Engineered Blood Clots Bring Lab Precision to Brain Hemorrhage Surgery Testing

Intracerebral hemorrhage, the deadliest and most disabling form of stroke, occurs when a ruptured vessel floods brain tissue with blood that rapidly coagulates into a hematoma. Surgeons increasingly turn to minimally invasive evacuation, threading catheters through the skull to remove the clot while sparing healthy tissue. Yet the devices designed for these procedures have long been tested against materials that only loosely resemble the real thing. A team at the University of California, San Francisco, led by neurosurgical researcher Prabhat Kumar and neurosurgeon Luis E. Savastano, has now built and rigorously characterized a family of engineered blood clot analogues designed to close that gap, publishing their findings in the Annals of Biomedical Engineering.

The central problem the researchers set out to solve is deceptively simple to state and notoriously difficult to execute. A blood clot is not a uniform blob but a composite material: a porous mesh of fibrin fibers, the protein scaffold that forms during coagulation, entrapping red blood cells, platelets, and plasma proteins. How such a clot deforms, fractures, or resists suction depends on the relative proportions and organization of these components. Previous analogue models, most of them developed for ischemic stroke thrombectomy, captured some of this behavior but were rarely validated against hemorrhagic clots across multiple modes of deformation, leaving device developers without a trustworthy benchmark.

To build their analogues, the team mixed packed red blood cells with fresh frozen plasma at three precise volumetric ratios: 85:15, 90:10, and 95:5 plasma to cells. Coagulation was triggered by adding calcium chloride, and the resulting clots were sectioned into roughly 150 milligram samples. The choice to prepare clots under near-static conditions was deliberate. Unlike arterial thrombi, which form under fast, dynamic flow that reshapes fibrin architecture, intracerebral hematomas accumulate in the relatively stagnant environment of brain parenchyma after a vessel ruptures, so static formation better mirrors the clinical scenario the devices must ultimately handle.

Mechanical fingerprinting then proceeded on three fronts. Oscillatory shear rheology on a rotational rheometer probed the clots’ viscoelastic character, steady flow sweeps measured how apparent viscosity changes with shear rate, and unconfined axial compression tracked how the materials stiffen under large deformation. All testing was performed at 37 degrees Celsius to approximate physiological conditions, with five independent samples per formulation. Scanning electron microscopy, at a standardized magnification of 6700 times, revealed the microstructural architecture underlying the measured properties.

The rheological results told a clear compositional story. Every formulation behaved predominantly as a solid, with the storage modulus exceeding the loss modulus across the entire strain range. Stiffness rose with fibrin content: the storage modulus at low strain increased from about 334 pascals in the cell-rich 85:15 mix to roughly 500 pascals in the fibrin-rich 95:5 mix, a statistically significant difference. Intriguingly, the loss tangent, a dimensionless ratio describing the balance between elastic and viscous behavior, was nearly identical across all three formulations and overlapped with values measured in patient-derived clots, suggesting the analogues reproduce the relative viscoelastic character of real hematomas even where absolute stiffness differs.

Flow behavior inverted that trend. The red-cell-rich 85:15 formulation exhibited by far the greatest resistance to flow, with a peak viscosity of about 7367 pascal-seconds, more than double that of the fibrin-rich formulations, which clustered near 2800 to 3400 pascal-seconds. All formulations showed non-Newtonian behavior, with viscosity peaking at low shear rates and collapsing at higher rates as the internal network broke down. This composition dependence matters for device design: a catheter-based aspiration tool must contend with very different flow resistance depending on whether it faces a cellular, sludge-like clot or a tough, fibrin-dominated one.

Compression testing exposed the limits of the current recipes. Mean compressive stress at roughly 20 percent strain climbed with plasma content, from about 0.38 kilopascals in the 85:15 mix to about 1.17 kilopascals in the 95:5 mix, but the trend fell just short of statistical significance. More strikingly, the three patient-derived ICH specimens, retrieved during surgery and tested within two hours of extraction, were dramatically stiffer, with compressive stresses ranging from about 3 kilopascals to nearly 19 kilopascals, a sixfold spread among just three patients. The analogues, in short, capture viscoelastic character but remain far softer than real hematomas under large deformation, a limitation the authors attribute to their simplified composition, which lacks platelets, leukocytes, extracellular DNA, and the clot contraction and crosslinking processes that stiffen native clots.

Electron microscopy tied the mechanical data to structure. The cell-rich formulation showed densely packed, rounded cellular aggregates with sparse fibrin connectivity, while the fibrin-rich formulation displayed a continuous, interconnected mesh of fibers with few cellular inclusions. The intermediate 90:10 recipe bridged the two, embedding cells within a connected fibrin matrix. Because patient specimens were prioritized for mechanical testing, direct microstructural comparison with human clots was not possible, so the imaging evidence speaks to structure-property relationships within the analogues rather than proving equivalence with native tissue.

The study’s most consequential contribution may be conceptual rather than material. Rather than chasing a single perfect surrogate, the authors propose a panel of three mechanically distinct formulations spanning softer, flow-resistant to stiffer, fibrin-dominant behaviors, allowing researchers to select the phenotype most relevant to a given experiment. The overlap in loss tangent with patient clots provides genuine clinical anchoring, while the compression shortfall defines a transparent target for next-generation recipes incorporating platelet-rich plasma, optimized thrombin, and Factor XIII-mediated crosslinking.

For a field where device success hinges on the messy interaction between hardware and clot, reproducible and tunable test materials are a quiet but genuine breakthrough. With a validated cadaveric hemorrhage model already developed by the same group, these analogues could standardize benchtop evaluation of minimally invasive evacuation technologies, shorten the path from prototype to operating room, and, if future iterations close the stiffness gap, give engineers a faithful rehearsal stage for one of neurosurgery’s most demanding procedures.

Subject of Research: Mechanical characterization of fibrin-based blood clot analogues for modeling intracerebral hemorrhage

Article Title: Mechanical Characterization of Fibrin-Based Blood Clot Analogues for Modeling Intracerebral Hemorrhage

Article References: Kumar, P., Liu, A., Senol, Y. C., Leong, D., Krishnan, N., & Savastano, L. E. (2026). Mechanical Characterization of Fibrin-Based Blood Clot Analogues for Modeling Intracerebral Hemorrhage. Annals of Biomedical Engineering. https://doi.org/10.1007/s10439-026-04398-x

Image Credits: AI Generated

DOI: 10.1007/s10439-026-04398-x

Keywords: intracerebral hemorrhage, blood clot analogues, fibrin, rheology, viscoelasticity, biomedical engineering, minimally invasive surgery, red blood cells, compression testing, scanning electron microscopy, stroke, hematoma evacuation