Related Experiment Video
Updated: Jun 23, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
42.6K
A Systematic Review of Deep-Learning Methods for Intracranial Aneurysm Detection in CT Angiography.
1Laboratory of Imaging Technologies, Faculty of Electrical Engineering, University of Ljubljana, 1000 Ljubljana, Slovenia.
Biomedicines
|November 25, 2023
Summary
Artificial intelligence algorithms show high accuracy in detecting large cerebral aneurysms on CT angiography. However, detecting smaller aneurysms remains a challenge, requiring further research and standardized metrics for improved diagnostic accuracy.
Area of Science:
- Neurology
- Radiology
- Medical Imaging
Background:
- Subarachnoid hemorrhage from cerebral aneurysm rupture is a major cause of death and disability.
- Early aneurysm detection via Computed Tomography Angiography (CTA) is critical for patient outcomes.
- Artificial intelligence (AI) can enhance aneurysm detection rates and reduce variability.
Purpose of the Study:
- To systematically review and meta-analyze the diagnostic accuracy of deep-learning AI algorithms for detecting cerebral aneurysms on CTA.
- To assess the performance of AI in identifying aneurysms of varying sizes.
Main Methods:
- Systematic search of PubMed, Embase, and Cochrane Library (Jan 2015-July 2023).
- Inclusion of studies using automated/semi-automatic deep-learning for aneurysm detection on CTA.
- Assessment using PRISMA and QUADAS-2 guidelines; meta-analysis of sensitivity, specificity, and false positives.
- Utilized enhanced FROC curves for study comparison.
Main Results:
- Fifteen studies were included, showing high pooled lesion-level sensitivity (0.87) for intracranial aneurysms.
- Sensitivity for small aneurysms (<3 mm) was notably low (0.56).
- Limited data on patient-level sensitivity and specificity due to definitional inconsistencies and lack of control groups.
Conclusions:
- Deep-learning AI demonstrates high accuracy for detecting cerebral aneurysms >3 mm on CTA.
- Significant need for research focused on improving detection of smaller aneurysms (<3 mm).
- Call for standardized test datasets and consistent performance metrics for AI in aneurysm detection.
