A Novel Hybrid Data Mining Framework for Early-Stage Aneurysm Detection and Risk Pattern Recognition

Abstract

Intracranial aneurysms pose a serious global health challenge due to the risk of rupture leading to subarachnoid hemorrhage, which causes significant morbidity and mortality rates[1]. Early identification of aneurysms before symptom onset is crucial for timely clinical intervention and better patient outcomes[2]. Advances in non-invasive imaging techniques such as Time-Of-Flight Magnetic Resonance Angiography (TOF-MRA) and Computed Tomography Angiography (CTA) have facilitated widespread screening, although the large and complex imaging datasets pose diagnostic challenges to clinicians[3][4]. Recent progress in deep learning (DL) and hybrid data mining frameworks—leveraging supervised, weakly supervised, and self-supervised learning—has spurred the development of automated aneurysm detection systems. These systems further integrate anatomical priors, attention mechanisms, transformer architectures, and heuristic post-processing. This paper reviews the latest hybrid frameworks, evaluates their diagnostic efficacy, and discusses persisting challenges and emerging opportunities toward robust, interpretable, and scalable aneurysm detection solutions .