| dc.description.abstract | Transarterial chemoembolization (TACE) treatment of liver tumors aims to reduce tumor perfusion and therefore relies on accurate perfusion imaging to monitor and evaluate treatment response. Although effective, contrast-enhanced imaging modalities suffer from treatment-induced artifacts that necessitate late follow-up imaging, typically 4-6 weeks post-treatment. The inability to assess treatment adequacy in real-time is a potential reason for lack of treatment response. Non-contrast perfusion ultrasound is a potential solution to this problem. However, perfusion ultrasound imaging without contrast is difficult and was shown to be ineffective for TACE treatment evaluation in the past. Without contrast enhancement, blood ultrasound signal is weak compared to tissue signal, and, because perfusion occurs in the smallest, most randomly oriented vessels, perfusion Doppler signal is particularly difficult to detect. Perfusion also constitutes the slowest flow and is comparable to tissue velocity caused by patient and sonographer hand motion. This causes a spectral overlap in the slow-time frequency domain which is conventionally used for separating and removing tissue signal. There is also evidence that this motion causes an overlap in the eigen-domain, affecting more advanced principal component analysis (PCA)-based tissue filtering techniques. Therefore, despite advances in Doppler processing, beamforming, and tissue filtering, tissue clutter interference with blood has remained a barrier for realizing non-contrast perfusion ultrasound imaging for real-time TACE treatment evaluation. To address this, we developed an adaptive time-domain tissue clutter demodulation scheme to reduce the tissue clutter bandwidth prior to tissue filtering to better separate tissue and blood. The method was validated and shown to potentially allow for blood velocities to be detected that were previously thought to be impossible to visualize without contrast. The technique was improved upon and combined with other perfusion-focused advancements. Additionally, an independent component analysis (ICA)-based tissue clutter filtering method was developed and integrated with other advancements, including the adaptive demodulation method, and was shown to be superior to PCA-based techniques. Finally, the method in combination with other perfusion-focused improvements was successfully applied during a preliminary study of patients undergoing TACE. | |
