Blood Vessel Model: Types and Uses in Medical Research

Blood vessel models span a wide range of technologies, from computer simulations built on patient scan data to living miniature arteries grown in a lab dish. Researchers use them to study how cardiovascular diseases develop, to test drugs before they reach a patient, and to plan complex surgeries. The reason no single model dominates is that each type captures something different about how real blood vessels behave, and every approach involves trade-offs between biological accuracy, cost, and throughput.

The Earliest Constructed Vessels

The idea of building a blood vessel outside the body dates to at least the late 1980s, when researchers assembled a multilayered tube from collagen, cultured smooth muscle cells, and endothelial cells reinforced with a synthetic mesh. Under electron microscopy the cells looked healthy and well-differentiated, and the construct withstood pressures similar to those inside an actual artery.

¹ That early proof of concept set the template for decades of refinement: start with a scaffold of some kind, seed it with the right cell types, and mature it under conditions that coax the cells into behaving as they would in the body.

Since then, the field has branched into several distinct model categories. Understanding the landscape means knowing what each category is good at and where it falls short.

Tissue-Engineered Scaffolds

One of the most mature approaches uses electrospinning, a technique that draws polymer solutions into extremely fine fibers and layers them into a tube. The resulting scaffold has a porous, fibrous architecture that loosely mimics the extracellular matrix surrounding real vessel cells. Researchers have tested scaffolds made from various biodegradable polymers. In one set of experiments, small-diameter tubes were implanted in rabbits and remained open and structurally intact over a seven-week observation period.

²

Electrospun scaffolds are attractive because the manufacturing parameters can be tuned to adjust fiber diameter, pore size, and overall mechanical stiffness. Different polymers and surface coatings can encourage or discourage cell adhesion, making it possible to tailor a graft for a specific application.

³ The limitation is biological complexity. A bare polymer tube does not replicate the signaling environment of a living vessel, and getting seeded cells to organize into the concentric layers of a real artery remains a challenge.

Bioreactors and Biomechanical Conditioning

Cells in arteries are constantly stretched, squeezed, and washed by pulsing blood flow. Reproducing those forces in the lab turns out to be critical for producing vessels with mechanical properties anywhere close to the real thing. Bioreactors do this by pumping culture medium through the lumen of a growing vessel construct, applying cyclic pressure that stretches the wall with every pulse.

⁴

One group embedded human skin fibroblasts in a fibrin gel, formed it into a tube, and subjected it to pulsatile flow in a bioreactor for seven to nine weeks. The fibroblasts remodeled the fibrin into circumferentially aligned collagen, a hallmark of real arterial walls. The resulting grafts achieved burst pressures of roughly 1,400 to 1,600 mmHg and had compliance comparable to native arteries, both numbers that matter because a vessel graft that cannot withstand normal blood pressure or flex with the pulse cycle is useless clinically.

⁵

Organ-on-a-Chip and Microfluidic Devices

Microfluidic blood vessel models, sometimes called “vessels-on-a-chip,” shrink the vascular environment down to a small plastic or glass device containing tiny channels lined with living endothelial cells. Fluid is pushed through these channels at controlled rates, producing the shear forces that vessel-lining cells experience in the body. The approach is inherently suited to high-throughput work because the devices are small, relatively cheap to manufacture, and can be run in parallel.

A recently described automated platform uses a custom 384-well plate format to generate over one hundred perfusable blood vessels in a single production run, each embedded in hydrogel and supported by stromal cells like fibroblasts and pericytes. Adding those stromal cells significantly improved the barrier function of the vessels and changed how the constructs responded to chemotherapy drugs and inflammatory challenges.

⁶ That kind of scalability is the selling point: if you need to test dozens of drug candidates against a living vascular barrier, a 384-well chip array can do what a tissue-engineered graft in a bioreactor cannot.

The field still needs better models that combine multiple cell types, more realistic extracellular matrix, and three-dimensional geometry, especially for disease states where the architecture of the vessel wall itself is part of the problem.

