Microvascular research最新文献

The morphology of nailfold capillaries serves as a crucial physiological parameter for analyzing human health status. However, during image acquisition, factors such as the nonplanar structure of the finger, lens depth-of-field limitations, and inaccurate focusing often cause defocusing and blurring, hindering physicians' observation of capillary structures and parameter measurements. To address this issue, this study proposes a wide-field nailfold capillary image deblurring method based on an improved MIMO-UNet architecture. In this study, a nailfold capillary dataset suitable for supervised learning was successfully constructed using image registration. During the deblurring process, a semantic residual feedback mechanism was introduced, which effectively enhanced the restoration accuracy of fine structures such as capillary loop edges and morphology. Additionally, a blurred-region attention module was designed to precisely identify blurred areas in nailfold images and prioritize the restoration of challenging regions, yielding clearer and more detailed capillary images. Experimental results demonstrated that the improved model achieves 3.82% and 0.22% higher PSNR and SSIM scores, respectively, compared with the original MIMO-UNet, while reducing MSE by 60.2%. Compared with other existing deblurring methods, this approach achieved the best performance in both accuracy and structural restoration of nailfold capillary images. Furthermore, static parameter measurements comparing deblurred images with real images show that differences in apical diameter, arterial limb diameter, and venous limb diameter are less than 0.995 μm, far below the physiological variation range. In summary, the proposed method demonstrates superior performance in both static parameter measurement accuracy and detailed restoration precision for nailfold images.

Red blood cells (RBC) deformability enables passage through capillaries, ensuring efficient microcirculatory flow and oxygen transport. Impaired deformability contributes to vascular complications and is associated with conditions such as sickle cell anemia, hereditary spherocytosis, and diabetes. Diamide and glutaraldehyde are commonly used in vitro to simulate pathological rigidity, yet their dose-dependent effects on RBC mechanics remain incompletely characterized. This study investigates the effects of diamide (10-200 μM) and glutaraldehyde (8-159,800 μM) on RBC morphology, deformability and viscosity at 20% hematocrit (HCT). Deformability was assessed by ektacytometry, viscosity by microfluidic viscometry, and morphology by defocusing microscopy with sphericity analysis. Osmolality and HCT were also measured to link the mechanical changes observed. Diamide produced a biphasic response: the maximum elongation index (EIₘₐₓ) decreased up to 140 μM, then partially recovered at higher concentrations. This coincided with increased sphericity, reduced cell volume and surface area, decreased HCT from hemolysis, and a non-monotonous viscosity profile. These effects reflect the combined influence of oxidative stress, vesiculation, and altered cell geometry. In contrast, glutaraldehyde induced an abrupt and irreversible loss of deformability at ≥7990 μM, while morphology and osmolality remained stable. HCT values were consistently lower across concentrations, which we attribute to reduced microtube filling efficiency by rigidified cells. Despite near-zero EIₘₐₓ at high concentrations, viscosity changes were modest due to extreme rigidification. These findings show that viscosity is governed not only by deformability but also by morphology, hemolysis and suspension dynamics.

3D fibrin bead angiogenesis assays are widely used to study endothelial sprouting in vitro, yet current analytical approaches are either time-consuming or poorly adaptable to complex imaging conditions, limiting quantitative assessment of co-cultures, spatial interactions, and nearest-neighbor-dependent angiogenic behavior. In this study, we developed a semi-automated user-interactive image analysis pipeline, Bead-based Endothelial Angiogenesis Data Suite (BEADS), to provide standardized quantitative bead-centric metrics of sprouting, migration, and spatial orientation in 3D fibrin angiogenesis assays. BEADS integrates automated bead detection with manual correction, followed by guided sprout and migratory-cell annotation across multi-channel image z-stacks. Novel analytical capabilities include co-culture designation, nearest-neighbor pairing, and circular statistics for sprout-directionality quantification. Performance was evaluated in assays using co-cultured male and female human pulmonary microvascular endothelial cell (HPMEC)-coated beads. BEADS reduced hands-on analysis time approximately sevenfold compared with manual tracing while preserving sprout-length accuracy against manual ground truth. BEADS provides a standardized, extensible platform for microvascular image analysis, supporting co-culture experimentation, spatial endothelial-interaction metrics, migratory-cell quantification, and high-throughput adaptation. This semi-automated workflow enables quantitative microvascular research by integrating computational precision with endothelial behavior and is broadly applicable to angiogenesis assays that incorporate co-cultures, perturbations, or multi-label experimental designs.

Tumor implantation on the chick chorioallantoic membrane (CAM) produces a characteristic radial or "starburst" microvascular pattern, classically attributed to tumor angiogenesis factor-driven sprouting. Here, we investigated the structural and functional basis of this pattern using intravital fluorescence microscopy, particle tracking, corrosion casting, scanning electron microscopy, and transcriptional profiling across 13 tumor cell lines. Eight cell lines successfully engrafted; six formed localized tumor masses associated with a rapid, well-defined radial vascular pattern evident within three days, while two exhibited diffuse, infiltrative growth with generalized capillary hypertrophy. Functional imaging revealed preferential flow paths, capillary loops, and U-turn particle trajectories at tumor margins. Structural analysis demonstrated that radial vessels arose predominantly from intussusceptive angiogenesis, with vessel duplication and intraluminal pillar formation, rather than endothelial proliferation alone. In contrast, diffuse tumors lacked radial organization despite robust capillary hypertrophy. Bulk transcriptional profiling failed to identify a distinctive angiogenic gene signature. These findings demonstrate that intussusceptive duplication underlies tumor-induced radial vascular patterning in the CAM.