3D printing of stents via two-photon polymerization

Abstract

Stents are medical devices used to treat the narrowing of the blood vessel, most commonly caused by atherosclerosis. Currently used bare-metal, drug-eluting stents are limited in size and architectural complexity, and there are a few risks associated with these medical devices. In some cases stents can cause thrombosis or even death. Furthermore, said risk increases while stenting relatively small vessels. This paper shows that vascular stents for relatively small vessels and/or for patient-specific stenting applications can be printed using two-photon polymerization (2PP). 2PP is an additive manufacturing technique with sub-$\mu$m resolution and unlimited 3D geometry potential. These capabilities were used to produce stents for small blood vessels, which are 5 mm tall and 0.7 or 0.9 mm in diameter, with 3D struts as thin as 50 $\mu$m. Several novel approaches were introduced to accommodate the printing of such a structure like voxel elongation and printing in stereolithography-like vat-sample holder configuration. Furthermore, the produced stents were tested mechanically proving their mechanical resilience to most common types of mechanical deformations. Experimental results were also compared to mathematical modeling, showing excellent agreement, hinting at the possibility of designing and testing complex micro-stent geometries before fabrication in silico. Finally, biocompatibility experiments were performed, in which rats survived the 7-day incubation period and showed no significant biocompatibility issues. Overall, the presented work gives an outlook on all aspects related to the 3D printing of stents - design, manufacturing, mechanical, and biological testing. We show that 2PP aligns very well with all these aspects and is a promising technique for the mass manufacturing of vascular stents for small vessels or specifically for the patient.

Introduction

According to the World Health Organization, one of the leading causes of death globally, taking up to 18 million lives each year, are cardiovascular diseases (CVDs)¹. Coronary artery disease (CAD) is a type of CVD condition in which the blood vessels become narrow and hardened because of vessel atherosclerosis². One of the treatment options for CAD is stenting. Nowadays, several types of stents are used for the treatment: bare-metal, drug-eluting, and bioresorbable stents³. The most widely used are drug-eluting stents (DES), a vascular implant releasing bioactive compounds that reduce the risk of restenosis or abnormal tissue growth⁴. Current stenting solutions, such as the newer generation of DES, offer promising treatment results. However, there are a few risks associated, such as late stent thrombosis, which can lead to myocardial infarction and even death⁵. Thus, there is ever preset need to further expand manufacturing capabilities, with the aim to expand the field and address mentioned challenges.

Currently, a new technique for producing stents has emerged - 3D printing. The main advantage of said technology is the capability to manufacture complex architectural structures out of a wide selection of materials⁶. As a result, 3D printing has gained recognition as a new option with the potential for manufacturing different structures for CVD treatment. However, several factors should be considered when printing structures for medical treatment applications. For example, when additive manufacturing is used for cardiovascular applications, the resolution of the printer, biocompatibility of a material, transparency, physical and mechanical properties, and durability should be considered⁷. These qualities play a crucial role in ensuring the mechanical stability and long-term efficacy of 3D printed stents in vivo. Another factor to consider is printing speed and resolution, which are vital for mass manufacturing of patient-specific treatment solutions⁸. Indeed, to produce medically viable stents, strut thickness should not exceed 100 $\mu$m, with surface roughness being as low as possible, ideally at the sub-$\mu$m level^9,10. Such requirements immediately eliminate numerous 3D printing techniques, severely limiting their selection. The mentioned properties are important because, ideally, if the 3D printing technology of choice becomes developed enough, it should allow the production of stents for each patient individually according to their needs, making it one step closer to individualized medicine. Apart from complex technical characteristics, there are also questions regarding the efficiency and standardization of such stents. Since there are various printing technology types¹¹, it is challenging to have streamlined and well-organized principles for such 3D printed structure certification and adoption. Price is also a concern because 3D printing technologies that are precise enough to accommodate such finite feature sizes needed for stents are usually rather slow, making structures manufactured relatively expensive. Regarding other non-additive manufacturing methods, femtosecond laser cutting can also be considered an attractive option. Femtosecond laser processing is renowned for its capability to control the thermal aspect of processing very well, resulting in superb cutting quality¹². Additionally, recent works showed that this can be combined with very high throughput if femtosecond bursts are employed¹³. During stent manufacturing, the capability to process both absorbing¹⁴ and transparent¹⁵ materials means a very high selection of possible materials, including those popular in medicine. Despite all these advantages, the necessity of using tubular stock and developing very high precision micro-lathe setups for processing makes such an approach somewhat more complicated and less flexible than the vat-based additive manufacturing technique. Also, due to the subtractive nature of the approach, the expected amount of technological waste is larger¹⁶, contributing to higher prices and a more significant environmental impact. Finally, the inherent demand to use cylindrical symmetry limits the exact 3D architecture of a stent. Although latter is not critical in the case of current stent production where cylindrical symmetry is physiologically preferred. Therefore, while subtractive femtosecond-based stent manufacturing is viable for producing some stent designs, the additive option brings a new level of flexibility and growth potential to the field. As a result, due to these complex requirements and interplay between them, commercialization efforts of these structure are exceptionally challenging. This follows the general tendency of the relatively slow application of 3D printing in the medical field despite the potential advantages it can bring.

The answer to the challenges of size, flexibility in manufacturing and 3D structural fidelity can be resolved with application of two-photon polymerization (2PP). It is an additive manufacturing technique based on nonlinear interaction between femtosecond laser light and photo-active resin¹⁷. Due to the specificity of such interaction, the technology achieves very high sub-$\mu$m resolution/feature size and allows the creation of highly complex 3D geometries. These capabilities allowed 2PP to be employed in a multitude of applications, including but not limited to photonics^18,19, microoptics^20,21, microfluidics^22,23, organ-on-chip^24,25 or micromechanics^26,27. Regarding the field of biomedicine, 2PP also showed substantial capabilities there. It was shown by numerous groups that materials applicable in 2PP can be biocompatible²⁸. Thus, by printing structures with features smaller than cell size, or bigger than single cells, it is possible to produce 3D objects for fundamental cell research²⁹, cell cultivation³⁰, or regenerative medicine³¹. Therefore, it seems natural to apply 2PP for stent manufacturing as well. However, the mass application of 2PP for stent printing process needs sufficient throughput. Here, classical 2PP struggles. It is a point-by-point printing technique with naturally limited printing speed. Thus, to make 2PP a viable 3D printing technique for mass manufacturing patient-specific stents, its throughput must be increased without a significant loss of resolution.

