Development of an Optimal Flow Diverter Stent Prediction Model Based on Parent Artery Morphology Analysis

Area of Science:

• Neurosurgery
• Biomedical Engineering
• Artificial Intelligence in Medicine

Background:

• Intracranial aneurysms (IA) pose significant rupture risks, necessitating effective treatment.
• Flow Diverter Stents (FDS) are a common treatment, but device selection is complex.
• Current FDS selection relies heavily on surgeon experience and intricate vascular imaging.

Purpose of the Study:

• To develop and validate a machine learning (ML) decision support system for predicting optimal Flow Diverter Stent (FDS) size and length.
• To integrate clinical data and detailed vascular measurements for improved FDS selection.
• To enhance treatment planning and patient outcomes for intracranial aneurysms (IA).

Main Methods:

• A machine learning model was developed using data from 94 internal carotid artery aneurysm cases.
• 61 features, including clinical parameters and vessel diameter measurements, were analyzed.
• Six regression algorithms were evaluated using Bayesian hyperparameter optimization and cross-validation.

Main Results:

• The ML system achieved 94.7% accuracy in predicting FDS size.
• Initial FDS length prediction accuracy was 78.9%, improving to 89% after feature selection.
• The system demonstrated strong potential for supporting clinical decision-making in FDS selection.

Conclusions:

• Machine learning can effectively predict appropriate Flow Diverter Stent (FDS) size and length for intracranial aneurysms (IA).
• This AI-driven approach can assist clinicians in optimizing FDS selection, potentially leading to better treatment outcomes.
• Further development and validation could integrate this system into routine clinical practice for aneurysm treatment.