Oxford Centre for Clinical Magnetic Resonance Research, Oxford Biomedical Research Centre National Institute for Health Research, Division of Cardiovascular (Q.Z., M.J.B., E.L., M.Shanmuganathan, I.A.P., C.N., R.M., K.W., E.H., A.B., S.D.P., H.C.W., S.N., V.M.F., S.K.P.).
Radcliffe Department of Medicine (Q.Z., M.J.B., E.L., M. Shanmuganathan, I.A.P., C.N., R.M., K.W., E.H., H.C.W., S.N., V.M.F., S.K.P.), University of Oxford, UK.
Circulation. 2021 Aug 24;144(8):589-599. doi: 10.1161/CIRCULATIONAHA.121.054432. Epub 2021 Jul 7.
Late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) imaging is the gold standard for noninvasive myocardial tissue characterization but requires intravenous contrast agent administration. It is highly desired to develop a contrast agent-free technology to replace LGE for faster and cheaper CMR scans.
A CMR virtual native enhancement (VNE) imaging technology was developed using artificial intelligence. The deep learning model for generating VNE uses multiple streams of convolutional neural networks to exploit and enhance the existing signals in native T1 maps (pixel-wise maps of tissue T1 relaxation times) and cine imaging of cardiac structure and function, presenting them as LGE-equivalent images. The VNE generator was trained using generative adversarial networks. This technology was first developed on CMR datasets from the multicenter Hypertrophic Cardiomyopathy Registry, using hypertrophic cardiomyopathy as an exemplar. The datasets were randomized into 2 independent groups for deep learning training and testing. The test data of VNE and LGE were scored and contoured by experienced human operators to assess image quality, visuospatial agreement, and myocardial lesion burden quantification. Image quality was compared using a nonparametric Wilcoxon test. Intra- and interobserver agreement was analyzed using intraclass correlation coefficients (ICC). Lesion quantification by VNE and LGE were compared using linear regression and ICC.
A total of 1348 hypertrophic cardiomyopathy patients provided 4093 triplets of matched T1 maps, cines, and LGE datasets. After randomization and data quality control, 2695 datasets were used for VNE method development and 345 were used for independent testing. VNE had significantly better image quality than LGE, as assessed by 4 operators (n=345 datasets; <0.001 [Wilcoxon test]). VNE revealed lesions characteristic of hypertrophic cardiomyopathy in high visuospatial agreement with LGE. In 121 patients (n=326 datasets), VNE correlated with LGE in detecting and quantifying both hyperintensity myocardial lesions (=0.77-0.79; ICC=0.77-0.87; <0.001) and intermediate-intensity lesions (=0.70-0.76; ICC=0.82-0.85; <0.001). The native CMR images (cine plus T1 map) required for VNE can be acquired within 15 minutes and producing a VNE image takes less than 1 second.
VNE is a new CMR technology that resembles conventional LGE but without the need for contrast administration. VNE achieved high agreement with LGE in the distribution and quantification of lesions, with significantly better image quality.
晚期钆增强（LGE）心血管磁共振（CMR）成像技术是无创性心肌组织特征描述的金标准，但需要静脉内造影剂给药。人们非常希望开发一种无造影剂的技术来替代 LGE，以实现更快和更便宜的 CMR 扫描。
使用人工智能开发了一种 CMR 虚拟原生增强（VNE）成像技术。用于生成 VNE 的深度学习模型使用多个卷积神经网络流来利用和增强原生 T1 图谱（组织 T1 弛豫时间的像素图谱）和心脏结构和功能的电影成像中的现有信号，将其呈现为 LGE 等效图像。VNE 生成器使用生成对抗网络进行训练。该技术首先在多中心肥厚型心肌病注册中心的 CMR 数据集上开发，以肥厚型心肌病为例。数据集被随机分为 2 个独立的组进行深度学习训练和测试。VNE 和 LGE 的测试数据由有经验的操作人员进行评分和描绘，以评估图像质量、视觉空间一致性和心肌病变负担定量。使用非参数 Wilcoxon 检验比较图像质量。使用组内相关系数（ICC）分析内部和观察者之间的一致性。使用线性回归和 ICC 比较 VNE 和 LGE 的病变定量。
共有 1348 例肥厚型心肌病患者提供了 4093 对匹配的 T1 图谱、电影和 LGE 数据集。随机化和数据质量控制后，使用 2695 个数据集进行 VNE 方法开发，使用 345 个数据集进行独立测试。VNE 在 4 名操作人员（n=345 个数据集；<0.001[Wilcoxon 检验]）评估的图像质量明显优于 LGE。VNE 以高视觉空间一致性显示出与 LGE 一致的肥厚型心肌病特征性病变。在 121 例患者（n=326 个数据集）中，VNE 与 LGE 在检测和定量高信号心肌病变（=0.77-0.79；ICC=0.77-0.87；<0.001）和中等强度病变（=0.70-0.76；ICC=0.82-0.85；<0.001）方面具有相关性。VNE 所需的原生 CMR 图像（电影加 T1 图谱）可在 15 分钟内采集，生成 VNE 图像不到 1 秒。
VNE 是一种新的 CMR 技术，类似于传统的 LGE，但无需造影剂给药。VNE 在病变的分布和定量方面与 LGE 具有高度一致性，并且图像质量显著提高。
