Xu Ziyi, Wang Xinghua, Qiao Nan, Zhang Tao
Department of Ultrasound, Shanxi Provincial People's Hospital Affiliated to Shanxi Medical University, Taiyuan, China.
Departments of Ultrasound, Second Hospital of Shanxi Medical University, Taiyuan, China.
Ren Fail. 2025 Dec;47(1):2525472. doi: 10.1080/0886022X.2025.2525472. Epub 2025 Jul 1.
Acute kidney injury (AKI) resulting from acute renal vein thrombosis (ARVT) is uncommon, yet it can progress swiftly, requiring prompt diagnosis and intervention. This study aimed to investigate the various multimodal ultrasound techniques, specifically conventional ultrasound (CUS), microvascular flow imaging (MFI), contrast-enhanced ultrasound (CEUS), and shear wave elastography (SWE), in conjunction with radiomics for early diagnosis and assessment of AKI resulting from ARVT using a rabbit model. Twenty healthy adult New Zealand white rabbits with 40 kidneys were included in this study. The left kidneys were designated as the experimental group ( = 20), whereas the right kidneys served as the control group( = 20). Throughout the study, multimodal ultrasound techniques were employed for image acquisition and analysis. The ultrasound images underwent processing, segmentation, feature extraction, and model construction. The dataset was randomly divided in a 7:3 ratio, and the performance of models was assessed through the Receiver Operating Characteristic Curve (ROC) analysis along with key performance metrics. In CUS images, the experimental group showed notable increases in renal volume, cortical thickness, and enhanced cortical echogenicity ( < 0.001, = 0.032, < 0.001). In the CDFI, MFI, and CEUS, the experimental group exhibited significant reductions in blood flow perfusion ( < 0.001). In SWE, Young's modulus values for the cortex, medulla, and sinus were significantly elevated in the experimental group ( < 0.001). The strongest correlations were found for creatinine, renal volume, peak systolic velocity of the arcuate artery, time from peak to half-value of sinus, and Young's modulus value for cortex minimum, with Area Under the Curve(AUC) values of 0.600, 0.868, 0.560, 0.503, and 0.982, respectively. The CUS, CDFI, MFI, CEUS, SWE, and CUS+CDFI+MFI+CEUS+SWE radiomics models demonstrated stronger performance, achieving AUC values of 0.899, 0.861, 0.899, 0.833, 0.861, and 0.734, respectively. Multimodal ultrasound combined with radiomics can significantly improve early diagnosis of AKI following ARVT, providing valuable insights for clinical research.
急性肾静脉血栓形成（ARVT）导致的急性肾损伤（AKI）并不常见，但病情进展迅速，需要及时诊断和干预。本研究旨在利用兔模型，研究多种多模态超声技术，特别是传统超声（CUS）、微血管血流成像（MFI）、超声造影（CEUS）和剪切波弹性成像（SWE），结合影像组学，对ARVT导致的AKI进行早期诊断和评估。本研究纳入了20只健康成年新西兰白兔，共40个肾脏。将左肾指定为实验组（n = 20），右肾作为对照组（n = 20）。在整个研究过程中，采用多模态超声技术进行图像采集和分析。对超声图像进行处理、分割、特征提取和模型构建。数据集按7:3的比例随机划分，并通过受试者操作特征曲线（ROC）分析以及关键性能指标评估模型的性能。在CUS图像中，实验组肾体积、皮质厚度显著增加，皮质回声增强（P < 0.001，P = 0.032，P < 0.001）。在CDFI、MFI和CEUS中，实验组血流灌注显著降低（P < 0.001）。在SWE中,实验组皮质、髓质和窦的杨氏模量值显著升高（P < 0.001）。肌酐、肾体积、弓形动脉收缩期峰值流速、窦部从峰值到半值的时间以及皮质最小值的杨氏模量值之间的相关性最强，曲线下面积（AUC）值分别为0.600、0.868、0.560、0.503和0.982。CUS、CDFI、MFI、CEUS、SWE以及CUS+CDFI+MFI+CEUS+SWE影像组学模型表现更强，AUC值分别为0.899、0.861、0.899、0.833、0.861和0.734。多模态超声结合影像组学可显著提高ARVT后AKI的早期诊断，为临床研究提供有价值的见解。
