Improving accuracy of myocardial T1 estimation in MyoMapNet.
Improving accuracy of myocardial T1 estimation in MyoMapNet.
复制标题
提高 MyoMapNet 中心肌 T1 估计的准确性。
DOI：
10.1002/mrm.29397
复制
发表时间：
2022
影响因子：
3.3
通讯作者：
Nezafat,Reza
中科院分区：
文献类型：
--
作者：
Guo,Rui;Chen,Zhensen;Amyar,Amine;El-Rewaidy,Hossam;Assana,Salah;Rodriguez,Jennifer;Pierce,Patrick;Goddu,Beth;Nezafat,Reza
目的提高MyoMapNet估计T1的准确性和鲁棒性。MyoMapNet是一种基于深度学习的方法，使用4张反转-恢复T1加权图像进行心脏T1标测。方法MyoMapNet是一种用于加速心脏T1标测序列T1估计的全连接神经网络，通过单个Look-Reversal反转-恢复实验（LL 4）收集4张T1加权图像。MyoMapNet最初是使用来自改良的Look-Maple反转恢复序列的体内数据进行训练的，这导致了对各种混杂因素的显著偏倚和敏感性。本研究试图使用数值模拟生成的信号和多个模拟混杂因素下的体模MR数据来训练MyoMapNet。然后通过使用不同于用于训练的新体模小瓶扫描的体模数据来评价训练模型。新模型的性能进行了比较，与修改后的Look-Recombination恢复序列和饱和恢复单次激发采集测量本地和对比后T1在25 subjects.ResultsIn体模study，T1值测量LL 4与MyoMapNet是高度相关的参考值从自旋回波序列。此外，估计的T1对翻转角和非共振的变化具有良好的鲁棒性。通过饱和恢复单次激发采集、改良Look-Map反转恢复序列和MyoMapNet测量的3 T下的自体和造影后心肌T1分别为1483 ± 46.6 ms和791 ± 45.8 ms、1169 ± 49.0 ms和612 ± 36.0 ms以及1443 ± 57.5 ms和700 ± 57.5 ms。相应的细胞外容积分别为22.90% ± 3.20%，28.88% ± 3.48%和30.65% ± 3.60%，respectively.ConclusionTraining MyoMapNet与数值模拟和体模数据将提高心肌T1值的估计，并增加其鲁棒性混淆，同时也减少了整体T1映射估计时间只有4心跳。
PurposeTo improve the accuracy and robustness of T1estimation by MyoMapNet, a deep learning–based approach using 4 inversion‐recovery T1‐weighted images for cardiac T1mapping.MethodsMyoMapNet is a fully connected neural network for T1estimation of an accelerated cardiac T1mapping sequence, which collects 4 T1‐weighted images by a single Look‐Locker inversion‐recovery experiment (LL4). MyoMapNet was originally trained using in vivo data from the modified Look‐Locker inversion recovery sequence, which resulted in significant bias and sensitivity to various confounders. This study sought to train MyoMapNet using signals generated from numerical simulations and phantom MR data under multiple simulated confounders. The trained model was then evaluated by phantom data scanned using new phantom vials that differed from those used for training. The performance of the new model was compared with modified Look‐Locker inversion recovery sequence and saturation‐recovery single‐shot acquisition for measuring native and postcontrast T1in 25 subjects.ResultsIn the phantom study, T1values measured by LL4 with MyoMapNet were highly correlated with reference values from the spin‐echo sequence. Furthermore, the estimated T1had excellent robustness to changes in flip angle and off‐resonance. Native and postcontrast myocardium T1at 3 Tesla measured by saturation‐recovery single‐shot acquisition, modified Look‐Locker inversion recovery sequence, and MyoMapNet were 1483 ± 46.6 ms and 791 ± 45.8 ms, 1169 ± 49.0 ms and 612 ± 36.0 ms, and 1443 ± 57.5 ms and 700 ± 57.5 ms, respectively. The corresponding extracellular volumes were 22.90% ± 3.20%, 28.88% ± 3.48%, and 30.65% ± 3.60%, respectively.ConclusionTraining MyoMapNet with numerical simulations and phantom data will improve the estimation of myocardial T1values and increase its robustness to confounders while also reducing the overall T1mapping estimation time to only 4 heartbeats.