In this pilot study, we present 3D-printed vascular stents using 2PP combined with spatial beam-shaping and horizontal voxel elongation techniques. Work logic is presented in Fig. 1. The goal is to produce extremely precise stents with complex 3D geometry at throughput, which would warrant its adoption for mass production of patient-specific structures. As a case study, we chose one of the current capability gaps in the field - stenting of very small (sub-1 mm) blood vessels- a task currently not addressed adequately by nowadays used stents³². Therefore, our stent prototypes were 0.7 mm and 0.9 mm in diameter and 5 mm in length. Stents were characterized using optical microscopy to evaluate their structural fidelity and surface properties. Their mechanical properties were also tested quantitatively and qualitatively, showing that stent properties can be pre-calculated by employing advanced in silico calculation methods. After successful mechanical testing, in vivo stent biocompatibility experiments on small rodents were conducted, giving initial information on how a living body reacts to 2PP printed structures of this kind.

Figure 1

The schematics present the study workflow. First, CAD models of stents were generated. The printing process followed, with the goal of preparing structures for mechanical and biological testing. Additionally, cylinder samples were prepared from the same material as the stents. Mechanical testing was performed in two steps - experimental material characterization and in silico modeling. Finally, stents were validated using rats in vivo.

Results

Printing of stents

We began our work by formulating technical specifications for stents of interest. As mentioned, metal or drug-eluting stents are already commercially available. These stents are most commonly used to treat the coronary vessels after CVDs. The stent design developed in this study was based on an open-cell architecture. Open-cell stents provide improved conformability to complex vascular anatomy, although some research has indicated that their greater strut spacing and flexibility may lead to increased arterial wall calcification. More critically, clinical evidence has shown that closed-cell stents are linked to a five-fold increased risk of in-hospital stroke and death after carotid artery stenting ³³. This is most likely because of the stents’ decreased flexibility and inadequate adaptability to irregular vessels. In order to provide mechanical performance with enhanced security outcomes in high-risk vascular circumstances, we decided to develop an open-cell stent design. Nevertheless, there is a capability gap related to treating small coronary vessels with stent implantation³⁴. For this reason, we concentrated on stents that cannot be produced with other manufacturing techniques. Namely, the parameters we chose for our stents were 5 mm tall, sub-1 mm diameter, and with strut thickness of 50 $\mu$m. Here, the combination of 3D resolution and speed of manufacturing is the key.

Fabrication throughput is crucial to making such a process feasible for mass adoption and commercialization. However, 2PP is notorious for being a precise but relatively slow 3D printing method. It is generally accepted that to achieve higher throughput objectives with lover numerical aperture (NA) should be used²⁷. The problem is that using objectives with NA < 0.4 voxel elongates in the Z axis much more than in the XY plane, with the voxel aspect ratio approaching 1:30 and more³⁵. This creates a risk that printing stents with complex features might result in a structure with no 3D geometry. In other words, instead of a 3D stent, a tube would be produced. At the same time, for NA > 0.4, printing times for mm-sized objects can be unacceptably long- twelve hours and more for a single structure. Thus, a custom solution was needed to fabricate such stents in a reasonable amount of time (i.e., no more than a few hours for a stent).

A possible solution to account for throughput challenge could be spatial light modulator (SLM) based beam shaping³⁶. Over the years SLM was applied in tandem with 2PP and general laser material processing numerous times either to improve throughput or better control processing parameters^37,38. In the former case, the most numerous studies were about either multiplexing 2PP process via formation of focal point arrays³⁹ or by creating structures in layer-by-layer fashion (somewhat similar to DLP)⁴⁰. Producing entire object in single exposure was also demonstrated⁴¹. First approach is simple to execute, but limits effective working area to the distance between focal points. Thus it is only suitable for period structures⁴² or for mass manufacturing of objects smaller than the distance between focal points⁴³. Single exposure also has numerous challenges. Necessity to generate complex phase masks puts substantial strain on computing hardware and can be limiting factor⁴⁴. Also, single exposure cannot be bigger than working field of the objective, meaning that for bigger structures stitching should be employed, which is undesirable due to mechanical and optical defects it introduces⁴⁵. Additionally, relatively low laser induced damage threshold of commercial SLMs dictate necessity to use high concentrations of photoinitiator in the material, which is not ideal, especially for bio applications. Finally, commercial SLMs have refresh rate of 60-120 Hz, meaning that time between layer formation is relatively significant negating some of the throughput gain. Dynamic mirror device (DMD) based 2PP systems solve some of these challenges⁴⁶, yet still require very powerful lasers, high photoinitiator concentrations and introduce additional challenges related to intensity contrast in the focal plane. Thus, unsurprisingly, applying of beam shaping to expedite 2PP manufacturing process is still mainly limited to academic domain.