⁷

3D Bioprinting

Bioprinting takes the organ-on-a-chip concept a step further by using specialized printers to deposit cells and biomaterials in precise spatial arrangements. A coaxial extrusion system, for example, can lay down concentric layers of bio-ink through a nozzle-within-a-nozzle arrangement, producing a hollow, perfusable tube in a single printing step.

⁸

More recently, embedded 3D printing strategies have made it possible to build freeform vascular networks that branch and curve through a surrounding tissue. One approach uses a microgel-based bio-ink that acts as both the structural material and a suspension medium for the printing process. Because the gel is shear-thinning, meaning it becomes fluid under the pressure of the print nozzle and then re-sets, the printer can trace out vessel paths in any direction without them collapsing. The technique was demonstrated by printing a ventricle model with a perfusable vascular network, a geometry that earlier printing strategies could not achieve.

⁹

Computational and In Silico Models

Not every blood vessel model involves living cells. Computational fluid dynamics (CFD) simulations reconstruct blood flow inside a vessel digitally, typically starting from a patient’s CT or MRI scan. The computer calculates velocity, pressure, and wall shear stress at thousands of points along the vessel surface, revealing hemodynamic details that cannot be measured directly in a living patient.

¹⁰

One application is modeling aortic coarctation, a narrowing of the aorta. Simulations showed that as the degree of narrowing increased, the pressure drop across the stenosis rose and the velocity profile changed dramatically. Wall shear stress peaked just upstream of the narrowing and dropped to its lowest levels within the narrowed segment, with exact locations depending on whether the stenosis was symmetric or asymmetric. Those details matter because regions of low or disturbed shear stress are where atherosclerotic plaques tend to form and where aneurysms may develop.

¹¹

A fair question is how accurate these simulations really are. A head-to-head comparison on an intracranial aneurysm measured wall shear stress using four methods: two physical flow-imaging techniques, in vivo MRI, and CFD. All four agreed on the overall flow pattern and the spatial distribution of shear stress across the aneurysm surface. However, there were large differences in absolute values. The MRI estimates of peak wall shear stress were two- to four-fold lower than those from the other methods.

¹² The practical takeaway is that CFD is reliable for predicting relative patterns (where shear stress is high versus low) but less reliable for pinning down exact magnitudes, at least when compared to imaging-based measurements.

Animal Models

Animal studies remain widely used for cardiovascular research, ranging from mice genetically engineered to develop atherosclerosis to large animals like pigs whose heart size and coronary anatomy more closely resemble a human’s. The advantage is obvious: a living animal has all the systemic complexity that any lab model lacks, including an immune system, hormonal regulation, and multi-organ interactions.

The downsides are equally clear. No single animal model perfectly recreates a human cardiovascular disease. Genetic and environmental factors create mismatches between the animal version of a disease and the human one, and practical concerns like cost, housing infrastructure, and the need for specialized surgical teams limit what most labs can do.

¹³ Methodological problems compound the biological ones. Many large-animal cardiovascular studies suffer from inadequate randomization, underpowered sample sizes, and insufficient blinding, which helps explain why therapies that look promising in preclinical work often fail to translate to humans.

¹⁴

Regulatory shifts are also reshaping the landscape. The FDA Modernization Act 3.0 and new NIH initiatives have increased funding for human-cell-based alternatives to animal testing, encouraging the development of organ-chip platforms as viable stand-ins for at least some animal studies.

¹⁵

Modeling Atherosclerosis and Thrombosis

Understanding how plaques form and blood clots develop is one of the most active applications of blood vessel models. In real arteries, atherosclerotic plaques tend to appear at bends, branch points, and other locations where blood flow is disturbed rather than smooth. Disturbed or oscillatory shear stress at these sites alters endothelial cell behavior in ways that promote plaque growth, inflammation, and vessel wall remodeling.

¹⁶

Experimental work has teased apart the contributions of shear stress magnitude versus direction. In one model, regions exposed to changes in shear magnitude alone developed plaques with a more dangerous, vulnerable character, while regions where both magnitude and direction were disrupted tended to form plaques with more stable features.

