Comparing slow-flow sensitivity in vivo between superharmonic and conventional ultrasound localization microscopy

Ultrasound localization microscopy (ULM) is a microvasculature imaging technique that can resolve vessels beyond the diffraction limit. A key step is the separation of microbubbles (MB) from tissue. Conventional ULM uses spatiotemporal filters, such as Singular Value Decomposition (SVD), that are insensitive to slow flow and easily degraded by motion. Superharmonic (SpH) imaging addresses these limitations by leveraging the nonlinear physics of MB interaction to separate the signals in the frequency domain while preserving slow and stationary flow. The slow-flow sensitivity of SpH ULM is compared with conventional ULM. Simultaneous 2-D SpH and 2-D B-mode images were acquired continuously at a frame rate of 400 Hz in an internal kidney, in a freely breathing rodent. Each acquisition contained 120 000 frames. MB localization and tracking were performed on SpH and SVD-filtered B-mode frames (with a singular value cutoff of 200/1000). Velocities and track counts were compared in manually segmented kidney regions. SpH ULM generated 13 and 2600 times more slow tracks (per unit area) ( < 10 mm/s) in the cortex and medulla, respectively, than conventional ULM. The mean slow track velocity in the cortex was additionally lower in SpH ULM. Conversely, conventional ULM resolved 26 times more fast tracks (10–30 mm/s) in the cortex. These results highlight the enhanced sensitivity of SpH imaging for slow-flow ULM (with a potential fast-flow tradeoff) toward future applications in kidney glomerular and tumor imaging.