A newly developed analytical platform enables standardized, quantitative analysis of tiny deep‑tissue blood vessels using ultrasound localization microscopy (ULM), supporting non‑invasive tracking of early disease‑related microvascular changes. Published in BME Frontiers, the open research establishes U‑VBA (ULM‑based Vascular Biomarker Automated Analysis), a ready‑to‑use framework bridging super‑resolution ultrasound reconstruction and automated vascular biomarker quantification.
Microvascular dysfunction serves as an early hallmark for glaucoma, cancer and multiple systemic illnesses. Although ULM breaks ultrasound’s diffraction limit and visualizes microvasculature several centimeters deep in tissue, the absence of unified processing pipelines hindered its real‑world clinical translation. U‑VBA incorporates motion correction and cascaded denoising workflows, and computes six complementary vascular biomarkers: density, intervessel distance, diameter, flow velocity, perfusion and tortuosity, capturing both structural and hemodynamic vessel properties.
Researchers validated U‑VBA across phantoms, animal models and patient samples. In chicken chorioallantoic membrane tests, the tool resolved capillaries down to 16.2 μm. In rabbit optic‑nerve‑injury glaucoma models, it longitudinally captured dynamic vascular remodeling during intraocular‑pressure elevation and partial recovery. Thirteen biomarkers exhibited significant post‑injury alterations, consistent with retinal neural damage observed in histology. U‑VBA showed high reproducibility (intra‑class correlation coefficient 0.981) and outperformed a benchmark algorithm by producing cleaner microvascular maps with minimal background noise.
The team further tested U‑VBA on 39 patients with three categories of cervical lymph‑node conditions: benign lesions, lymphoma and metastatic carcinoma. Four biomarkers displayed significant inter‑group differences. A support‑vector‑machine classifier built on these multiparametric features achieved 85% classification accuracy in five‑fold cross‑validation, demonstrating potential to reduce reliance on invasive biopsies. U‑VBA is compatible with commercial clinical ultrasound systems and fits into existing clinical workflows.
“U‑VBA delivers standardized multiparametric vascular readouts that were previously hard to obtain from ULM data,” said Dr. Xuejun Qian. The team notes limitations including small animal sample size, 2D imaging and contrast‑agent dependence and plans further improvements toward 3D and contrast‑free imaging.
BME Frontiers
A General-Purpose Framework for Microvascular Reconstruction and Quantitative Analysis in Ultrasound Localization Microscopy
3-Aug-2026