In his latest work presented at IEEE IUS 2025, Lauge and collaborators investigate how spatial smoothing can improve the accuracy of super-resolution ultrasound (SRUS) vessel width estimates when compared with micro-CT.
Using both ultrasound localization microscopy (ULM) and super-resolution using erythrocytes (SURE) in a rat kidney model, the team compared SRUS-derived vessel widths to micro-CT data with 5.45 μm resolution. They found that modest Gaussian smoothing (σ = 15 μm)—matched to the system’s effective resolution—reduced width underestimation by up to 73%, yielding near-unbiased results across vessel classes.
These findings demonstrate that resolution-matched preprocessing can mitigate systematic errors in SRUS structural quantification, bringing ultrasound-based microvascular imaging even closer to micro-CT accuracy.
