• Yupeng Zhang, Chao Ma, Shikai Liang, Peng Yan, Fei Liang, Feng Guo, and Chuhan Jiang.
• Department of Interventional Neuroradiology, Beijing Neurosurgical Institute and Beijing Tiantan Hospital, Capital Medical University, Beijing, China.
• World Neurosurg. 2018 Nov 1; 119: e934-e940.

ObjectiveTo explore whether computed morphologic features can be used as independent predictors of incomplete occlusion of aneurysms treated with the Pipeline embolization device.MethodsFrom January 2016 to September 2017, 58 patients with 58 aneurysms were treated with the Pipeline embolization device. Aneurysms were manually segmented from the Digital Imaging and Communications in Medicine file, and we calculated 16 shape features voxel by voxel on the segmented aneurysm image. Along with 13 other clinical and radiographic variables, we performed univariate and multivariate analysis to explore predictors of incomplete occlusion.ResultsAt last angiographic follow-up (median 6.2 months), complete occlusion was achieved in 41 aneurysms (70.7%). In multivariate analysis, malapposition of stent (odds ratio = 0.03; 95% confidence interval, 0.00-0.32; P = 0.004) and higher elongation value (odds ratio = 0.03; 95% confidence interval, 0.01-0.17; P < 0.001) were independently associated with incomplete occlusion of aneurysms. Compared with aneurysms with complete occlusion, incompletely occluded aneurysms had higher elongation values (median 0.890 vs. 0.766; P < 0.001); the optimal cutoff value of elongation for occlusion status classification was 0.862. Predicting accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and area under the curve of the logistic regression model were 0.879, 0.902, 0.824, 0.925, 0.778, and 0.872.ConclusionsMalapposition of stent and higher elongation value were independent negative predictors of aneurysm occlusion following flow diversion.Copyright © 2018 Elsevier Inc. All rights reserved.

25 keywords...

hide…