HR-VWI Radiomics for Predicting 1-Year Adverse Events in Symptomatic ICAS: A Comparison of Plaque-Only and Integrated...

Area of Science:

• Radiology
• Cardiovascular Imaging
• Artificial Intelligence in Medicine

Background:

• Intracranial atherosclerotic stenosis (ICAS) poses a significant risk for stroke.
• Accurate prediction of adverse clinical events in ICAS patients is crucial for timely intervention.
• Current imaging and clinical scoring methods have limitations in predicting these events.

Purpose of the Study:

• To compare the predictive performance of an integrated "Plaque-Vessel" radiomics model with a "Plaque-only" model.
• To evaluate the models' ability to predict 1-year adverse clinical events in symptomatic ICAS patients.
• To assess the radiomics models against established clinical scores and qualitative imaging features.

Main Methods:

• Retrospective analysis of 272 symptomatic ICAS patients with a temporal split for training/validation and independent testing.
• Radiomic features extracted from T1-weighted and contrast-enhanced T1-weighted high-resolution vessel wall imaging (HR-VWI).
• Two radiomics models developed: "Plaque" (plaque region of interest only) and "Plaque-Vessel" (plaque, vessel wall, lumen). Performance assessed using AUC and Decision Curve Analysis (DCA).

Main Results:

• The "Plaque-Vessel" radiomics model achieved the highest predictive performance with an Area Under the Curve (AUC) of 0.800 in the test cohort.
• This model significantly outperformed qualitative plaque features (AUC = 0.631), Essen Stroke Risk Score (ESRS; AUC = 0.554), and NIH Stroke Scale (NIHSS; AUC = 0.543).
• Decision Curve Analysis (DCA) indicated greater net clinical benefit for the "Plaque-Vessel" model compared to the "Plaque-only" model and clinical scores.

Conclusions:

• The integrated "Plaque-Vessel" radiomics model demonstrates superior performance in predicting 1-year adverse events in symptomatic ICAS compared to conventional methods.
• While showing a trend towards improvement over the "Plaque-only" model, the statistical difference was inconclusive, suggesting exploratory findings.
• Further large-scale validation is necessary before routine clinical implementation for guiding treatment decisions in ICAS.