Quantitative mapping of developing fetal organs using dynamic MRI and artificial neural networksare Technologies

使用动态 MRI 和人工神经网络对发育中的胎儿器官进行定量绘图技术

• 批准号: 2271380
• 金额: --
• 依托单位: King's College London
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2019
• 资助国家: 英国
• 起止时间: 2019 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2271380
• 关键词: Quantitative; mapping; developing; fetal; organs

Quantitative T1 and T2 mapping of the fetus has potential to allow characterisation of subtle changes in tissue properties. Normal developmental processes such as myelination, changing microstructure and water content are known to have profound effect on T1 and T2 relaxation times1, which reflect fundamental biological properties of the tissues. Identifying subtle changes to these values resulting from pathology, inflammation or injury would require fully quantitative measurements.However, the fetal motion presents a serious challenge to established quantitative methods. Many state-of-the-art neuroimaging qMRI methods require multiple lengthy 3D acquisitions and steady-state sequences2 that are easily disrupted by motion, and are thus incompatible with fetal imaging. Recent advancements have produced high quality 3D non-quantitative anatomical images of moving fetuses by using a "slice to volume registration" (SVR) approach, which assembles a 3D image from multiple 2D images using image registration3 (see Fig. 1 and 2).The aim of the project is to develop innovative motion resistant slice-selective sequences that can be reconstructed into high resolution consistent quantitative maps. Artificial intelligence methods (deep neural networks) will be employed to provide a tractable solution to a complex inverse problem of estimation of relaxation times from the non-steady state sequences by modelling interaction with fetal motion.The newly developed motion resistant quantitative techniques will be applied to fetal brain and other organs to assess their diagnostic value and facilitate comparison of fetal brain development with pre-term neonatal population.

胎儿的定量 T1 和 T2 绘图有可能描述组织特性的细微变化。众所周知，髓鞘形成、微观结构变化和含水量等正常发育过程对 T1 和 T2 弛豫时间 1 具有深远影响，这反映了组织的基本生物学特性。识别由病理、炎症或损伤引起的这些值的细微变化需要完全定量测量。然而，胎动对现有的定量方法提出了严峻的挑战。许多最先进的神经成像 qMRI 方法需要多次冗长的 3D 采集和稳态序列2，这些序列很容易被运动破坏，因此与胎儿成像不兼容。最近的进展通过使用“切片到体积配准”(SVR) 方法产生了移动胎儿的高质量 3D 非定量解剖图像，该方法使用图像配准从多个 2D 图像组装 3D 图像（见图 1 和 2）。该项目的目的是开发创新的抗运动切片选择序列，可以将其重建为高分辨率一致的定量图。将采用人工智能方法（深度神经网络），通过对与胎儿运动的相互作用进行建模，为从非稳态序列估计松弛时间的复杂反问题提供易于处理的解决方案。新开发的抗运动定量技术将应用于胎儿大脑和其他器官，以评估其诊断价值，并促进胎儿大脑发育与早产新生儿群体的比较。