Automated Diffusion-Weighted MRI Analysis for Exploratory Risk Stratification of Malignant Cerebral Edema After Acute...

Area of Science:

• Neurology
• Radiology
• Medical Imaging

Background:

• Malignant cerebral edema (MCE) is a rare but severe complication of acute ischemic stroke (AIS).
• Early identification of patients at high risk for MCE remains a significant clinical challenge.
• Current diagnostic methods lack sufficient accuracy for timely MCE risk assessment.

Purpose of the Study:

• To evaluate SCR-U1.8, an automated imaging biomarker, for exploratory risk stratification of MCE in AIS patients.
• To assess the performance of SCR-U1.8 in integrating stroke lesion burden and cerebrospinal fluid reserve.
• To compare SCR-U1.8-based prediction with simpler threshold-based methods.

Main Methods:

• Retrospective analysis of AIS patients who underwent diffusion-weighted imaging (DWI).
• Automated segmentation of stroke lesions (U1.8 lesion volume) and cerebrospinal fluid volume (CSFV).
• Calculation of SCR-U1.8 (U1.8/CSFV) and evaluation of logistic regression models.

Main Results:

• SCR-U1.8 logistic regression (SCR-U1.8-LR) demonstrated high performance across 100 train-test splits, achieving superior accuracy, precision, F1 score, and ROC-AUC.
• SCR-U1.8-LR showed significantly improved accuracy, precision, and F1 score compared to a U1.8 volume threshold.
• While sensitivity and NPV slightly decreased, SCR-U1.8-LR offered a more robust risk stratification.

Conclusions:

• SCR-U1.8 offers a simple, physiologically interpretable biomarker for assessing stroke lesion burden relative to CSF reserve.
• The biomarker shows potential for exploratory risk stratification of MCE after AIS.
• Further external validation is required due to the limited number of MCE events observed in this study.