Implementation and prospective clinical validation of AI-based planning and shimming techniques in cardiac MRI.
Implementation and prospective clinical validation of AI-based planning and shimming techniques in cardiac MRI.
复制标题
心脏MRI中基于AI的计划和光滑技术的实施和前瞻性临床验证。
作者：
心血管磁共振（CMR）是心血管疾病治疗的重要诊断工具。先进 CMR 技术与人工智能 (AI) 相结合的出现，有望简化成像、在不影响图像质量 (IQ) 的情况下缩短图像采集时间，并提高磁场均匀性。在这里，我们的目标是实施两种基于人工智能的深度学习技术，用于自动切片对齐和心脏匀场，并评估它们在临床心脏磁共振成像（MRI）中的性能。在预先获取的心脏 MRI 数据集（> 500 名受试者）上开发、训练和验证了两个深度神经网络，以实现 CMR 的自动切片规划和匀场（在扫描仪中实现）。为了检查我们的自动心脏计划 (EasyScan) 和基于人工智能的垫片 (AI shim) 的性能，随后进行了两项前瞻性研究。为了进行 EasyScan 验证，10 名健康受试者接受了两个相同的 CMR 协议：手动心脏计划和基于 AI 的 EasyScan，以在 1.5 T 临床 MRI 扫描仪上评估协议扫描时间差异和心脏平面处方的准确性。为了进行 AI 垫片验证，总共招募了 20 名受试者：10 名健康患者和 10 名转诊进行 CMR 检查的心脏肿瘤患者。使用标准心脏容量垫片和 AI 垫片获得电影图像，以评估信噪比 (SNR)、对比度噪声比 (CNR)、整体 IQ（清晰度和 MR 图像退化）、射血分数 (EF) 和绝对壁增厚。采用非参数（Wilcoxon）和参数（t 检验）评估的混合统计方法进行统计分析。与手动心脏计划的协议相比，基于人工智能的平面处方 EasyScan 的 CMR 协议最大限度地减少了操作员依赖性，并将总体扫描时间缩短了 2 分钟以上（快 ∼13%，p < 0.001）。 EasyScan 平面处方还证明了比之前报道的策略更准确的心脏平面（由经过认证的心脏 MRI 技术人员手动处方的平面的平面角度误差更少）。此外，AI 垫片还改善了 B0 场均匀性。使用 AI 垫片获得的电影图像显示，左心室 (LV) 心肌的 SNR (12.49%；p = 0.002) 明显高于使用体积垫片获得的图像（体积垫片：32.90 ± 7.42 与 AI 垫片：37.01 ± 8.87）。使用 AI 垫片进行电影成像 (149.02 ± 39.15) 时，左心室心肌 CNR 比体积垫片 (132.49 ± 33.94) 高 12.48%。使用 AI shim 获得的图像比使用体积 shim 获得的图像更清晰 (p = 0.012)。 LVEF 和绝对壁增厚也表明两种匀场方法之间存在差异。在两组中，AI 垫片的 LVEF 略大于体积垫片的 LVEF：健康组的 AI 垫片高 2.87%，患者组的 AI 垫片高 1.70%。左心室壁绝对增厚（单位：毫米）还显示，各组匀场方法之间存在差异，患者组观察到的变化较大（健康组：3.31%，p = 0.234，患者组：7.29%，p = 0.059）。与采用手动心脏计划的 CMR 协议相比，使用 EasyScan 进行心脏计划的 CMR 检查显示加速了心脏检查。与体积匀场相比，使用 AI 匀场技术还实现了改进且更均匀的 B0 磁场均匀性。
Cardiovascular magnetic resonance (CMR) is a vital diagnostic tool in the management of cardiovascular diseases. The advent of advanced CMR technologies combined with artificial intelligence (AI) has the potential to simplify imaging, reduce image acquisition time without compromising image quality (IQ), and improve magnetic field uniformity. Here, we aim to implement two AI‐based deep learning techniques for automatic slice alignment and cardiac shimming and evaluate their performance in clinical cardiac magnetic resonance imaging (MRI). Two deep neural networks were developed, trained, and validated on pre‐acquired cardiac MRI datasets (>500 subjects) to achieve automatic slice planning and shimming (implemented in the scanner) for CMR. To examine the performance of our automated cardiac planning (EasyScan) and AI‐based shim (AI shim), two prospective studies were performed subsequently. For the EasyScan validation, 10 healthy subjects underwent two identical CMR protocols: with manual cardiac planning and with AI‐based EasyScan to assess protocol scan time difference and accuracy of cardiac plane prescriptions on a 1.5 T clinical MRI scanner. For the AI shim validation, a total of 20 subjects were recruited: 10 healthy and 10 cardio‐oncology patients with referrals