Predicting Cerebral Aneurysms Rupture From Hemodynamics.

Area of Science:

• Biomedical Engineering
• Cardiovascular Research
• Medical Imaging Analysis

Background:

• Cerebral aneurysms are common vascular abnormalities with potentially fatal consequences, often leading to subarachnoid hemorrhage upon rupture.
• Accurately predicting which cerebral aneurysms are at high risk of rupture remains a significant clinical challenge.
• Current risk assessment methods may not fully capture the complex biomechanical factors contributing to aneurysm rupture.

Purpose of the Study:

• To investigate the efficacy of combining computational fluid dynamics (CFD) with statistical testing and classification models for assessing cerebral aneurysm rupture risk.
• To identify key hemodynamic, geometric, and patient-specific features associated with aneurysm rupture.
• To enhance the accuracy of rupture risk prediction by integrating diverse data sources.

Main Methods:

• Performed CFD simulations for 103 cerebral aneurysms to extract hemodynamic metrics.
• Combined hemodynamic data with geometric, anatomical, and patient-specific information.
• Utilized PERMANOVA and classification models to differentiate between ruptured and unruptured aneurysms.

Main Results:

• Univariate analyses showed limited predictive power (best accuracy 65%).
• Integrating non-CFD data improved prediction accuracy to 67%.
• Incorporating CFD-derived hemodynamic features boosted rupture prediction accuracy to nearly 79%.
• Specific hemodynamic variables showed relevance during distinct cardiac cycle phases, highlighting transient flow effects.

Conclusions:

• Combining hemodynamic analysis with statistical techniques provides a robust method for assessing cerebral aneurysm rupture risk.
• The study offers valuable insights into the biomechanical mechanisms underlying aneurysm rupture.
• The developed approach shows potential for clinical translation to support individualized patient risk stratification.