Phys Med Biol. 2026 Jul 14. doi: 10.1088/1361-6560/ae8aac. Online ahead of print.
ABSTRACT
OBJECTIVE: Coronary microvascular dysfunction plays a central role in major cardiovascular diseases, yet non-invasive imaging of the human coronary microcirculation remains an unresolved challenge in clinical practice. Ultrasound Localization Microscopy (ULM) enables microvascular imaging beyond the diffraction limit, but whole-heart volumetric ULM is hindered by cardiac motion, rib-induced acoustic aberrations, limited control over microbubble (MB) dynamics, and hardware-related field-of-view constraints. Experimental optimization under clinical conditions is therefore challenging.
APPROACH: We developed a dedicated simulation framework for three-dimensional coronary ULM by extending a previously validated brain ULM simulation framework to the human coronary circulation. A synthetic coronary vascular network with physiologically realistic hemodynamics was generated using a space-colonization algorithm combined with Murray's law and Poiseuille flow assumptions. Ground-truth MB dynamics were simulated and convolved with experimentally measured in vitro point spread functions, including configurations with human ribs to reproduce acoustic aberrations. The framework generates realistic four-dimensional (x, y, z, t) ultrasound datasets for quantitative benchmarking of MB detection, localization, and tracking performance.
MAIN RESULTS: Systematic simulations demonstrated that both cardiac motion and rib-induced aberrations significantly degrade coronary ULM performance, particularly in small vessels, leading to reduced precision, sensitivity, and Jaccard index. Cardiac gating partially restored performance but increased acquisition time, underscoring critical trade-offs between robustness and efficiency.
SIGNIFICANCE: This simulation framework provides a realistic and flexible benchmarking platform for the evaluation of volumetric coronary ULM under controlled conditions. It enables quantitative assessment of acquisition strategies, probe designs and image-processing algorithms, thereby supporting the development and clinical translation of three-dimensional coronary ULM.
PMID:42447911 | DOI:10.1088/1361-6560/ae8aac