for a CMR examination. Cine images were obtained with standard cardiac volume shim and with AI shim to assess signal‐to‐noise ratio (SNR), contrast‐to‐noise ratio (CNR), overall IQ (sharpness and MR image degradation), ejection fraction (EF), and absolute wall thickening. A hybrid statistical method using of nonparametric (Wilcoxon) and parametric (t‐test) assessments was employed for statistical analyses. CMR protocol with AI‐based plane prescriptions, EasyScan, minimized operator dependence and reduced overall scanning time by over 2 min (∼13 % faster, p < 0.001) compared to the protocol with manual cardiac planning. EasyScan plane prescriptions also demonstrated more accurate (less plane angulation errors from planes manually prescribed by a certified cardiac MRI technologist) cardiac planes than previously reported strategies. Additionally, AI shim resulted in improved B0 field homogeneity. Cine images obtained with AI shim revealed a significantly higher SNR (12.49%; p = 0.002) than those obtained with volume shim (volume shim: 32.90 ± 7.42 vs. AI shim: 37.01 ± 8.87) for the left ventricle (LV) myocardium. LV myocardium CNR was 12.48% higher for cine imaging with AI shim (149.02 ± 39.15) than volume shim (132.49 ± 33.94). Images obtained with AI shim resulted in sharper images than those obtained with volume shim (p = 0.012). The LVEF and absolute wall thickening also showed that differences exist between the two shimming methods. The LVEF by AI shim was shown to be slightly larger than LVEF by volume shim in two groups: 2.87% higher with AI shim for the healthy group and 1.70% higher with AI shim for the patient group. The LV absolute wall thickening (in mm) also showed that differences exist between shimming methods for each group with larger changes observed in the patient group (healthy: 3.31%, p = 0.234 and patient group: 7.29%, p = 0.059). CMR exams using EasyScan for cardiac planning demonstrated accelerated cardiac exam compared to the CMR protocol with manual cardiac planning. Improved and more uniform B0 magnetic field homogeneity also achieved using AI shim technique compared to volume shimming.
登录
查看更多内容
DOI：
10.1186/s12968-018-0484-5
发表时间：
2018-09-20
期刊：
Journal of cardiovascular magnetic resonance : official journal of the Society for Cardiovascular Magnetic Resonance
影响因子：
--
作者：
Puntmann VO;Valbuena S;Hinojar R;Petersen SE;Greenwood JP;Kramer CM;Kwong RY;McCann GP;Berry C;Nagel E;SCMR Clinical Trial Writing Group
通讯作者：
SCMR Clinical Trial Writing Group
影响因子：
3.9
作者：
Osadebey ME;Pedersen M;Arnold DL;Wendel-Mitoraj KE
通讯作者：
Wendel-Mitoraj KE
影响因子：
4.4
作者：
Thiele, H;Nagel, E;Fleck, E
通讯作者：
Fleck, E
影响因子：
4.4
作者：
Dietrich, Olaf;Raya, Jose G.;Schoenberg, Stefan O.
通讯作者：
Schoenberg, Stefan O.
影响因子：
24
作者：
Heckbert, Susan R.;Post, Wendy;Bluemke, David A.
通讯作者：
Bluemke, David A.