Automatic Wide-Field Optical Coherence Tomography for Assessment of Transplant Kidney Viability

摘要 世界范围内供移植的肾脏短缺，主要原因是没有可靠的方法。 以确定可供移植的肾脏的存活能力。有未得到满足的临床需求来减少&gt；8,000 每年因未能找到可供移植的可行肾脏而死亡的人数，以减少近4- 每年在捐赠者等待名单上的时间，并减少失败的移植数量。目前的筛查程序 已故供者肾脏使用两种方法：1)根据活检的解剖特征进行病理评分 (肾小管、肾小球等)2)根据捐赠者的病史得出的肾脏捐赠者概况指数(KDPI) (高血压、糖尿病、体重等)。然而，临床研究表明，这些目前的方法有 有限的鉴别力。光学相干层析成像(OCT)是一种可以获得 高分辨率、非侵入性、原位和实时的生物组织断层图像。我们有 OCT可以提供肾脏的非侵入性、实时、组织病理学信息，即 使用任何其他已知的程序都无法获得。我们已经证明了人体的OCT成像 移植前后的肾脏组织病理学检查可以用来预测移植后的肾脏 功能。此外，这些初步试验表明，人类供体肾脏的OCT成像 在手术室使用手持设备是安全的，整个肾脏可以在相对较小的范围内进行评估。 一小段时间。从最近完成的对169个人类移植肾脏的临床研究中，我们发现在 扩大标准供体(ECD)肾(或边缘肾)，增大管腔直径即可 植入前预测移植物功能延迟(DGF)。在这个方案中，我们将开发一种新型的OCT设备 通过智能扫描和深度学习在移植前评估供体肾脏的存活能力。通过扫描 对于整个肾脏表面，我们的设备旨在消除活组织检查/KDPI范例造成的不确定性。这个 OCT成像研究将与移植后的肾功能相关联，以建立成像算法 以及肾脏移植前OCT成像指南。我们的中心假设是 OCT测量的综合形态参数可用于确定移植后 肾功能。该方案的具体目标是：1)开发机器人辅助的自动三维扫描OCT 用于人体肾脏移植前成像的成像设备。2)发展基于深度学习的形象 定量评估OCT参数作为功能状态指示器的处理算法 肾脏，以组织学评分系统为金标准。3)得出用于评估的诊断标准 并进行前瞻性临床研究，以评估预测术后肾功能的准确性。 使用OCT实现移植功能。

Summary There is a worldwide shortage of kidneys for transplantation due mainly to the fact that there is no reliable means to determine the viability of kidneys available for transplant. There are unmet clinical needs to reduce the >8,000 deaths that occur each year from failures of finding viable kidney matches for transplant, to reduce the nearly 4- year time on donor waitlists, and to reduce the number of failed transplants. The current process for screening deceased donor kidneys uses two methods: 1) pathological scores based on anatomical features from a biopsy (tubules, glomeruli, etc.) and 2) the Kidney Donor Profile Index (KDPI) derived from the donor’s medical history (hypertension, diabetes, weight, etc.). However, clinical research indicates that those current methods have limited discriminative power. Optical coherence tomography (OCT) is an imaging technology that can obtain high-resolution, non-invasive, cross-sectional images of biological tissues in situ and in real time. We have demonstrated that OCT can provide non-invasive, real-time, histopathological information of the kidney that is impossible to obtain using any other known procedure. We have demonstrated that OCT imaging of human kidney histopathology both prior to and following their transplant can be used to predict post-transplant renal function. Furthermore, these preliminary trials have demonstrated that OCT imaging of human donor kidneys with a hand-held unit in the operating room is safe and that the entire kidney can be evaluated within a relatively short period of time. From a recently finished clinical study with 169 human transplant kidneys, we found that in the expanded criteria donor (ECD) kidneys (or marginal kidneys), increased tubular lumen diameter was able to predict delayed graft function (DGF) prior to implantation. In this proposal, we will develop a novel OCT device with intelligent scanning and deep learning to evaluate donor kidney viability before transplant. By scanning the whole kidney surface, our device intends to eliminate the uncertainty created by the biopsy/KDPI paradigm. The OCT imaging studies will be correlated with post-transplant renal function in order to establish imaging algorithms and guidelines for OCT imaging of kidneys prior to their transplant. Our central hypothesis is that more comprehensive morphological parameters as measured by OCT can be used to determine post-transplantation renal function. The specific aims of this proposal are: 1) Develop robot-assisted automatic 3D scanning OCT imaging device to image human kidneys prior to their transplantation. 2) Develop deep-learning-based image processing algorithms to quantitatively assess the OCT parameters as indicators of the functional status of kidneys, using histology scoring system as the gold standard. 3) Derive the diagnostic criteria for assessing transplant kidney function and perform prospective clinical studies to assess the accuracy of predicting post- transplant function using OCT.