Latent class analysis of 12 routine post-procedural clinical and laboratory indicators identified three biologically distinct phenotypes (Low-burden, Metabolic, Hypercoagulable) among EVT-treated patients with divergent 3-month functional outcomes, reproduced in a temporal validation cohort, though multicenter prospective validation is needed before clinical use.
Latent class analysis of 825 EVT-treated patients identified three phenotypes based on 12 routine post-procedural indicators.
The Low-burden phenotype had the lowest inflammatory and coagulatory burden and the best functional outcomes.
The Metabolic phenotype, characterized by hyperglycemia and elevated neutrophil-to-lymphocyte ratio, had a markedly lower favorable outcome rate.
The Hypercoagulable phenotype, defined by markedly elevated D-dimer and atrial fibrillation, had the lowest favorable outcome rate and the highest sICH rate.
Both the Metabolic and Hypercoagulable phenotypes were independently associated with unfavorable 3-month outcome after multivariable adjustment.
The class distribution and outcome associations were reproduced in an independent temporal validation cohort.
Approximately half of EVT patients fail to achieve functional independence at 3 months despite successful recanalization, and the authors conclude multicenter prospective validation is needed before clinical application.
Even when doctors successfully clear a blocked brain artery with endovascular thrombectomy, about half of stroke patients do not regain functional independence by three months. This study asked whether routine blood tests and clinical measures taken after the procedure could sort patients into biologically distinct groups with different outlooks. Using 12 routine post-procedure measures from 825 patients, a statistical grouping method found three patterns. A 'low-burden' group (about 44%) had low inflammation and clotting activity, and about 71% of them had good functional outcomes. A 'metabolic' group (about 43%) had high blood sugar and a high neutrophil-to-lymphocyte ratio (an inflammation marker), and about 45% did well. A 'hypercoagulable' group (13%) had very high D-dimer (a clotting marker) and atrial fibrillation, and about 40% did well. Both of the latter groups had significantly lower odds of good outcomes even after accounting for other factors. The same patterns appeared in a separate group of 211 later patients. This research suggests that post-procedure biology may help explain why outcomes differ after successful thrombectomy, and might eventually help identify patients who could benefit from different management. However, the study was done in a limited setting, and the authors emphasize that larger multicenter prospective studies are needed before these groupings can be used in clinical practice.
Yang L, Chen C, Sun L, Li M, Han L. (2026). Latent class analysis identifies three biologically distinct phenotypes among patients undergoing endovascular thrombectomy: a derivation and temporal validation study.. Frontiers in neurology. https://doi.org/10.3389/fneur.2026.1855033