Advantages of 3D registration technology (3DRT) in clinical application of unruptured intracranial aneurysm follow-up: A novel method to judge aneurysm growth - 03/09/22

Doi : 10.1016/j.neurad.2022.08.004

Jiewen Geng ^a,^b, Yadong Wang ^a,^b,^c, Zhe Ji ^a,^b, Wenzhi Wang ^d, Yin Yin ^d, Guangming Yang ^d, Xinxin Fan ^a,^b,^e, Tianyang Li ^f, Peng Hu ^a,^b, Chuan He ^a,^b, Hongqi Zhang ^a,^b,^⁎
^a Department of Neurosurgery, Xuanwu Hospital, Capital Medical University, Beijing, China
^b China International Neuroscience Institute, Beijing, China
^c Department of Neurosurgery, Weihai Municipal Hospital, Weihai, Shandong, China
^d Department of R&D, UnionStrong (Beijing) Technology Co., Ltd., Beijing, China
^e Department of Neurosurgery, Xi'an No.3 Hospital, the Affiliated Hospital of Northwest University, Xi'an, Shanxi, China
^f Capital Medical University, Beijing, China

^⁎Corresponding author at: Department of Neurosurgery in Xuanwu Hospital, Capital Medical University, Beijing, China.Department of Neurosurgery in Xuanwu HospitalCapital Medical UniversityBeijingChina

Sous presse. Épreuves corrigées par l'auteur. Disponible en ligne depuis le Saturday 03 September 2022
Cet article a été publié dans un numéro de la revue, cliquez ici pour y accéder

Abstract

Background and purpose

Currently available methods for determining aneurysm growth are not accurate enough. Therefore, we introduced a more intuitive and accurate 3D registration technology (3DRT) to judge the growth of aneurysms.

Materials and methods

We developed an in-house technique for 3DRT and calculated its derivative parameters, voxel change rate (VCR), maximum growth vector (MGV), and parent artery coincidence (PAC). To verify the accuracy, growing aneurysms and stable aneurysms matching 1:3 were selected, and a 3DRT measurement was performed. We calculated the sensitivity, specificity, and accuracy of cases with VCR > 20%, MGV > 1 mm, and combined indicator of VCR > 20% + MGV >1 mm. In addition, we analyzed the cause of the poor registration effect, where the registration effect of PAC > 0.7 was considered acceptable. We also collected 24 consecutive aneurysms for agreement analysis of 2D manual measurement and 3DRT.

Results

Twenty-seven growing aneurysms and 81 stable aneurysms were included in the normal model group, and 88 aneurysms with good registration effect in the adjusted model group. For aneurysms with VCR > 20%, the sensitivity and the specificity were the highest at 81.48% and 91.35%, respectively, while in the adjusted model group, the sensitivity and the specificity increased to 94.44% and 94.29%, respectively. When using VCR > 20% as the growth metric, the AUC value in the normal and the adjusted model group was 0.856 and 0.947, respectively. The ICC between 2D manual measurements and the 3DRT was 0.95 (95%CI: 0.88-0.98), and the time spent between the two groups had a significant difference (10.96 min vs. 3.44 min, p<0.01, 95% CI, 6.49-8.53).