A spatial coherence filtering method to increase resolution in color-flow ultrasound imaging

Unmet Need

Color-flow ultrasound imaging is a technique that allows for visualization of blood flow within tissues and is frequently used to diagnose cardiovascular abnormalities, guide needle biopsies, examine tumor vascularity, and measure blood flow velocity. In the case of carotid arterial disease, a precursor to stroke affecting about 5% of the US population, ultrasound assessment and peak blood flow velocity measurements derived therein are often the only criteria used to determine whether intervention is necessary. Despite its use in several vital medical procedures, color-flow ultrasound imaging remains notoriously poorly resolved, due to the inability of these systems to accurately filter out ‘clutter’, or background sound waves from the environment. Existing ultrasound systems rely on static filters that attempt to filter out clutter based on standard thresholds of motion and magnitude of background sound waves. However, these simple thresholds have proven insufficient to effectively discriminate clutter, especially when background waves have similar magnitude and motion to ultrasound waves of interest. Furthermore, when these thresholds are set too low or too high, vital diagnostic measurements such as blood flow velocity become skewed. As a result, poor ultrasound quality and resolution can drastically affect diagnosis of potentially fatal heart conditions and resulting medical interventions chosen for patients. Given this, there is a need for a filtering technology that assesses clutter on an ongoing basis rather than as a static feature and can effectively identify and filter out clutter.

Technology

Duke inventors have developed an improved ultrasound imaging method that more accurately visualizes cardiovascular abnormalities and improves the accuracy of blood flow velocity measurements. This is intended to be used as a new data analysis platform for ultrasound color flow imaging data, which has become the standard mode of imaging in almost all diagnostic ultrasound systems. Specifically, ensemble echo data is passed through a bank of different clutter filters and a unique feature of waves called spatial coherence is measured from the output of each filter. Spatial coherence has been demonstrated to be a more reliable metric than motion or magnitude for discriminating between ultrasound waves and clutter. The filter which yields the maximum spatial coherence is selected for subsequent use. With this methodology, multiple features of the clutter waves are used to dynamically select for the optimum filter to discriminate against them, rather than relying on a standard threshold. This introduces an added dimension to better discriminate between flow and clutter signals and improve filtering, particularly under conditions where motion and magnitude-based methods fail. Compared to existing filtering technologies, coherence-adaptive clutter filtering more effectively suppresses bias from clutter while more accurately assessing flow rates as compared to ground truth. Because of this added resolution, coherence-adaptive filtering also resolved important biological structures such as small blood vessels not otherwise seen with standard imaging technologies. This has been demonstrated in simulation experiments using existing ultrasound data, and in in vivo liver and fetal imaging in patients.

Advantages

• Dynamic system that adjusts to different background conditions to effectively filter out clutter
• Increased resolution during imaging allows visualization of otherwise unseen biological features
• Easily integrated into existing ultrasound imaging platforms
• Can be used live, or on previously acquired ultrasound data