Quality assessment and clinical value of rapid segmentation of computed tomography angiography images of aortic root based on deep learning
HUANG Xiaobin
WU Zhengcan
HUANG Ying
CHEN Qian
CHAI Hao
YE Peng
XIE Guanghui
XU Hui
MA Chenming
YANG Jian
ZHANG Haibo
CHEN Guozhong
Abstract：Objective To explore the reliability and clinical value of fast and accurate automatic segmentation and evaluation of aortic root based on deep learning compared with computed tomography angiography(CTA)image analysis.Methods CTA images of 68 patients undergoing transcatheter aortic valve replacement(TAVR)in Nanjing First Hospital from August 2021 to October 2022 were collected and analyzed using 3mensio and Cvpilot platform,respectively.The analysis results were analyzed retrospectively and subjectively by experts.Results Of the 68 patients subjectively evaluated by experts,63(92.6％)reached Grade Ⅲ,4(5.9％)reached Grade Ⅱ,and 1(1.5％)reached Grade Ⅰ.There were no statistically significant differences between the measurement results of Cvpilot platform and 3mensio(P＞0.05).Compared with junior doctors,the measurement results of the Cvpilot platform were closer to the evaluation results of experts.In addition,the Cvpilot platform significantly reduced the total measured time and mouse travel distance(P＜0.01).Conclusion For patients undergoing TAVR,the Cvpilot platform has a good correlation with conventional computed tomography measurements(r=0.883).The reliable evaluation quality of the platform makes it have broad clinical application prospects in the future.
Keywords：aortic valvetranscatheter aortic valve replacementautomatic segmentationdeep learningcomputed tomography
Publication Date：2025-02-27
Online Publishing Date：2026-08-26(First online date of this platform, not the publication date of the document)
Pages：5( 187-191 )