What question did this study set out to answer?

The study aimed to develop a nomogram that integrates multimodal CT imaging features to predict hemorrhagic transformation after endovascular thrombectomy for acute ischemic stroke.

May 27, 2026Open Access

Nomogram prediction model for hemorrhagic transformation after endovascular thrombectomy in acute ischemic stroke based on multimodal CT imaging features

Key Points

The study aimed to develop a nomogram that integrates multimodal CT imaging features to predict hemorrhagic transformation after endovascular thrombectomy for acute ischemic stroke.
Retrospective single-center study including patients with acute anterior circulation ischemic stroke undergoing mechanical thrombectomy from December 2020 to December 2024.
Patients classified into hemorrhagic transformation and non-hemorrhagic groups based on follow-up imaging at 24–72 hours post-procedure.
Multivariable logistic regression used to identify factors predicting hemorrhagic transformation.
Of 437 patients, 132 (30.2%) developed hemorrhagic transformation.
Key predictors included atrial fibrillation (OR 1.790), baseline NIHSS score (OR 1.093), blood glucose (OR 1.129), and ASPECTS score (OR 0.733).
The nomogram showed an AUC of 0.866, outperforming the clinical-only model's AUC of 0.781 (P < 0.001).

Abstract

Background The occurrence of hemorrhagic transformation (HT) represents a significant risk in patients suffering from acute ischemic stroke (AIS) who are undergoing endovascular thrombectomy (EVT). Multimodal CT imaging features may improve risk prediction. This study aimed to construct a nomogram incorporating multimodal CT parameters to predict HT. Methods This retrospective single-center study included consecutive patients suffering from acute anterior circulation ischemic stroke who underwent mechanical thrombectomy from December 2020 to December 2024. Based on follow-up imaging at 24–72 h post-procedure, patients were classified into HT group (hemorrhagic infarction type 2 HI2 or parenchymal hematoma PH1/PH2) and non-HT group (no hemorrhage or hemorrhagic infarction type 1 HI1). Demographic characteristics, vascular risk factors, admission clinical status, laboratory tests, treatment-related factors, and multimodal CT imaging features were collected. Imaging included non-contrast CT (Alberta Stroke Program Early CT Score ASPECTS, hyperdense middle cerebral artery sign), CT angiography (collateral score, occlusion site), and CT perfusion (core infarct volume, hypoperfusion volume, mismatch volume, relative cerebral blood flow rCBF, relative cerebral blood volume rCBV, permeability surface PS). Results Of 437 patients, 132 (30.2%) developed HT. Multivariable logistic regression identified atrial fibrillation (OR 1.790), baseline National Institutes of Health Stroke Scale (NIHSS) score (OR 1.093), blood glucose (OR 1.129), ASPECTS score (OR 0.733), collateral score (OR 0.486), core infarct volume (OR 1.044), and PS (OR 3.040) as factors that independently contribute to the prediction. Nomogram showed good discriminatory power (AUC = 0.866), which significantly outperformed a clinical-only model (AUC = 0.781, P < 0.001), along with excellent calibration and positive net benefit on DCA. Conclusion A nomogram integrating multimodal CT imaging features with clinical variables provides a promising tool for risk estimation of HT after EVT, which may facilitate early risk stratification and individualized management.

Read Full Paperexternally

Mark Helpful

Bookmark

Relay

View Full Paper