In a 20 mouse glioma (a type of brain tumor) study, super resolution ultrasound tracked four vascular parameters at once as tumors grew denser, shorter, slower, and more disordered over three weeks.
Ultrasound localization microscopy can now read four vascular parameters at once inside a living mouse brain: vessel density, average vessel length, blood flow speed, and the disorder of flow direction. In a 2026 study of 20 mice bearing orthotopic GL261 gliomas, those four measurements shifted in the same direction over roughly three weeks of tumor growth, producing a microvascular fingerprint that no single number captures.
GL261 is a standard mouse glioma cell line, and the study does not involve human patients.
Conventional ultrasound cannot resolve individual vessels smaller than roughly half a millimeter because of the diffraction limit. ULM breaks that ceiling by tracking injected microbubbles, which are gas-filled spheres the size of red blood cells, as they move through the vasculature frame by frame. Each microbubble appears as a bright point. Localizing thousands of those points across frames yields maps of single vessels at micrometer resolution. The technique has been used for several years in animal vasculature and, more recently, in selected human studies. The new paper, published in Annals of Biomedical Engineering, applies the method to mouse gliomas and reads four quantitative parameters from the same imaging session: structure and function in one scan.
On the tumor side, vessel density rose about 30% relative to the opposite (contralateral) hemisphere at day 14 and about 56% at day 21. Average vessel length shortened: roughly 124 micrometers versus 193 micrometers on the contralateral side at day 14, and 87 micrometers versus 234 micrometers at day 21. By day 21, flow velocity had fallen to 3.7 millimeters per second inside the tumor compared with 4.6 millimeters per second on the opposite side. The flow-direction distribution also became more disordered at both timepoints, a marker of chaotic, branching vasculature rather than the orderly flow seen in normal tissue.
Two validation steps anchor the structural numbers. ULM-derived vessel density correlated strongly with CD31 immunostaining, the standard histology marker for blood-vessel endothelium, and with perfusion imaging using the fluorescent dye DiI. The cross-checks serve a methodological purpose. A single density measurement can be confounded by imaging angle, vessel overlap, or contrast timing. Four independent parameters, each validated against histology, narrow what "vascular remodeling" can mean in a live animal and turn the scan from a single number into a coordinated signature.
A 2025 medRxiv preprint reports the first intraoperative ULM scans in human brain tumors, and a 2025 ScienceDirect study examined microcirculatory heterogeneity in human glioblastoma. The new paper adds a multi-parameter mouse-mode demonstration to that set; the human-side evidence is still early.
The GL261 day-14-to-day-21 signature is the demonstration, not the conclusion. Whether the same four-parameter pattern holds across other mouse glioma lines and across the molecular subtypes of human glioblastoma is the open question the field will have to answer before any clinical reading can rest on it.