AUTHOR=An Xingwei , He Jiaqian , Di Yang , Wang Miao , Luo Bin , Huang Ying , Ming Dong 

TITLE=Intracranial Aneurysm Rupture Risk Estimation With Multidimensional Feature Fusion

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 16 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2022.813056

DOI=10.3389/fnins.2022.813056

ISSN=1662-453X

ABSTRACT=The rupture of aneurysms is the main cause of spontaneous subarachnoid hemorrhage, which is a serious life-threatening disease with high mortality and permanent disability rates. It is therefore highly desirable to evaluate the rupture risk of aneurysms. In this paper, we propose a novel semi-automatic prediction model for the estimation of intracranial aneurysm rupture risk. It consists of multi-dimensional feature fusion, feature selection, and the construction of classification models. In the multi-dimensional feature fusion, we extract four kinds of features and combine them into the feature set, including morphological features, radiomics features, clinical features, and deep learning features. Specifically, we extract and analyze the classification capability of three types of different deep learning features with the feature extractor 3D EfficientNet-B0. In the experiment, we construct five distinct classification models, among which the logistic regression classifier shows the best performance for aneurysm rupture risk estimation, reaching an F2-score of 0.758. Our results suggest that the full use of multi-dimensional feature fusion can enhance the performance of aneurysm risk assessment. Compared with other methods, our method achieved the state-of-the-art performance for aneurysm risk assessment methods based on CADA 2020.