., Heitkamp, C., Winkelmeier, L., Geest, V., Heitkamp, A., Jungnitz, M., Peter, G., Meucci, L., Faizy, T., Nawabi, J., Brekenfeld, C., Flottmann, F., Schell, M., Hadjilaou, A., Hanning, U., … German Stroke Registry Endovascular Treatment (2026). Reperfusion‐Dependent Outcomes After Endovascular Thrombectomy Stratified by NIHSS ‐ ASPECTS Clinical‐Core Mismatch. Annals of Clinical and Translational Neurology, 13(9), 1807-1816. https://doi.org/10.1002/acn3.70358

Image Credits: AI Generated

DOI: 10.1002/acn3.70358

Keywords: mechanical thrombectomy, acute ischemic stroke, reperfusion, clinical-core mismatch, NIHSS, ASPECTS, mTICI, functional outcome, German Stroke Registry, large vessel occlusion, penumbra, neurointervention