Improved Quantification of Myocardium Scar in Late Gadolinium Enhancement Images: Deep Learning Based Image Fusion Approach.
Improved Quantification of Myocardium Scar in Late Gadolinium Enhancement Images: Deep Learning Based Image Fusion Approach.
复制标题
DOI：
10.1002/jmri.27555
复制
发表时间：
2021-07
期刊：
影响因子：
--
通讯作者：
Nezafat R
中科院分区：
文献类型：
--
作者：
Fahmy AS;Rowin EJ;Chan RH;Manning WJ;Maron MS;Nezafat R
由于瘢痕与背景对比度低和图像质量低，在晚期钆增强（LGE）心脏磁共振成像中量化心肌瘢痕可能具有挑战性。为了解决模糊的LGE区域，有经验的阅片师通常使用传统的电影序列来准确识别心肌边界。开发一种深度学习模型，用于结合LGE和电影图像，以提高LGE疤痕量化的鲁棒性和准确性。回顾展共191例肥厚型心肌病患者：1）来自两个研究中心的162例患者随机分为训练组（50%; 81例患者）、验证组（25%，40例患者）和测试组（25%; 41例患者）; 2）来自第三个研究中心的外部测试数据集（29例患者）。1.5T，反转-恢复分段梯度回波LGE和平衡稳态自由进动电影序列训练两个卷积神经网络（CNN）进行心肌和瘢痕分割，一个进行LGE-Cine融合，一个不进行LGE-Cine融合。对于具有融合的CNN，输入是在匹配的心脏相位和解剖位置处的两个对齐的LGE和电影图像。对于没有融合的CNN，仅使用LGE图像作为输入。使用数据集的手动分割作为参考标准。使用Pearson线性相关系数（r）和Bland-Altman分析评估LGE瘢痕负荷和心肌体积的手动和基于CNN的定量。两种CNN模型均显示出与LGE瘢痕负荷和心肌体积的手动定量的高度一致性。采用LGE-Cine融合的CNN比不采用LGE-Cine融合的CNN更稳健，允许成功分割显著更多的切片（603 [95%] vs. 635个切片中的562（89%）; P < 0.001）。此外，与未进行LGE-Cine融合的CNN相比，采用LGE-Cine融合的CNN与LGE瘢痕负荷的手动量化的一致性更好（%ScarLGE‐cine = 0.82 × %Scarmanual，r = 0.84 vs. %ScarLGE = 0.47 × %Scarmanual，r = 0.81）和心肌体积（体积LGE-电影= 1.03 ×体积手动，r = 0.96 vs.体积LGE = 0.91 ×体积手动，r = 0.91）。基于CNN的LGE-Cine融合可以提高自动疤痕量化的鲁棒性和准确性。3 1
Quantification of myocardium scarring in late gadolinium enhanced (LGE) cardiac magnetic resonance imaging can be challenging due to low scar‐to‐background contrast and low image quality. To resolve ambiguous LGE regions, experienced readers often use conventional cine sequences to accurately identify the myocardium borders. To develop a deep learning model for combining LGE and cine images to improve the robustness and accuracy of LGE scar quantification. Retrospective. A total of 191 hypertrophic cardiomyopathy patients: 1) 162 patients from two sites randomly split into training (50%; 81 patients), validation (25%, 40 patients), and testing (25%; 41 patients); and 2) an external testing dataset (29 patients) from a third site. 1.5T, inversion‐recovery segmented gradient‐echo LGE and balanced steady‐state free‐precession cine sequences Two convolutional neural networks (CNN) were trained for myocardium and scar segmentation, one with and one without LGE‐Cine fusion. For CNN with fusion, the input was two aligned LGE and cine images at matched cardiac phase and anatomical location. For CNN without fusion, only LGE images were used as input. Manual segmentation of the datasets was used as reference standard. Manual and CNN‐based quantifications of LGE scar burden and of myocardial volume were assessed using Pearson linear correlation coefficients (r) and Bland–Altman analysis. Both CNN models showed strong agreement with manual quantification of LGE scar burden and myocardium volume. CNN with LGE‐Cine fusion was more robust than CNN without LGE‐Cine fusion, allowing for successful segmentation of significantly more slices (603 [95%] vs. 562 (89%) of 635 slices; P < 0.001). Also, CNN with LGE‐Cine fusion showed better agreement with manual quantification of LGE scar burden than CNN without LGE‐Cine fusion (%ScarLGE‐cine = 0.82 × %Scarmanual, r = 0.84 vs. %ScarLGE = 0.47 × %Scarmanual, r = 0.81) and myocardium volume (VolumeLGE‐cine = 1.03 × Volumemanual, r = 0.96 vs. VolumeLGE = 0.91 × Volumemanual, r = 0.91). CNN based LGE‐Cine fusion can improve the robustness and accuracy of automated scar quantification. 3 1
登录
查看更多内容
影响因子：
14
作者：
Disertori, Marcello;Rigoni, Marta;Ravelli, Flavia
通讯作者：
Ravelli, Flavia
影响因子：
3.3
作者：
Leong, Chen Onn;Lim, Einly;Liew, Yih Miin
通讯作者：
Liew, Yih Miin
影响因子：
5.4
作者：
Wong, Timothy C.;Piehler, Kayla M.;Schelbert, Erik B.
通讯作者：
Schelbert, Erik B.
影响因子：
2.6
作者：
Turkbey, E. B.;Nacif, M. S.;Bluemke, D. A.
通讯作者：
Bluemke, D. A.
影响因子：
10.9
作者：
Karim, Rashed;Bhagirath, Pranav;Rhode, Kawal
通讯作者：
Rhode, Kawal