To truly exploit the capabilities of SLM for efficient stent printing we chose a novel approach of horizontal voxel elongation. The principle of this approach applied for stent production is illustrated in Fig. 2. SLM is used to extend focal point in one horizontal direction while keeping intensity distribution close to standard Gaussian shape in perpendicular horizontal direction. Voxel expansion in Z direction due to such elongation is also minimal, maintaining possibility to produce complex 3D structures in vertical domain. Then standard, hatching-based layer-by-layer printing is performed. After each layer the scanning direction and focal point is rotated by $90^o$ allowing for mechanically stronger print. Advantages of such solution lays in simplicity. First, standard printing algorithms can be used. Phase masks needs to be calculated once without the need to use advanced computer hardware capable of fast and precise calculations of multiple phase masks. The single generated phase mask can just be rotated around to position the voxel elongation perpendicularly to the translation direction for maximum efficiency. Voxel elongation in one horizontal direction also does not require massive increase in laser power to maintain sufficient intensity for polymerization - just around as much as the voxel was elongated in comparison standard focusing. In our case we chose 10 $\mu$m elongation, as it represented the best compromise between throughput increase and requirement to produce 50 $\mu$m thick stent struts. In direction perpendicular to the elongation voxel size/resolution width remained at 1 $\mu$m, in Z direction - around 20 $\mu$m. Such feature size/resolution is very close to what could be expected from conventional 2PP²⁷. Finally, as it is just an enhancement for regular slicing and hatching it also allow stitch-free fabrication, enabling favorable optical and mechanical properties of final structures.

Figure 2

Schematics of application of elongated voxels for high throughput stent manufacturing. (a) and (b) shows the progression of the process. Expansion of voxel to the width of 10 $\mu \text {m}$ allows to reduce the number of translations needed to manufacture a single layer. At the same time, a regular hatching and slicing algorithm is applied, and the structure is printed in a layer-by-layer fashion. Therefore, printing is exceptionally fast but realized in a fundamentally straightforward fashion.

One of the advantages of the voxel elongation strategy is it’s flexibility and throughput. With it, single 5 mm long and 0.7 mm and 0.9 mm radius stent can be produced in less than 2 hours. The only required post processing step after printing is development, which was fast (up to 5 minutes, including sample handling) and as simple as submerging sample to the developer batch. Thus, it had negligible impact on the throughput. The study’s goal, application field, and planned insertion method dictated the stent design’s rationale. To test the capability to 3D print stents using voxel elongation, a complex 3D shape with very thin (50 $\upmu$m width) struts with an overall diameter of less than 1 mm was chosen. Additionally, the chosen architecture minimized stent surface area, which is important to minimize the risk of blood clots and overall volume, assuring that if there are any photoinitiator/monomer left due to non-100% polymerization degree⁴⁷, it has minimal chances of causing any adverse effects. As mentioned already, it should be quite suitable for small vasculature treatment. Finally, a straight, thin stent with reinforced rings was preferred due to the chosen endovascular, open-type stent implantation approach for controlled in vivo testing in an animal model. Of course, different architectures can be used and easily adapted for printing on-demand in the future due to the inherent flexibility of additive manufacturing employed in this work. To sum up the planned in vivo testing, the stent design was optimized to be inserted using an intravenous catheter. For this reason, diameters of 0.7 mm and 0.9 mm were chosen, as they correspond to the 26G catheter size. Having these dimensions, there were no challenges in printing 30 stents (15 of each type) to facilitate full needs of biological experiments while ensuring that additional stents would be available if unforeseen events occurred. To demonstrate the flexibility and scalability of this approach, using the same 10 $\mu$m wide voxel stents of 10 mm lenght and 1 mm radius were produced. The pipeline for such experiment was exceptionally simple, as scaling required just to resize the 3D model in the printer control setup. As an additional test for the approach we also printed a stent that was 50 mm tall with a 2.5 mm radius. For it a wider 25 $\mu$m voxel was used. No additional tuning of the system was needed, once again showing how easy it is to switch between usage of different beam shaped voxels. 50 mm tall stent was printed in around eight hours and could be used not only for blood vessels but, for example, for stenting the urinary tract. All three stents side by side are shown in Fig. 3. Overall, from a purely printing standpoint, elongated voxel-based 2PP basically trivializes the printing of stents with single features from tens of $\mu$m to overall size towards several cm. Additionally, performed SEM analysis Fig. 3 (b) showed that the surface roughness of stents is around 1 $\upmu$m RMS, which is a tolerable value for such a structure. Finally, it is important to mention that the wettability properties of such structures are also important. However, the ideal sample for such characterization should be a flat plate⁴⁸. Thus, due to the complex 3D stent architecture, such measurement could not be performed and/or yield proper, correct contact angle values. As a result, contact angle measurement was not performed in this study.

Figure 3

Stents of various sizes. (a) - from left to right, the image shows vascular stents printed using voxel elongation technology, with heights of 50 mm, 10 mm, and 5 mm, (b) - 0.7 mm diameter stent SEM micrograph, (c) - 0.7 mm diameter stent SEM micrograph visualizing the 50  $\mu$m strut thickness.

Mechanical testing

The mechanical testing consists of two parts, in the first part, tests for visual assessment of the mechanical effect were performed, in the second part, compression tests of the workpiece were performed, during which force and displacement were measured. The theoretical study was intended to be validated by the results of the second part of the mechanical testing.

The initial tests involved visual assessment of the samples. Printing experiments showed that stents of relevant sizes can be produced using 2PP and voxel elongation. However, they must exhibit appropriate mechanical and biological properties. Thus, next step in this study was to determine mechanical properties of the printed 3D objects. Mechanical experiments started by qualitatively testing stent resilience for the most common types of mechanical stress present during the insertion procedure and while in the body - bending, crushing, and pulling. For it, the stent was placed under an optical microscope, and the polymeric probe was used for bending and crushing experiments while pulling was performed with metal tweezers. Stents showed strong mechanical resilience during the experiment, maintaining their structural integrity throughout the process [Fig. 4]. While minor cracks could be observed, no spalling occurred, showing that the material is relatively flexible and, at the same time, tends to return to its original configuration even after sustaining some mechanical damage. These results are essential for medical applications where spalling within blood vessels would create hazardous particles in the bloodstream.

