To develop and internally evaluate a probability-level stacked ensemble model integrating DWI and ADC MRI-based predictions with clinical predictions to predict 90-day poor functional outcomes in acute ischemic stroke patients.
Key Findings:
The fusion model achieved the highest AUC of 0.951 in the held-out internal test set, with sensitivity and specificity of 0.889 and 0.911, respectively.
The fusion model significantly outperformed the imaging model (p < 0.05) but not the clinical model or Wouters 2018 model.
Interpretation:
The fusion model demonstrated high internal discrimination and improved performance compared to imaging alone.
Limitations:
The study was conducted at a single center, limiting generalizability.
External validation, recalibration, and prospective evaluation are required before broader clinical use.
Conclusion:
The fusion model shows promise for early risk stratification in acute ischemic stroke but requires further validation.