¹⁷ High shear stress, meanwhile, can drive overexpression of growth factors and nitric oxide in the vessel wall, which in turn promotes the sprouting of fragile new blood vessels within the plaque itself, a feature associated with plaque vulnerability.

¹⁸

For thrombosis specifically, microfluidic devices have become powerful testing grounds. One atherothrombosis-on-a-chip recreated the geometry of a stenosed vessel and showed that platelet aggregation tracked closely with local shear rates. The device was sensitive enough to detect the antiplatelet effects of aspirin, with therapeutic doses delaying platelet adhesion and reducing the area of thrombus in a dose-dependent manner.

¹⁹ A separate device focused on generating fully occlusive clots at arterial shear rates and used the time it took for the channel to completely block as an objective readout. When the antiplatelet drug eptifibatide was tested, the time to occlusion increased significantly compared to untreated controls.

²⁰

Aneurysm Risk Assessment

Computational models have a particularly compelling application in cerebral aneurysms, where the clinical question is often whether an unruptured aneurysm is likely to burst. A study using patient-specific CFD models of over a hundred intracranial aneurysms found that ruptured aneurysms tended to have complex or unstable flow patterns, small impingement regions, and narrow jets. Aneurysms with small impingement sizes were about six times more likely to have ruptured than those with large impingement sizes.

²¹ That kind of patient-specific hemodynamic profiling could eventually help clinicians decide which aneurysms warrant preemptive treatment and which can be monitored safely.

Cancer and Tumor Vascularization

Tumors cannot grow beyond a few millimeters without recruiting their own blood supply, a process called angiogenesis. Modeling this process in the lab is important both for understanding how cancers spread and for testing drugs designed to starve them of blood.

A vascularized lung cancer-on-a-chip embedded tri-cellular spheroids (tumor cells, endothelial cells, and fibroblasts) in a lung-derived hydrogel with perfusable channels mimicking small blood vessels. When the anticancer drug doxorubicin was tested on this platform, its dose-dependent killing effect was more pronounced than in flat, two-dimensional cell cultures, underscoring how the three-dimensional vascular context changes drug response.

²²

Another group engineered the microenvironment around a tumor to induce controlled angiogenesis and found that the direction of interstitial flow governed which way new capillaries sprouted: they grew opposite to the flow direction. Adding lung fibroblasts improved the continuity and lumen formation of those sprouted capillaries. The resulting perfusable vascular network allowed enhanced delivery of both anticancer drugs and immune cells to the tumor spheroids compared to unvascularized controls.

²³

The newest iterations incorporate patient-derived tumor organoids into the chip. One such platform used tumor-specific microvasculature and showed that highly metastatic tumor cells actively induced nearby vessels to sprout while simultaneously migrating toward those blood vessels, recapitulating a key step in metastasis.

²⁴

Blood-Brain Barrier Models

The blood-brain barrier is a specialized vascular structure where the endothelial cells lining brain capillaries form extremely tight junctions, restricting what can pass from the bloodstream into brain tissue. This makes drug delivery to the brain notoriously difficult, and it also means that standard vascular models are not useful for brain-targeted work. Researchers need models that specifically replicate the barrier’s restrictiveness.

Microfluidic blood-brain barrier chips have been built using brain endothelial cells derived from human stem cells, co-cultured with astrocytes on opposite sides of a porous membrane under constant perfusion. These platforms maintained barrier function for up to ten days and yielded permeability measurements for test compounds like caffeine and doxorubicin that were comparable to values measured in living animals.

²⁵ The ability to match in vivo permeability is what makes these chips useful as drug-screening tools, since permeability values measured in flat cell cultures often overestimate how easily a drug crosses the barrier in a real brain.

²⁶

Neonatal blood-brain barriers behave differently from adult ones, and a dedicated neonatal barrier chip demonstrated this concretely. When brain endothelial cells were cultured alone, the barrier was quite leaky. Adding astrocyte-conditioned medium or co-culturing with astrocytes reduced permeability to a large molecule by roughly 15- to 40-fold, highlighting just how dependent barrier function is on the supporting cell environment.