Cylindrical specimen compression is tested mechanically and theoretical modeling studies are included in the second part. Presented qualitative testing, while promising, only showed a general tendency of the 3D printed structure. Thus, quantitative testing was needed. It was performed in two main steps. First, the material itself was tested. Then stent mechanical properties were deduced. Here, using material properties from a general datasheet might not be applicable, as 2PP printing parameters, especially exposure, have a tremendous impact on the mechanical properties of the polymer^49,50. Therefore, every printed structure should be characterized independently. In our case, to test the material with which stents were printed we produced cylinders that were 3 mm in diameter and 1 mm in height Fig. 5 (a). We chose such a configuration because such cylindrical specimens are commonly used in experimental studies of the compressive properties of materials. The quantitative parameters tested were tensile and compressive forces induced by the universal digital force gauge. The compression was recorded during the experiment at 10 $\mu$m intervals from 0 to 150  $\mu$m, resulting in 16 measurement points. The compressive load was read individually at each step (every 10 $\mu$m), and the maximum value obtained was 17.5 N. The curve of the specimen’s strain and the compressive load variation is shown in Fig. 5 (b). It shows that the strain of the cylindrical specimen is minor and has an approximately linear dependence on the compressive force. Compression experiments on the samples have shown that the material has hyperelastic properties. This aligns well with qualitative testing, which showed stents tending to return to their original shape after deformation was exerted on them.

Figure 4

Tests to evaluate samples visually. Still images taken out of the video observe mechanical manipulation of a 3D printed 5 mm tall stent. Bending, crushing, and pulling were attempted using either plastic probes or metal tweezers. Although all tests were performed in sequence and mechanical defects, such as cracks, could accumulate, however, stent still maintained its structural integrity. Subsequent tests with other stents showed the result to be highly repeatable.

Quantitative characterization of stents followed. Here, we were faced with a challenge. Due to their relatively small size, the mechanical characteristics of 3D-printed stents were difficult to measure using classical laboratory measuring devices. To bypass this limitation without creating a mechanical testing procedure from scratch, we instead chose the in silico approach, i.e., use mathematical modeling to determine the precise mechanical properties of the structures. Therefore, the next step of this study was to use experimental measurements acquired using cylindrical specimens, create virtual material for calculations, and then apply the properties of the virtual material found to study the stents’ mechanical properties. As a material model, the Mooney-Rivlin two-parameter hyperelastic material model with a nearly incompressible version was chosen when the isochoric strain energy density was determined. Eq. 1 is written in terms of the two isochoric invariants of the elastic right Cauchy–Green deformation tensors $I_1(\overline{C}_{el})$ and $I_2(\overline{C}_{el})$, and the elastic volume ratio $J_{el}$:

$$\begin{aligned} W_{\text {iso}} = C_{10}(\overline{I}_1 - 3) + C_{01}(\overline{I}_2 - 3) + \frac{1}{2} \kappa (J_{\text {el}} - 1)^2. \end{aligned}$$

(1)

The material parameters $C_{10}$ and $C_{01}$ are related to the Lamé parameter (shear modulus) $\mu = 2(C_{10} + C_{01})$. Volumetric strain energy density – quadratic, the pressure formulation was mixed, and the Bulk modulus $\kappa$ was 100 times the equivalent shear modulus. The Nitsche contact model in incomplete formulation was used in the first experiment. Nitsche’s method for contact is an extension of the general method to impose Dirichlet conditions in a weak sense without having to add Lagrange multipliers. Moreover, a small static coefficient of friction was used in the contact zone to stabilize against lateral shear. The specimen was placed between two planes, and the displacement of the upper plane was determined based on the results of a real experiment. One of the numerical experiment results is shown in Fig. 5 (d). The values of the found parameters $C_{10}$ and $C_{01}$ are shown at the bottom of the right photo in the Fig. 5 (d). The force acting on the specimen in the contact zone is shown in Fig. 5 (c) as the reaction force vs strain. In the Fig. 5 (c) the red rectangles represents experimental results presented as a Least-Squares fit curve of all five experimental measurements, and the blue curve shows the reaction force of the numerical experiment. It can be stated that the characteristics of the material used in the real experiment are very close to the characteristics obtained from parametric modeling. These properties will be applied to the resulting material in further numerical experiments.

Figure 5

Numerical experiment results for specimen compression and experimental and numerical results comparison.: (a) specimen photo, (b) Strain vs. compression force curve for a cylindrical specimen, (c) Experimental and numerical results comparison, (d) Results of the numerical experiment.

The second phase of the numerical experiment consisted of creating a virtual model of the stent and performing four virtual experiments with it. The first ones are compression and tensile tests, the second a bending test, and the third a local compression test (with tweezers). The final, fourth experiment involves the interaction of the stent with the vein and blood flow. The simulation was performed in the time domain, in 2.4 seconds, during three cardiac rhythm periods. The reaction force and elastic strain energy were selected as the resulting parameters in all experiments. In compression (or any other type of stress), every material suffers some deformation that changes the average relative position of its atoms and molecules. As a result of reversible deformation, when the compressive forces disappear, a reaction force is generated that opposes the compressive forces and may eventually balance them. Thus, we can determine the force of the external load on the sample for any form of external action. For a hyperelastic material, the strain energy density function is the fundamental quantity from which stresses are derived. The hyperelastic model determines the form of the strain energy density function. A global variable containing the integration and summation of all elastic strain energy density contributions at the physical interface is used as the resultant. Fig. 6 shows the total reaction force and total elastic strain energy of compression (a) and tension (b) simulations of the stent, respectively. It can be seen that the properties of the stent under loading in the longitudinal direction are pretty elastic. Its lattice geometry partly determines this. With lateral loading on the specimen, as demonstrated in Fig. 6 (c), we see an increase in the reaction force. This indicates that the specimen has a stiffer structure when bent. The third numerical experiment simulated the local compression of the specimen with tweezers. As can be seen from the results depicted in Fig. 6 (d), the stiffness of the structure in this case is even greater. Finally, the last simulation involving the computational-fluid dynamic model of blood structure vein-stent interaction is presented in Fig. 6 (e). The length of the vein is 9 mm, a hyperelastic material Mooney-Rivlin, two parameters model for the vein was used, and the blood is calculated as a Newtonian fluid. Fig. 6 (f) shows the maximum blood velocity in the center of a vein in the presence of a stent. It shows that there is some stepwise increase and decrease of blood velocity in the channel at some moments. Based on these calculations, it can be concluded that the stent design has good prospects for medical applications. The elastic properties of the design in the longitudinal direction and good stiffness in the radial and shear directions make it a good candidate for bypass surgery.

