Abstract
INTRODUCTION: Spinal cord injury (SCI) leads to irreversible neurological deficits driven largely by secondary pathophysiological processes, including microvascular dysfunction and glial activation. These changes occur over days to weeks following the initial trauma and play a critical role in lesion expansion and functional deterioration. Conventional imaging methods cannot resolve capillary-scale vascular changes in vivo, limiting clinicians’ ability to assess vascular integrity and therapeutic response. Ultrasound Localization Microscopy (ULM), a super-resolution imaging technique, offers real-time, non-invasive visualization of spinal microvasculature at resolutions below 10 μm, enabling quantification of vascular density, perfusion speed, and vessel morphology. METHODS: Twelve adult female Sprague-Dawley rats (250–300 g) were used. Nine underwent T9–T11 contusion injury using a 200 kdyn impactor and were divided into three post-injury groups: acute (Day 1), subacute (Day 7), and chronic (Day 21). Two uninjured animals served as controls. B-mode ultrasound localized the injury epicenter, while ULM with intravenous Definity® microbubbles was used to reconstruct high-resolution microvascular maps and assess vessel density (mVD), flow speed (mFS), and tortuosity in dorsal and ventral spinal cord regions. Behavioral recovery was evaluated using the Basso, Beattie, and Bresnahan (BBB) locomotor scale, and glial activation was quantified using GFAP immunohistochemistry. RESULTS: ULM enabled visualization and quantification of dynamic vascular alterations with high spatiotemporal fidelity. Venous mVD at the injury site decreased significantly from 3.7 in uninjured controls to 0.57 in the acute phase, 0.09 in the subacute phase, and 0.22 in the chronic phase (p < 0.001). Arterial mVD at the injury site was less affected, decreasing from 1.9 (control) to 0.29 (acute), partially recovering to 0.47 (subacute) and 0.83 (chronic). Rostral and caudal regions exhibited similar trends, with the lowest values observed in the subacute group. Mean blood flow speed (mFS) also declined from a baseline of 1.31 mm/s in controls to 1.03 mm/s (acute), 1.04 mm/s (subacute), and 1.10 mm/s (chronic) (p < 0.01). Arterial flow speed at the injury site showed a biphasic pattern, dropping to 0.90 mm/s (acute) and rising to 1.25 mm/s (chronic), while venous mFS was lowest in the subacute group (0.71 mm/s) before partial recovery. Vessel tortuosity increased during injury progression, particularly in dorsal venous structures. GFAP-positive area increased progressively from 6.1% (acute) to 12.3% (subacute) and 18.7% (chronic) (p < 0.01). Strong correlations were observed between venous mVD and GFAP expression (ρ = –0.89), arterial mFS and GFAP (ρ = 0.83), and BBB score with mVD and mFS (ρ > 0.8). BBB scores improved from 5.3 ± 0.9 (acute) to 12.7 ± 1.1 (chronic), reflecting partial functional recovery. CONCLUSIONS: This study establishes ULM as a transformative, non-invasive modality for tracking microvascular and glial remodeling in SCI with high spatial and temporal resolution. ULM-detected reductions in venous density and flow velocity were early and persistent indicators of tissue compromise, while arterial flow showed partial recovery aligned with glial activation and functional improvement. These imaging biomarkers offer strong potential for SCI stratification, monitoring therapeutic response, and guiding vascular-targeted interventions. Future clinical translation of ULM could revolutionize spinal cord injury management by enabling bedside assessment of neurovascular health in real time.