²⁷

Surgical Planning with 3D-Printed Replicas

A completely different category of blood vessel model serves surgeons rather than bench scientists. Using patient CT scan data, 3D printers can produce life-size replicas of a person’s vasculature in plastic or silicone. The surgeon then uses the replica to rehearse a procedure, select the right catheter and guidewire sizes, and anticipate tricky anatomy before ever touching the patient.

In one reported case involving multiple splenic artery aneurysms, the surgical team printed a model of the patient’s arteries ahead of the endovascular procedure. They used the model to determine the ideal puncture site, select the optimal combination of guide catheter, base catheter, and microcatheter, and practice navigating the anatomy. During the actual procedure, the catheter combination chosen during testing worked on the first attempt, and the need for repeated X-ray views to figure out angles was minimized.

²⁸

For aortic aneurysm repair, where custom-fenestrated stent grafts must align precisely with branch arteries, 3D-printed models let the clinical team test the device in a simulation that mimics the X-ray environment of the operating room. In at least one case, rehearsal on a patient-specific phantom led to a complication-free procedure.

²⁹ These printed models are also useful for testing new commercial devices before they go near a patient, and for training residents on complex vascular anatomy.

³⁰

Drug Screening and Toxicity Testing

Pharmaceutical companies need to know whether a drug candidate damages blood vessels before it reaches human volunteers. Traditional testing uses animal studies or flat-dish cell cultures, but tissue-engineered blood vessels offer a middle ground: they are human, three-dimensional, and can be made at moderate throughput. Small-scale endothelialized constructs have shown promise for screening drug candidates, though challenges remain in scaling them up for the large compound libraries that pharma pipelines demand and in maintaining the differentiated state of the vessel wall cells over time.

³¹

Models of vascular aging add another dimension. One group induced senescence in a tissue-engineered blood vessel by exposing it to hydrogen peroxide for a week, simulating chronic oxidative stress. Both the endothelial cells and the supporting mural cells became senescent, measured by increased markers of cell-cycle arrest and reduced production of nitric oxide synthase. The functional consequences on vasoreactivity, however, were specific to the endothelial layer, illustrating how a layered vessel model can tease apart cell-type-specific effects in a way that a single-cell-type culture cannot.

³²

Toward Implantable Vessels and Regenerative Medicine

The ultimate ambition for many tissue-engineered blood vessel models is implantation: creating a graft that can replace a damaged or blocked artery in a patient. The challenge is not just making a tube that withstands pressure but ensuring it connects to the host’s own circulation and stays open over time.

Pre-vascularization is one strategy to improve the odds. Rather than implanting a bare scaffold and hoping host blood vessels will grow into it, researchers seed the construct with endothelial cells arranged into cord-like patterns before surgery. In mouse studies, grafts containing pre-patterned endothelial cords recruited blood and formed larger-diameter vessels compared to grafts where endothelial cells were distributed homogeneously. Vessels in the patterned grafts averaged close to 20 micrometers in diameter, while the unpatterned versions were mostly under 10 micrometers.

³³ The same study found that engraftment success depended heavily on the host tissue site, a reminder that even a well-engineered construct interacts with its biological surroundings in ways that are hard to predict.

A range of pre-vascularization methods are now in development, including bioprinting vascular channels directly into tissue constructs, micropatterning to guide vessel growth, and cell sheet engineering that stacks thin layers of vessel-forming cells.

³⁴

Artificial Intelligence in Vascular Modeling

Machine learning is beginning to overlay existing vascular models with predictive capability. AI algorithms trained on patient imaging data can identify patterns that predict disease progression, aneurysm growth, or response to treatment, often without researchers needing to specify in advance which variables matter. Reviews of the field highlight applications in aortic aneurysm management, peripheral arterial disease, and carotid stenosis, with the broader goal of developing precision-medicine approaches tailored to individual patients’ vascular anatomy and risk profiles.

³⁵ The combination of patient-specific computational flow models with AI-driven risk prediction is one of the more active frontiers, though clinical validation of these tools is still in its early stages.