For all simulations, the following boundary conditions and assumptions were established: fluid-structure interaction (FSI) with fully coupled interfaces, laminar flow with incompressible fluid properties, non-suppressed backflow at the inlet and outlet, no-slip boundary conditions of the wall, pressure boundary conditions at the inlet and the outlet, deforming domain and Yeoh mesh adaption method with smoothing factor C₂ = 10. Viscoelasticity was considered using generalized Maxwell model, the used solver method is PARDISO, and the implicit solver method was used with a time step of 0.01 s, following the generalized alpha model. The pressure was considered as the boundary condition for the inlet and outlet. The plot of pressure distribution at the vein inlet was taken from actual medical measurements. For both the inlet and outlet, this option allowed us to obtain the normal stress, which is approximately equal to the pressure in most cases. Also, the outlet’s tangential stress component was set to zero. If the reference pressure, defined at the physics interface level, is equal to 0, the value of the pressure p₀, at the boundary is the absolute pressure. Otherwise, p₀ is the relative pressure at the boundary. The suppress backflow option adjusts the inlet pressure locally to prevent fluid from exiting the domain through the boundary. Considering we used the non-suppressed backflow option, the inlet boundary could become an outlet depending on the pressure field in the rest of the domain. The parameters of the interaction of the vein wall with the blood flow and the boundary conditions were taken from our previous studies and based on information from other authors.

Only the elasticity of the vein was considered in this simulation. The calcification of the vein was not taken into account because, from the point of view of boundary conditions, it is a non-deformable vein wall. However, any model has its limitations. In our model, we did not consider the adhesion and deadhesion of the stent with the vein wall and the contact interaction between the stent and the vein wall. In the real system, some drift of the stent is possible. We assumed it had good in situ stability since it has a reticulated shape. Based on this assumption, we did not consider using the contact model. Another assumption was that blood is modeled as a Newtonian fluid, although it may change its physical properties at different flow stages. At low blood flow velocities in these veins (up to 14 cm/sec), the influence of non-Newtonian fluid is minimal, so we chose the Newtonian blood model for our calculation.

From our perspective, there are potential other directions for numerical calculation in this field. First, muscle contractions must be considered in order to conduct a more detailed investigation of the influence of additional factors on the interaction between the stent and vein wall pair. Muscle contractions cause the veins to squeeze, which in turn pushes blood towards the heart. This is crucial in the legs, where blood needs to travel against gravity. Moreover, in this instance, contact interaction may need to be considered due to additional deformations of the vein elastic wall. The next one concerns the impact of calcification of the vein wall in a specific location. Future research could focus on evaluating contact interactions in this area or beyond, including adhesion and de-adhesion. It seems that contact interaction will significantly influence the results in this instance. To conclude this chapter, we would like to draw attention to certain limitations that are used in these numerical calculations. The Table 1 summarizes the key assumptions used in the numerical experiment in a brief manner.

Figure 6

The second experimental phase of numerical testing. (a) - compression test, (b) - tension test, (c) - bending test, (d) - local compression test, (e) - computational-fluid dynamic model of blood structure, (f) - maximum blood velocity in the center of a vein in the presence of a stent.

Table 1 Comparison of real boundary conditions with assumptions and their impact on accuracy.

Biological testing

Successful mechanical testing and modeling showed promise in 3D-printed stents for bio-compatibility testing. The goal was to determine if such small stents could be used for blood vessel treatment. The Wistar rats were chosen as model animals because experiments with cell cultures using Ormoclear FX material have been done and were not cytotoxic and did not affect the cell-to-cell junctions, which is vital to tissue growth^51,52. Indeed, this material was tested extensively in biological settings throughout the last two decades ^53,54,55,56, including in vitro testing and including cardiovascular scaffold research in vivo using rats. Therefore, we evaluated whether handling and stenting procedures are possible with such small stents. The standard medical pipeline was adapted to make this stenting procedure feasible without complicating the procedure. It is important, because capitalizing on existing equipment and practices would allow a seamless and fast introduction to medical practice.

Animal testing was performed to evaluate the stenting procedure and its efficacy. The stents were successfully inserted into the carotid artery of 10 male and female Wistar rats and stabilized with a surgical suture. The stenting procedure used an intravenous catheter size 26G, proving that even such small stents can be applied with currently available medical assets and procedures. All test rats survived 7 days, during which time they did not show any changes from the typical behavior characteristic of the species, showing that the material of stents was not inherently toxic. This aligns well with previous knowledge from literature where hybrid organic-inorganic photo-polymers were shown on numerous occasions to be bio-compatible^57,58,59,60. The main question, therefore, was how blood vessels fared with 3D-printed stent prototypes after the procedure. Subsequently, the rats were euthanized, and then their artery samples were extracted and prepared for histology analysis [Fig. 7]. The result is promising and shows no immediate challenges in continued design and material testing. At the same time, further investigations focusing on the immune response profile are necessary to determine the factors that impact inflammation. More biologically sound materials, such as hydrogels and low surface roughness structures, should be tested in the future - both of which are available while employing 2PP-based 3D printing^61,62. The final goal would be to use drug-eluting biodegradable materials for complete blood vessel regeneration. While additional polymer development should be carried out, initial experiments of material-releasing 2PP-processable bio-materials were already demonstrated⁶³. It is important to note that current results reflect a very early stage of pilot experiments. Future experiments combining in vitro testing of more materials, especially targeting interaction with fibroblasts and endothelial cells alongside long-term (6 months and more) in vivo experiments will be performed in future research. Apart from that the next phase of experiments will prioritize the optimization of stent design and material composition. Auxetic structures, known for their remarkable mechanical strength and flexibility, can be used to optimize stent design. These designs allow the stent to expand during deployment, matching the patient’s vascular shape. These optimizations could improve biocompatibility and endothelial integration and include more comprehensive investigations encompassing both the systemic immune response–characterized by markers such as IL-6, TNF-$\alpha$, IL-8 (CXCL8), and MCP-1 (CCL2)–and, more specifically, the endothelial and vascular-specific response. A subsequent step should involve monitoring changes in the concentrations of VCAM-1, ICAM-1, E-selectin, vWF, endothelin-1, and miR-126/miR-92a, among other relevant biomarkers, to provide a detailed understanding of local vascular inflammation and healing dynamics.

