Advances in ultrasound imaging for tissue strain and perfusion quantification
Yizhou Huang defended his PhD thesis at the Department of Electrical Engineering on February 3rd.
Ultrasound imaging has become an indispensable tool in modern medicine, offering a cost-effective, time-efficient, non-invasive, and safe alternative to other diagnostic modalities. Over the past eight decades, continuous advancements in ultrasound technology have revolutionized the quantification of tissue strain and perfusion, leading to the establishment of clinical standards in cardiology and oncology. Notably, strain imaging is now integral for myocardial deformation assessment, while perfusion imaging plays a key role in tumor diagnosis. Inspired by these successes, Yizhou Huang expanded the application of ultrasound imaging to novel clinical challenges, specifically in fertilization, cardiology, and neurosurgery.
Infertility, affecting approximately 17.5% of the global adult population, presents a significant medical and psychological burden. While various biological, lifestyle, and age-related factors contribute to infertility, uterine contractility plays a crucial role in reproductive success. Although transvaginal ultrasound enables visualization of uterine peristaltic contractions, clinical adoption has been hindered by a lack of quantitative metrics.
To address this gap, developed a dedicated ultrasound-based tool utilizing speckle tracking and strain imaging to quantify uterine contractions. This tool measures contraction frequency, amplitude, direction, propagation velocity, and coordination, providing critical insights into uterine dynamics. Additionally, a machine-learning framework leveraging these contraction features was designed to predict in-vitro fertilization (IVF) success prior to embryo transfer. With an area under the receiver operating characteristic curve exceeding 0.8, this predictive model enables clinicians to tailor treatment plans, ultimately improving pregnancy outcomes. This advancement is particularly valuable given that IVF is often the last resort for women with fertility issues, costing between $5,000 and $10,000 per cycle, with a success rate of only around 30%. The developed tool also holds promise for diagnosing uterine dysfunctions such as adenomyosis. Its clinical relevance is further highlighted by its application in hospital settings, where it has been used in studies that have earned high academic recognition.
Advanced coronary blood flow imaging in cardiology
In cardiology the assessment of myocardial contraction using ultrasound strain imaging is a standard diagnostic tool for coronary artery disease (CAD), a leading cause of mortality worldwide. However, conventional strain imaging only provides indirect indicators of perfusion deficits. Recent innovations in ultrafast Doppler imaging, capable of imaging microvessels at over 10 kHz, offer direct visualization of perfusion within organs such as the kidneys and prostate. Despite this potential, transthoracic cardiac imaging using ultrafast Doppler presents challenges due to strong tissue motion artifacts.
To overcome these limitations, this PhD research introduces advanced clutter-filtering techniques, extending singular value decomposition (SVD) filtering into a higher-dimensional space. The novel higher-order SVD (HOSVD) clutter filter incorporates an additional angular dimension, significantly enhancing contrast-to-noise ratio (CNR) and enabling clear visualization of coronary blood flow in a Langendorff-perfused swine heart model. Further improvements, including blood subspace estimation and noise rejection strategies, yielded an average CNR increase of 10.8 dB for cardiac muscle suppression and 5 dB for noise reduction. These advancements lay a strong foundation for real-time coronary flow screening in clinical practice. The importance of this research is underscored by a formal collaboration with GE Healthcare, which aims to implement these techniques into clinical ultrasound scanners for broader medical use.
Real-time cerebrovascular perfusion monitoring in neurosurgery
Beyond cardiovascular applications, ultrasound imaging holds promise for neurosurgical interventions, particularly in preventing surgery-related stroke (SrS). Current intraoperative monitoring techniques, such as indocyanine green angiography and Charbel probes, are limited in providing real-time, continuous assessments of cerebral perfusion.
This work presents the first documented use of ultrafast Doppler imaging for real-time collateral recruitment monitoring in human subjects undergoing temporary cerebral artery clipping. By adapting the HOSVD clutter filter, the microvascular perfusion of occluded arteries was visualized in ten patients, capturing perfusion dynamics during vessel occlusion, hyper-perfusion post-clip removal, and subsequent recovery. These findings align with clinical expectations and demonstrate the potential of ultrafast Doppler imaging for intraoperative stroke prevention. The significance of this work is reflected in its pioneering nature, as these represent the first intraoperative brain perfusion measurements of this kind, conducted in collaboration with Utrecht UMC.
New possibilities for future applications
These innovations, validated in ex-vivo animal models and human subjects, have the potential to improve clinical diagnostics, enhance treatment planning, and contribute to better patient outcomes. Furthermore, the methodologies developed in this work offer valuable insights into biomechanical and hemodynamic processes, opening new possibilities for future applications in other medical fields. The impact of this research of Huang has been recognized through multiple prestigious awards, including the Young Investigator Award from the European Federation of Societies for Ultrasound in Medicine and Biology (EFSUMB, 2022), further underscoring its significance in advancing medical ultrasound technology.
Title of PhD thesis: . Promotor: Prof. Massimo Mischi. Co-promotors: Dr. Ruud van Sloun and Prof. Dick Schoot (Catharina Ziekenhuis).