[home] [Personal Program] [Help]

---

Session: Vascular III
Session starts: Friday 27 January, 14:00
Presentation starts: 14:00
Room: Room 559

---

Yizhou Huang (Eindhoven University of Technology)
Emilia Badescu (Philips Research France)
Ruud van den Sloun (Eindhoven University of Technology)
Massimo Mischi (Eindhoven University of Technology)

Abstract:
The advancement of ultrafast Doppler imaging enables more accurate flow measurements compared to conventional Doppler. Most studies have focused on ultrafast coronary imaging combined with singular value decomposition (SVD) for clutter filtering with favorable acquisition settings (open-chest, linear high-frequency probe, and plane waves). Here, we aim at verifying the feasibility of using phased-array diverging wave imaging to measure coronary flow at increased depth, for non-negligible cardiac tissue motion, mimicking a condition similar to clinical transthoracic echocardiography. This imposes an extra challenge on SVD for separating blood signals from clutter. To obtain the optimal Doppler performance, a set of acquisition and processing parameters, including the number of transmitting angles, the ensemble length, and different compounding strategies, along with different SVD thresholding methods, are examined in this study. To mimic coronary flow measurements, a 15% gelatin phantom with a 3-mm diameter channel (20 degrees, 9-12 cm depth) is connected to a syringe pump generating flow at 10 cm/s. A Philips S5-1 phased array probe operating at 3.125-MHz is mounted on a sinusoidal axial motion generator with a speed of ±6 cm/s. A Verasonics 256 system is programmed to generate alternate 1-, 2-, 4-, 8, and 14-angle (between ±7°) acquisitions with a pulse repetition frequency of 4400 Hz. The performance of different compounding strategies, including coherent compounding (CC), CC with a sliding window of 2 angles (CC-S), and linear interpolation of CC frames, is compared. Six different ensemble lengths from 0.025 to 1.2 s are analyzed. SVD with three literature thresholding methods (Baranger et al., TMI, 2018; Maresca et al., TUFFC, 2018) are investigated. A novel method to determine the clutter-blood threshold is proposed that relies on the minimum distance of the singular value curve from the origin, and an extensive threshold search method is also implemented as a reference. Contrast-to-Noise ratio (CNR) is employed for quantitative comparison. Using an alternate 2-angle acquisition, CC-S, an ensemble length of 0.1 s, with the proposed SVD thresholding method, returns the highest CNR. These promising results prove the feasibility of measuring coronary flow with ultrafast diverging wave imaging.