Figure 7

Representative photographs of cross-sections of carotid arteries obtained from Wistar lineage rats after a 7-day period. In the control group (a), all major layers are visible, and their relative thickness is within normal bounds. Test groups (b), (c), and (d) exhibit vasculitis, or mild to moderate inflammation. The architecture of the vessel wall is unaffected, but there is a minor infiltration of granulocytes and lymphocytes (b). There is a reduction in the boundaries between the tunica externa and tunica media, as well as a more intense infiltration of immune system cells (c). More intense damage to the integrity of the vessel wall. The endothelial layer is affected (d). The conventional hematoxylin-eosin (H&E) technique was used to stain the sections.

Discussion

As already discussed, employing 3D printing for stent manufacturing brings numerous advantages regarding adaptability to specific cases or architectures. Enhanced throughput 2PP presented in this work expands these capabilities even further, as it allows to reach $\sim$ $\mu$m level resolution/feature size while maintaining the possibility to print stents in tens of minutes as the technology evolves. While this is an excellent technology to produce stents for relatively small vasculature reconstruction, it also enables employing complex 3D geometries towards mechanically programmable stents. 2PP was already a superb tool for producing complex 3D architectures that can act as mechanical meta-materials with programmable functional mechanical properties. Numerous times, 2PP performed as an excellent tool for such structure manufacturing^64,65. Modern metal stents use such architectures to some extent, but they are limited by manufacturing technologies with inherent cylindrical symmetry and available sub-mm processing resolution. With 2PP-based 3D printing, both of these limitations can be bypassed. Then, it is possible to envision mm-sized stents with single structural 3D features going down to $\mu$m size with built-in application-specific mechanical properties. These can precisely control the Poisson index so stents can be introduced into the body with minimal invasiveness and collateral damage. Another option would be stents with gradient mechanical properties that can support specific parts of vasculature in patient-specific ways. Many capabilities should be explored in the future. Additionally, due to the simplicity and speed of the printing process, it is possible to test the mechanical properties of the materials very quickly, allowing fast development of numerical models. It is important to note that for every material and parameter set, such experiments must be performed separately and from scratch because the mechanical properties of 2PP processed materials heavily depend on the printing settings^49,50. Afterward, it is possible to perform most of the experimentation in silico Indeed, there is a clear necessity of expediting 3D structure testing in the field. One way or another, potential mechanical failures, especially over long time, is one of the main worries in the field of 3D printed stents. In this work we show that in silico can somewhat simplify the pipeline and make it cheaper, allowing to do mass screening of stent designs even before printing them for experimentation. That way time and resources needed for long-term pre-clinical trials is significantly reduced and only the best designs can be tested in in vivo experimentation. In this study, we demonstrated the short-term biocompatibility of the implanted stents, with no significant adverse effects observed within the seven-day evaluation period. These findings suggest that the stents are well-tolerated in the acute phase, providing an important first step toward their further development. However, we acknowledge that this study does not assess the long-term functional efficacy or safety of the stents. Future studies should include extended implantation periods to evaluate endothelialization, mechanical integrity, and potential inflammatory or thrombotic responses over time. Therefore, this work can be considered a blueprint for a pipeline in further 2PP 3D printed stent development.

Currently, the techniques used for stent printing tend to be slow, which makes them unsuitable for mass production, and they need more fidelity for more complex architectures. Therefore, 2PP with elongated voxel technology could solve problems related to 3D bioprinted stents. One of the main advantages of elongated voxel technology is that it enables printing large structures with feature sizes up to 1 $\mu \text {m}$. It was observed that the vascular stent’s architecture plays a crucial role in the treatment efficacy⁶⁶. For example, the efficacy of drug-eluting stents depends not only on the drug used but on the architecture of the stent⁶⁷. Currently, there are various stent designs in the market, although, using elongated voxel technology, we are moving one step closer to printing stents according to the patient’s needs. The patient-specific stents can achieve the identical shape of the patient’s blood vessel after placement and expansion⁶⁸. This minimizes the risk of stent restenosis, which is one of the primary reasons stented vessels re-narrow after treatment⁶⁹. While printing patient-specific stents could be challenging, elongated voxel could speed up printing. It would allow printing complex architecture stents while maintaining speed by printing small and precise features with less elongated voxel and more bulk features with more elongated voxels. In the future, the 2PP combined with elongated voxel technology has the potential to print even larger patient-specific stents, which could be used for larger vasculature treatment. However, the presented process is just a pilot test designed to show how much application of elongated voxel can expedite stent production with 2PP. While the current 2-hour manufacturing time is not entirely suitable for industrial mass production, further scale-up is possible. The most obvious path would be the application of Galvo scanners, which should allow an increase in translation velocity to 1 m/s. SLM-based beam shaping was shown to be compatible with scanners for 2PP. Furthermore, beam multiplexing should be employed if multiple of the same stents are needed, thus parallelizing the process. At the same time findings related to voxel elongation lay the foundation for further dramatic increase of 2PP throughput. Indeed, the general regard in the additive manufacturing field is that 2PP is a precise but relatively slow processing technique that cannot be easily upscaled. This work showed that with a dedicated, beam-shaping-based optical chain, it is possible to expedite the printing throughput by at least one order of magnitude, allowing produce stents at a base price of around 500 Euros for one structure. More advances to are expected in the future with full technology calibration and a better understanding of how to generate phase masks dynamically for architecture-specific voxel shape manipulation. Then, applying all mentioned techniques, we predict that printing throughput could be increased by more than one order of magnitude, allowing to manufacture 5 to 10 stents per hour, subsequently reducing the price by 50% to 75%, depending on the exact material and stent architecture.

