AUTHOR=You Wei , Sun Yong , Feng Junqiang , Wang Zhiliang , Li Lin , Chen Xiheng , Lv Jian , Tang Yudi , Deng Dingwei , Wei Dachao , Gui Siming , Liu Xinke , Liu Peng , Jin Hengwei , Ge Huijian , Zhang Yanling 

TITLE=Protocol and Preliminary Results of the Establishment of Intracranial Aneurysm Database for Artificial Intelligence Application Based on CTA Images

JOURNAL=Frontiers in Neurology

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2022.932933

DOI=10.3389/fneur.2022.932933

ISSN=1664-2295

ABSTRACT=Background and purpose
Unruptured intracranial aneurysm (UIAs) are increasingly being detected in clinical practice. Artificial intelligence has been increasingly used to assist diagnostic techniques and show encouraging prospects. In present study, we report the protocol and preliminary results of the establishment of intracranial aneurysm database for artificial intelligence application based on CTA images.
Methods
Through a review of picture archiving and communication system, we collected CTA images of aneurysm patients between January 2010 and Mar 2021. The radiologists performed manual segmentation of all diagnosed aneurysms on subtraction CTA as the basis for automatic aneurysm segmentation. Then, artificial intelligence will apply to two stages of aneurysm treatment: automatic aneurysm detection and segmentation model based on CTA image and aneurysm risk prediction model.  
Results
Three medical centers have been included in this study so far. A total of 3190 cases of CTA examinations with 4124 aneurysms were included in the database. All identified aneurysms from CTA images that enrolled in present study were manually segmented on subtraction CTA by 6 readers. We developed a structure of 3D-Unet for aneurysm detection and segmentation in CTA images. The algorithm was developed and tested using a total of 2272 head CTAs with 2938 intracranial aneurysm. The recall and FP/case of this model for detecting aneurysms were 0.964 and 2.01 and the Dice values for aneurysm segmentation were 0.783.
Conclusions
This paper introduces the protocol and preliminary results of the establishment of intracranial aneurysm database for artificial intelligence application based on CTA images. The establishment of a multi-center database based on CTA images of intracranial aneurysms is the basis for the application of AI in the diagnosis and treatment of aneurysms. In addition to segmentation, AI should have a great potential for aneurysm treatment and management in the future.