Cardiovascular

Risk Stratification for Early Neurological Deterioration After Mechanical Thrombectomy and Development of a 24-h Postprocedural Reassessment Model in Patients With Acute Anterior-Circulation Large-Vessel Occlusion Based on Computed Tomography Perfusion Mismatch Volume.

TL;DR

CTP mismatch volume was related to END following mechanical thrombectomy, and preprocedural and 24-h reassessment models demonstrated acceptable discrimination for predicting early neurological deterioration after internal validation.

Key Findings

Early neurological deterioration (END) occurred in 19.9% of patients treated with mechanical thrombectomy for acute anterior-circulation large-vessel occlusion.

  • END occurred in 87 of 438 patients (19.9%)
  • 78 patients experienced neurological worsening and 9 died within 72 h
  • END was defined as a ≥4-point increase from preprocedural NIHSS score or death from any cause during the first 72 h after thrombectomy
  • Study period was January 2021 to December 2024 with consecutive patients included retrospectively

The association between mismatch volume and END was nonlinear, with paradoxically higher END rates observed at lower mismatch volumes.

  • P for nonlinearity = 0.023 based on restricted cubic spline (RCS) analysis
  • Observed END frequencies were 33.0%, 18.9%, and 12.5% for mismatch volumes <75 mL, 75-105 mL, and >105 mL, respectively
  • P for trend <0.001 across mismatch volume categories
  • The counterintuitive pattern of higher END at lower mismatch volumes suggests that small mismatch may reflect limited salvageable tissue or severe baseline injury

A preprocedural model comprising mismatch volume, ischemic core volume, admission NIHSS score, and collateral score achieved an AUC of 0.789 for predicting END.

  • AUC of 0.789 (95% CI, 0.735–0.843)
  • Bootstrap-corrected AUC was 0.774 after 1000 bootstrap resamples for internal validation
  • Only information available before thrombectomy was entered into the preprocedural model
  • END occurred in 8.0%, 20.7%, and 38.5% of patients assigned to low-, intermediate-, and high-risk categories of the simplified preprocedural score

A 24-h landmark reassessment model for delayed END (occurring >24–72 h) achieved an AUC of 0.838 among patients who remained END-free through 24 h.

  • Among 387 patients eligible for 24-h reassessment, 36 developed delayed END
  • AUC of 0.838 (95% CI, 0.773–0.903)
  • The 24-h model added successful recanalization status and hemorrhagic transformation subtypes to the baseline predictors
  • Model discrimination and calibration were evaluated by ROC analysis, calibration assessment, and decision curve analysis (DCA)

Delayed END was associated with parenchymal hematoma but not with hemorrhagic infarction as a hemorrhagic transformation subtype.

  • This distinction was identified in the 24-h reassessment cohort of 387 patients
  • Hemorrhagic transformation subtypes were incorporated as predictors only in the 24-h landmark model
  • Parenchymal hematoma, representing more severe hemorrhagic transformation, was specifically linked to delayed neurological deterioration after 24 h

The study constructed two temporally anchored risk models corresponding to distinct clinical decision points: preprocedural and 24-h post-thrombectomy.

  • The preprocedural model used only information available before thrombectomy
  • The 24-h landmark model was restricted to patients free of END at 24 h, forming a reassessment cohort
  • Internal validity was assessed using 1000 bootstrap resamples for both models
  • The authors note that independent external validation remains necessary

What This Means

This research studied a serious complication called early neurological deterioration (END) — meaning a significant worsening of stroke symptoms or death within 72 hours — in patients who underwent a clot-removal procedure called mechanical thrombectomy for stroke. The study looked at 438 patients treated over four years and found that about 1 in 5 patients (19.9%) experienced END. Researchers focused on a brain imaging measurement called CT perfusion (CTP) mismatch volume, which estimates how much brain tissue is at risk of dying but could potentially be saved. Surprisingly, patients with smaller mismatch volumes (less than 75 mL) had the highest rates of END (33%), while those with larger mismatch volumes had lower rates, suggesting the relationship between this imaging measure and outcomes is more complex than previously assumed. The researchers built two predictive tools to help identify which patients are at highest risk. The first tool, used before the procedure, combined the mismatch volume with three other factors — the size of already-damaged brain tissue (ischemic core), stroke severity on admission, and the quality of blood flow through backup vessels (collateral score). This model performed reasonably well at identifying high-risk patients, with about 38.5% of those classified as high-risk actually experiencing END compared to only 8% in the low-risk group. A second tool was designed for use 24 hours after the procedure for patients who had not yet deteriorated, adding information about whether the blood vessel was successfully reopened and whether any bleeding into the brain had occurred. Notably, one specific type of brain bleeding (parenchymal hematoma) was linked to later deterioration, while a milder type (hemorrhagic infarction) was not. This research suggests that CT perfusion imaging data, combined with other clinical and imaging factors, can help doctors identify stroke patients most likely to deteriorate after clot removal — both before the procedure and again at the 24-hour mark. Having two separate risk assessment windows may allow for more tailored monitoring and potentially earlier intervention. However, the authors caution that these models were tested only on the same patients used to build them (internal validation), and testing in separate patient groups at other hospitals will be needed before these tools can be widely adopted in clinical practice.

Have a question about this study?

Citation

Du W, He A, Jiang L. (2026). Risk Stratification for Early Neurological Deterioration After Mechanical Thrombectomy and Development of a 24-h Postprocedural Reassessment Model in Patients With Acute Anterior-Circulation Large-Vessel Occlusion Based on Computed Tomography Perfusion Mismatch Volume.. Revista de neurologia. https://doi.org/10.31083/RN53745