Transvalvular pressure drops are assessed using Doppler echocardiography for the diagnosis of heart valve disease. However, this method is highly user-dependent and may overestimate transvalvular pressure drops by up to 54%. This work aimed to assess transvalvular pressure drops using velocity fields derived from blood speckle imaging (BSI), as a potential alternative to Doppler. A silicone 3D-printed aortic valve model, segmented from a healthy CT scan, was placed within a silicone tube. A CardioFlow 5000MR flow pump was used to circulate blood mimicking fluid to create eight different stenotic conditions. Eight PendoTech pressure sensors were embedded along the tube wall to record ground-truth pressures (10 kHz). The simplified Bernoulli equation with measured probe angle correction was used to estimate pressure drop from maximum velocity values acquired across the valve using Doppler and BSI with a GE Vivid E95 ultrasound machine and 6S-D cardiac phased array transducer. There were no significant differences between pressure drops estimated by Doppler, BSI and ground-truth at the lowest stenotic condition (10.4 ± 1.76, 10.3 ± 1.63 vs. 10.5 ± 1.00 mmHg, respectively; p > 0.05). Significant differences were observed between the pressure drops estimated by the three methods at the greatest stenotic condition (26.4 ± 1.52, 14.5 ± 2.14 vs. 20.9 ± 1.92 mmHg for Doppler, BSI and ground-truth, respectively; p < 0.05). Across all conditions, Doppler overestimated pressure drop (Bias = 3.92 mmHg), while BSI underestimated pressure drop (Bias = -3.31 mmHg). BSI accurately estimated pressure drops only up to 10.5 mmHg in controlled phantom conditions of low stenotic burden. Doppler overestimated pressure drops of 20.9 mmHg. Although BSI offers a number of theoretical advantages to conventional Doppler echocardiography, further refinements and clinical studies are required with BSI before it can be used to improve transvalvular pressure drop estimation in the clinical evaluation of aortic stenosis.

中文翻译：

---

血斑成像与传统多普勒超声在主动脉血流体模中跨瓣压降估计的比较

使用多普勒超声心动图评估跨瓣压降以诊断心脏瓣膜疾病。然而，这种方法高度依赖于用户，可能会高估高达 54% 的跨瓣压降。这项工作旨在使用源自血斑成像 (BSI) 的速度场评估跨瓣压降，作为多普勒的潜在替代方案。从健康的 CT 扫描中分割出来的硅胶 3D 打印主动脉瓣模型被放置在硅胶管内。CardioFlow 5000MR 流量泵用于循环血液模拟流体以产生八种不同的狭窄状况。八个 PendoTech 压力传感器沿管壁嵌入，以记录真实压力 (10 kHz)。使用带有 GE Vivid E95 超声机和 6S-D 心脏相控阵换能器的多普勒和 BSI，使用带有测量探头角度校正的简化伯努利方程来估计通过瓣膜获得的最大速度值的压降。在最低狭窄条件下，多普勒、BSI 和地面实况估计的压降之间没有显着差异（分别为 10.4 ± 1.76、10.3 ± 1.63 和 10.5 ± 1.00 mmHg；p > 0.05）。在最大狭窄条件下，三种方法估计的压降之间存在显着差异（多普勒、BSI 和地面实况分别为 26.4 ± 1.52、14.5 ± 2.14 和 20.9 ± 1.92 mmHg；p < 0.05）。在所有条件下，多普勒高估了压降（偏差 = 3.92 mmHg），而 BSI 低估了压降（偏差 = -3.31 mmHg）。BSI 准确估计在低狭窄负担的受控体模条件下压力下降仅高达 10.5 mmHg。多普勒高估了 20.9 mmHg 的压降。尽管 BSI 与传统的多普勒超声心动图相比具有许多理论优势，但在将 BSI 用于改善主动脉瓣狭窄临床评估中的跨瓣压降估计之前，还需要对 BSI 进行进一步的改进和临床研究。