Ormoclear FX was used in our work for stent printing. This material was chosen because of its good printability properties using 2PP. It was important to use well-known materials for testing. Then, while testing voxel elongation for functional structure printing, we can be sure that it worked not due to some specific material properties but because the approach itself was sound. Additionally, since it was the first time vascular stents were produced using 2PP, it was logical to choose a material that had been successfully used to print other structures, such as scaffolds⁵². Ormoclear FX material is mainly used in producing microfluidic systems or optics⁷⁰. OrmoClear FX is a commercially available UV-curable, inorganic-organic hybrid polymer material based on sol-gel synthesis. Due to the inorganic backbone from SiO moieties, a specific combination of material features, such as optical clarity and temperature stability, is achieved. The material is intentionally designed for the wafer-scale production of micro-optical components, such as micro-lenses or diffractive optical elements. However, its resistance to many chemicals, solvents, and aqueous solutions makes it a good candidate for life-science applications. As it is a commercial material, its general properties can be found in a dedicated datasheet prepared by the manufacturer⁷⁰. Stents that pose shape memory properties are essential to their applicability in the clinical field. In the future, these properties can allow minimally invasive implantation, minimizing the risk of vessel injury during the deployment procedure⁷¹. After implantation, the stent could better withstand the mechanical forces present in the vessel, which reduces the risk of stent migration. These properties are tied together with the material used and printed architecture. Thus, it is vital to test stents’ flexibility and mechanical stability. Stents from Ormoclear FX material demonstrated good mechanical stability in stressful environments and did not break during or after the stenting process in the animal model. However, we plan to use other materials for future experiments to improve our results further. The need for other, mode advanced bio-materials arises because OrmoClear FX is not ideal for biological applications, especially BSV stents, as it is not biodegradable.

As bioresorbable vascular stents (BVS) start to gain more attention in the biomedical field, the choice of materials becomes crucial⁷². 2PP, combined with voxel elongation technology, could open new avenues for producing BVS stents. Therefore, one of the challenges is the material selected for their production using 2PP. Currently used materials for BVS production include bioresorbable polymers such as poly-l-lactic acid (PLLA) and poly-d,l-lactic acid (PDLLA)⁷³ being the most popular. Among metals, magnesium, iron, or zinc-based BVS material is most commonly used⁷⁴. The ideal candidates to produce BVS using 3D bioprinting are polycaprolactone (PCL) and polylactic acid (PLA) because they possess good mechanical properties, are biodegradable and biocompatible^75,76. Although, some challenges are included- BVS produced from polymeric materials fail to provide sufficient radial strength, making them inadequate for implantation. Also, some BVS that were mechanically stable collapsed after implantation⁷⁷. The use of polymeric materials for BVS production presents several challenges. These issues can be solved with rapidly evolving technology and integrating different scientific disciplines. By using 2PP and different polymers for printing, 3D-printed BVS is possible, and we are planning to work on it in the future. Regarding future challenges and research directions, the first thing to consider is that most commonly, stents already on the market are implanted using interventional angiography, where the stent is inserted into the damaged blood vessel using catheters and expanded or otherwise pushed in. The main challenge in using this type of stent for implantation will be to create or adapt implantation kits for angiography operations. For open-type implantations, it is necessary to adapt the surface of the 3D-printed stents so that they do not disintegrate during suturing when the needle pierces its layers. Overall, the work showed great advances toward stents with well-defined structural features within tens of $\upmu$m. Additionally, it was shown that stents with such precision can be manufactured in relatively short time, with possibility to improve throughput even further. Nevertheless, the question of material remains, as the final 3D printed BVS stent material should be biodegradable and release drugs as it decays in the body. Further work in the material domain is needed to realize this role.

Conculsions

We succeeded in 3D printing a vascular stent using 2PP and voxel elongation technology. Using this technology, we managed to print stents faster, thus cutting the production time by 20 times. After successful stent printing, we evaluated their mechanical characteristic by performing a three-step experimentation plan. Firstly, we evaluated the stents’ mechanical resilience qualitatively, showing that stents were resilient to environmental stress. Only minimal cracking was seen, and stents returned to their original form after deformation. Then, quantitative experimentation was performed where material properties were tested, and numerical experimentations were performed afterward. These results showed that stents had a good design and were promising tools for medical applications. After that, biological testing followed, and thirty stents were printed. The stenting procedure was performed in 10 female and male Wistar rats. After a 7-day incubation period, rats did not show any changes from the typical behavior characteristics of the species, showing that the material was not inherently toxic. This work shows 2PP’s potential to produce patient-specific stents for small vasculature stenting and offers an alternative to currently available manufacturing techniques. While the results are encouraging, further investigation into materials’ biocompatibility and durability is needed to validate these stents for medical use.

Methods

For 3D printing, a variation of Vital Light 3D (Vital3D Technologies) 2PP setup was applied. Schematic of the system is provided in Fig. 8. The light source in the system is fs laser “Biolit 2” (Litilit), emitting 80 fs pulses at a 40 MHz repetition rate. II harmonic crystal follows, converting fundamental wavelength to 522 nm radiation needed for printing. Then, the light is directed to the power attenuator LPA (Optogama), polarization rotator MRO (Optogama), and 3X beam expander (Optogama). The beam is expanded to fill the full apertures of the SLM and objective lens. SLM used in the setup is Leto (Holoeye), with a matrix size of 1920x1080 pixels. Phase masks for elongated voxels were generated using built-in Holoeye SLM pattern generator software (version 5.2). Two lenses forming a 4F system follow, with the iris being placed in the focal plane acting as a spatial filter. The dichroic mirror then directs the laser beam to the 20x 0.8 NA objective lens (Zeiss), which focuses on the sample. The sample is then submerged into a material vat and positioned in 3D. The setup also has a built-in imaging sub-system consisting of a red LED placed in the sample holder above the sample and an additional lens and CMOS camera opposite the dichroic mirror. Due to this sub-system, the printing process can be observed in situ in real-time.

