Clinical Prediction Model for Futile Reperfusion in Acute Ischemic Stroke

390 patient records from Ningbo No.2 Hospital were used to develop a multidimensional prediction model for futile reperfusion after endovascular thrombectomy. The model integrates nine clinical, imaging, and laboratory variables and demonstrated a pooled test AUC of 0.795. This dataset, authored by Sisi Jiang and last updated in April 2026, contains the supplementary file for the study.

Use Cases

• Validate the nine-variable prediction model for futile reperfusion based on NIHSS score, CTA-SI ASPECTS, and time from onset to reperfusion.
• Analyze the prognostic value of laboratory markers like C-reactive protein, glucose, and white blood cell count for post-EVT outcomes.
• Develop individualized risk prediction tools, such as nomograms, using the integrated clinical, imaging, and laboratory risk architecture.

Strengths

• Model development used a rigorous two-step predictor selection process with a 100% selection frequency across five imputed datasets.
• The final model demonstrated good discriminative ability with a pooled test AUC of 0.795 and a Brier score of 0.178.
• Data handling addressed missing values through multiple imputation, and model performance measures were pooled following Rubin’s rules.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data reflects a single-center study from Ningbo No.2 Hospital, which may limit generalizability.

Provenance

Source

figshare

Collection Method

Retrospective study of patients who underwent endovascular thrombectomy.

Freshness

Last updated 2026-04-30 04:16:29; freshness should be verified.

Geography

Ningbo, China (Ningbo No.2 Hospital).