Clinical multidimensional prediction model for futile reperfusion in acute ischemic stroke after endovascular thrombectomy

 What followup research did you do ensure reperfusion will work completely every time? Oh NO, YOU INCOMPETENTLY DID NOTHING, right? Predicting failure is totally fucking useless! I don't think you identified cause and effect properly, measuring Rankin scores has nothing directly to do with reperfusion! My god, the blithering stupidity out there is astounding!

But it probably is because you did NOTHING to stop the 5 causes of the neuronal cascade of death in the first week and thus letting die hundreds of millions to billions of neurons!

You really don't  know what the fuck you are doing in stroke, so get the hell out!

Clinical multidimensional prediction model for futile reperfusion in acute ischemic stroke after endovascular thrombectomy

• 
  

• • 

    Sisi Jiang 1

  • W

    Weinv Fan 1

  • 

    Yunqin Wu 1

  • 

    Xiaoxia Liu 1

  • 

    Da Li 1,2

  • O

    Ou Zhang 1,2

  • X

    Xiaofeng Xie 1

  • F

    Feiyu Chen 2,3

  • Y

    Yindan Yao 1,2*

• 1. Department of Neurology, Ningbo No.2 Hospital, Wenzhou Medical University, Ningbo, China

• 2. Department of Cerebrovascular Diseases, Ningbo No.2 Hospital, Wenzhou Medical University, Ningbo, Zhejiang, China

Abstract

Background: 

Previous Studies on prediction models for futile reperfusion after endovascular thrombectomy (EVT) in acute ischemic stroke (AIS) related to large vessel occlusion (LVO) have yielded inconsistent results. This inconsistency may be largely attributed to methodological limitations, particularly in variable selection and missing data handling. Consequently, the prognostic value of several key clinical predictors remains to be fully elucidated.

Methods: 

This retrospective study included 390 patients with AIS who underwent EVT at Ningbo No.2 Hospital. All of them achieved successful reperfusion with modified Thrombolysis in Cerebral Infarction (mTICI) score ≥ 2b. Futile reperfusion was defined as a modified Rankin Scale score of 3–6 at 90-day. Missing data were handled with multiple imputation. Logistic regression models were built using a two step predictor selection process: first univariable screening with p < 0.2; then further selection based on event count constraints. Only variables that were selected in all five imputed datasets, meaning a 100% selection frequency, were retained. Model performance measures were pooled following Rubin’s rules.

Results: 

Based on preoperative assessments integrating clinical, imaging, and laboratory markers, the final model comprised nine variables: National Institutes of Health Stroke Scale (NIHSS) score, Computed Tomography angiography-source images Alberta Stroke Program Early Computed Tomography Score (CTA-SI ASPECTS), time from onset to reperfusion (OTR), collateral circulation scores (CCS), C-reactive protein (CRP), glucose, white blood cell (WBC) count, neutrophil count, and monocyte count. The final model demonstrated good discriminative ability, with a pooled test AUC of 0.795 and a Brier score of 0.178. At the optimal threshold (mean 0.457), the model achieved a specificity of 0.822 and accuracy of 0.761, with positive net benefit across clinically relevant threshold probabilities on decision curve analysis. A nomogram incorporating the nine consistently selected predictors was developed to facilitate individualized risk prediction.

Conclusion: 

We developed a multidimensional model integrating clinical, imaging, and laboratory markers to predict futile reperfusion following EVT in patients with anterior circulation stroke. Each marker provides independent prognostic information; collectively, they represent the multidimensional risk architecture underlying postprocedural outcomes.

More at link.