Figure 8

Schematics of 3D printing 2PP setup used in the work.

Stents were produced according to the protocol prepared in the Vital3D Technologies laboratory. 3D models of stents were prepared using Autodesk Fusion 360 software (v.16.12.0.2384). The model was exported to STL format and uploaded to the printing software. For manufacturing - Ormoclear FX (Micro Resist Technology) material was used. Before fabrication, it was warmed up to room temperature and poured into the material vat. 1 mm thick substrates were used for the printing. They were placed in an inverted sample holder and submerged in the liquid resin. Inverted architecture allowed the extraction of samples with printed stent from material vat after each layer was produced. Voxel was elongated to 10 $\mu$m in the horizontal plane for fabrication, and the average laser power used for experiments - was 10 mW. Development step followed. Before the stent development, the substrate is removed from the sample holder and immersed into a pentanone bath for 1-2 minutes. Then, the slide with the stent is rinsed with the pentanone for 1- 1.5 minutes using a single-use pipette. The final development step is rinsing the stents with acetone for approximately 1-2 minutes. Stents were imaged using optical microscope MRCL700 3D Imager Pro (Microqubic) and SEM TM-3000 (Hitachi).

Stent mechanical testing was carried out in two ways - qualitative bending and later quantitative results was acquired using mathematical modeling. The qualitative measurement was executed using a Microqubic MRCL700 3D Imager Pro microscope and a plastic probe (PLA material) or stainless-steel tweezers. Both implements were operated manually. During this testing, the stent was on a glass substrate, which was further attached to the microscope using adhesive tape, so there would be no undesirable movement of the sample after a mechanical force or pull was applied, as it would have skewed testing results. Regarding quantitative in silico testing, we have to start by stressing that stents produced in this work had relatively small diameters, thus, mechanical testing was performed virtually using COMSOL Multiphysics software (version 6.3). Firstly, the stent material had to be mechanically tested, and the results were transferred to the software. For that, cylinders in diameter of 3 mm and height of 1 mm were printed using the same material, sample preparation, and development parameters as for stents. Cylinder mechanical properties were subsequently tested at Vilnius Tech University. There were three main steps in performing virtual stent testing. Firstly, the material properties are assessed, and secondly, the virtual material is created based on the mechanical test of cylinders, lastly, numerical simulations are performed. The experimental set-up for cylinder testing consisted of a universal digital force gauge for measuring tensile and compressive forces with an integrated load cell Sauter FH 100 with a maximum force range of 100 N, a tolerance of up to 0.5%, and a displacement value of 0.05 N, and a stand with a displacement gauge Sauter TVL, with a measuring range of 200 mm and a displacement value of 10 $\upmu$m. The cylindrical specimen was placed on the experimental table so that the axial lines of the specimen and the digital force gauge were aligned and the surface of the specimen was parallel to the surface of the measuring head. The bench displacement gauge was set to zero when the measuring head touched the surface of the specimen. A schematic diagram of the experimental set-up is given in [Fig. 9]. After cylinder testing, the results were uploaded to COMSOL Multiphysics software (version 6.3) and virtual material according to the parameters was created. later on, virtual stent testing was performed.

Figure 9

Schematics of setup used to perform mechanical indentation experiments of 2PP produced 3 mm diameter cylinders.

All animal procedures were done in The Lithuanian University of Health Sciences (LUHC) Biological Research Center, authorized by the State Food and Veterinary Services and all methods were reported in accordance with ARRIVE guidelines. The study used 8–12-week-old and 250-350 grams male and female Wistar rats. Wistar albinos’ rats were provided from LUHC VA Vivarium, Tilzes str. 18, Kaunas, breeders No. LT 59-101. Animals were randomly allocated to all groups. Ten rats were included in the study group, five animals were used for control analyses, and in total, 15 rats were used for survival analyses. Before the surgical procedure, animals were housed in an acclimatized room and fed a standard diet and water for 24 hours. During surgical procedures, animals were anesthetized using ketamine chloride 90 mg/kg (Ketamidor, Richter Pharma, Austria) and medetomidine chloride 1 mg/kg (Cepetor, CP-Pharma, Germany) IP injection for analgesia buprenorphine 0,05 mg/kg (Bupaq, Richter Pharma, Austria) were injected. During all procedures, animal vital parameters, SpO2, and heart rate were monitored using a Drager Vista monitor (Draeger, Germany). For study animals under anesthesia left neck side incision were done, left carotid artery were found and lifted using surgical sutures. Intravenous catheter of size 26G (BBraun, Germany) were used for polymeric stent introduction. Stents were pulled to the side, and a catheter tip was inserted inside. The catheter needle insertion was smooth and successful, for the control group, only catheter insertion and retraction were made. After the stent was inserted inside the artery, the plastic catheter part was retracted, and sterile gauze and slight pressure were applied on the artery wall in order to stop bleeding, after that, all tissue layers were closed using surgical sutures. After anesthesia, animals were left in a calm and warm environment for proper awakening and moved to cages and animal housing room. After 7 days, animals were euthanized using pentobarbital sodium 100 mg/kg (Exagon, Richter Pharma, Austria) IP injection, after animals were confirmed dead, the same incision was made in order to take artery samples. Stents before surgery were disinfected using chlorhexidine 0,02 %, which was made in a university pharmacy for 30 minutes. LUHC Biological research Center veterinary approval No. Lt-61-19-004 for animal use in research facility applied on 2024-04-10 to the Lithuanian Experimental Animal Ethics Commission of the State Food and Veterinary Service, application for animal research, the number G5-(8.3 Mr)- 17654 was assigned. Within 40 working days, permission to conduct an animal experiment was received, entitled “3D printed implants efficiency and compatibility studies - VITAL3D”, in which 30 Wistar line rats are allowed to be used, permission No. G2-279. Methods used in this study were carried out in accordance The Law on Animal Welfare and Protection of the Republic of Lithuania, implementing Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes, as last amended.