What can we learn from machine learning studies on flow diverter aneurysm embolization? A systematic review

Area of Science:

• Neuroendovascular surgery
• Medical artificial intelligence
• Cerebrovascular disease management

Background:

• Flow diverter use for aneurysms is increasing, complicating outcome prediction.
• Accurate prediction of flow diverter success is crucial for patient management.

Purpose of the Study:

• To review machine learning algorithms for predicting flow diverter success in aneurysm occlusion.
• To guide neurointerventionalists in utilizing AI for treatment planning.

Main Methods:

• Systematic review following PRISMA guidelines.
• Screening of PubMed, Embase, Scopus, and Web of Science databases.
• Inclusion of original research on machine learning for flow diverter outcome prediction.

Main Results:

• Five studies evaluated various machine learning algorithms.
• Algorithms analyzed patient, aneurysm, and device-related features.
• Accuracy ranged from 61% to 90% with algorithms like Gaussian SVM achieving the highest rates.

Conclusions:

• Machine learning models demonstrate high accuracy (up to 90%) in predicting flow diverter success.
• AI integration can enhance neurointerventional decision-making for aneurysm treatment.