Analyzing the Efficacy of Computer-Aided Detection in Cerebral Aneurysm Diagnosis Using MRI Modality: A Review

Computer-aided detection (CAD) models play a critical role in the clinical diagnosis of cerebral aneurysms, significantly contributing to the reduction of mortality rates associated with this condition. This article provides a comprehensive overview of the evolution of CAD models for aneurysm detection, with a particular focus on MRI modalities. It explores the motivations behind CAD systems, the methodologies employed, and their respective advantages and limitations, offering valuable insights into the current state-of-the-art (SOTA) CAD systems. The research papers selected for this review focus on research utilizing TOF MRA as the imaging modality and emphasize computer-aided detection through both traditional and deep learning techniques, with a particular emphasis on Convolutional Neural Networks (CNNs). CNNs have proven to be a crucial component in improving the accuracy and efficiency of aneurysm detection by automatically learning features from raw imaging data, bypassing the need for manual feature extraction. The article also presents a detailed experimental analysis of deep learning models, benchmarked using TOF MRA datasets. Key research gaps are identified, including the need for large training samples, challenges in Maximum Intensity Projection (MIP) imaging, limitations of 2D architectures, and issues related to overfitting and computational complexity. The review also observes that shallow networks and pretrained models are effective in addressing these challenges. In addition to identifying these gaps, the review outlines future directions for the development of CAD systems, aiming to further advance CAD models for aneurysm detection.