Abstract
Abdominal aortic aneurysms (AAA) may remain clinically silent until they enlarge and patients present with a potentially lethal rupture. This necessitates early detection and elective treatment. The goal of this study was to develop an easy-to-train algorithm which is capable of automated AAA screening in CT scans and can be applied to an intra-hospital environment. Three deep convolutional neural networks (ResNet, VGG-16 and AlexNet) were adapted for 3D classification and applied to a dataset consisting of 187 heterogenous CT scans. The 3D ResNet outperformed both other networks. Across the five folds of the first training dataset it achieved an accuracy of 0.856 and an area under the curve (AUC) of 0.926. Subsequently, the algorithms performance was verified on a second data set containing 106 scans, where it ran fully automated and resulted in an accuracy of 0.953 and an AUC of 0.971. A layer-wise relevance propagation (LRP) made the decision process interpretable and showed that the network correctly focused on the aortic lumen. In conclusion, the deep learning-based screening proved to be robust and showed high performance even on a heterogeneous multi-center data set. Integration into hospital workflow and its effect on aneurysm management would be an exciting topic of future research.
Highlights
Abdominal aortic aneurysm (AAA) is a potentially life-threatening condition [1,2,3].Possible rupture is associated with high mortality exceeding 50% [4,5,6]
The aim of this study is to develop and validate an trainable and fully automated deep learning 3D AAA screening algorithm, which can run as a background process in the clinic workflow
layer-wise relevance propagation (LRP) is applied to the best network from Experiment 1 to validate that the network decision is based on the correct region of interest (ROI)
Summary
Abdominal aortic aneurysm (AAA) is a potentially life-threatening condition [1,2,3].Possible rupture is associated with high mortality exceeding 50% [4,5,6]. The focus on other clinical questions and the time-consuming nature of a detailed AAA analysis might lead to underreporting and delayed diagnosis [7]. Patients might become discharged without detection of an early AAA. This may potentially lead to a delay in treatment since surveillance programs have been shown to be of benefit [8]. We modified the network to make it suitable for 3D classification by reducing the number of filters and employing only two fully connected layers. In contrast to the AlexNet it uses multiple 3 × 3 convolutions replacing large kernel-sized filters. The first convolutional layer uses a stride of two in every spatial direction
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
