Development of a novel nomogram to predict hemorrhagic transformation following endovascular treatment in patients with acute ischemic stroke

Abstract

Background: Hemorrhagic transformation (HT) is a critical complication of endovascular therapy (EVT) in acute ischemic stroke (AIS), significantly worsening patient outcomes. Although various risk factors have been identified, existing predictive models often fail to account for the multimodal nature of EVT and the complex interplay of clinical, imaging, and laboratory variables.

Objective: This study aimed to develop and validate a nomogram-based predictive model to estimate the risk of HT in AIS patients undergoing EVT, incorporating clinical, imaging, and laboratory data to provide a comprehensive risk assessment.

Methods: A retrospective analysis was performed on 154 AIS patients who underwent EVT at a single center between 2018 and 2023. The least absolute shrinkage and selection and operator (LASSO) and multivariate logistic regression were used to identify the independent predictors of HT. A nomogram was constructed and evaluated using the area under the receiver operating characteristic curve (AUC-ROC), calibration curves, and decision curve analysis (DCA).

Results: Among the 154 patients, 34.4% experienced HT. The nomogram demonstrated excellent discriminatory ability, with an AUC-ROC of 0.82 (95% CI: 0.752-0.888), and strong calibration, as indicated by calibration curves. DCA confirmed the model's clinical utility when the threshold probability was <0.8. Six independent prediction factors of HT were identified: atrial fibrillation (OR: 6.152), albumin (OR: 1.145), baseline NIHSS score (OR: 1.081), diastolic blood pressure (OR: 1.057), Trial of ORG 10172 in Acute Stroke Treatment (TOAST) Classification (TOAST_2, cardioembolic stroke subtype, OR: 0.201), and the location of obstructed blood vessel_5 (basilar artery occlusion, OR: 0.081).

Conclusion: The developed nomogram provides an accurate, individualized risk assessment of HT in AIS patients undergoing EVT. This tool enables personalized risk stratification, aiding clinicians in optimizing treatment strategies and improving patient outcomes. Further multicenter validation is warranted to generalize these findings.

Keywords: acute ischemic stroke; endovascular therapy; hemorrhagic transformation; nomogram; prediction factors.

Figure 1

Selection of predictive variables using LASSO regression. (A) Characteristics of variable coefficient changes. This path diagram illustrates the evolution of the regression coefficients as the regularization parameter (λ) increases. (B) Determining the optimal value for parameter λ in the LASSO regression model using the cross-validation method.

Figure 2

Nomogram for predicting HT in AIS patients undergoing EVT. The predictors include Obstructed blood vessel_5 (basilar artery occlusion), TOAST_2 (cardioembolic stroke), atrial fibrillation, albumin (ALB) levels, NIHSS score, and diastolic blood pressure (DBP). HT, hemorrhagic transformation; AIS, acute ischemic stroke; EVT, endovascular therapy.

Figure 3

ROC curve and AUC of the nomogram for predicting HT in AIS patients undergoing EVT. ROC, receiver operating characteristic; AUC, area under the ROC curve; HT, hemorrhagic transformation; AIS, acute ischemic stroke; EVT, endovascular therapy.

Figure 4

Calibration curve for evaluating the agreement between the nomogram predicted probability and the actual probability.

Figure 5

DCA of the nomogram for evaluating the clinical applicability of the model. DCA, decision curve analysis.