Computational Data Science and Engineering, North Carolina A&T State University, Greensboro, North Carolina, USA.
Joint Department of Biomedical Engineering, North Carolina State University and the University of North Carolina at Chapel Hill, Raleigh, North Carolina, USA.
Ultrasound Med Biol. 2023 Mar;49(3):761-772. doi: 10.1016/j.ultrasmedbio.2022.10.015. Epub 2022 Nov 30.
Early stages of diabetic kidney disease (DKD) are difficult to diagnose in patients with type 2 diabetes. This work was aimed at identifying contrast-enhanced ultrasound (CEUS) perfusion parameters, a microcirculatory biomarker indicative of early DKD progression. CEUS kidney flash-replenishment data were acquired in control, insulin resistant and diabetic vervet monkeys (N = 16). By use of a mono-exponential model, time-intensity curve parameters related to blood volume (A), velocity (β) and flow rate (perfusion index [PI]) were extracted from 10 concentric kidney layers to study spatial perfusion patterns that could serve as strong indicators of disease. Mean squared error (MSE) was used to assess model performance. Features calculated from the perfusion parameters were inputs for the linear regression models to determine which features could distinguish between cohorts. The mono-exponential model performed well, with average MSEs (±standard deviation) of 0.0254 (±0.0210), 0.0321 (±0.0242) and 0.0287 (±0.0130) for the control, insulin resistant and diabetic cohorts, respectively. Perfusion index features, with blood pressure, were the best classifiers between cohorts (p < 0.05). CEUS has the potential to detect early microvascular changes, providing insight into disease-related structural changes in the kidney. The sensitivity of this technique should be explored further by assessing various stages of DKD.
糖尿病肾病（DKD）早期在 2 型糖尿病患者中难以诊断。本研究旨在确定对比增强超声（CEUS）灌注参数，这是一种反映早期 DKD 进展的微循环生物标志物。在对照、胰岛素抵抗和糖尿病恒河猴（N=16）中获取了 CEUS 肾脏闪烁补充数据。通过使用单指数模型，从 10 个同心肾层中提取与血容量（A）、速度（β）和流量（灌注指数[PI]）相关的时间-强度曲线参数，以研究可作为疾病强指标的空间灌注模式。均方误差（MSE）用于评估模型性能。从灌注参数计算的特征是线性回归模型的输入，以确定哪些特征可以区分队列。单指数模型表现良好，平均 MSE（±标准偏差）分别为 0.0254（±0.0210）、0.0321（±0.0242）和 0.0287（±0.0130），用于对照、胰岛素抵抗和糖尿病队列。灌注指数特征与血压相结合，是队列之间的最佳分类器（p < 0.05）。CEUS 具有检测早期微血管变化的潜力，可深入了解肾脏与疾病相关的结构变化。通过评估 DKD 的各个阶段，应进一步探索该技术的敏感性。
