3D printed anisotropic tissue simulants with embedded fluid capsules for medical simulation and training

Abstract

Human tissues are primarily composed of collagen and elastin fiber networks that exhibit directional mechanical properties that are not replicable by conventional tissue simulants manufactured via casting. Here, we 3D print tissue simulants that incorporate anisotropic mechanical properties through the manipulation of infill voxel shape and dimensions. A mathematical model for predicting the anisotropy of single- and multimaterial structures with orthogonal infill patterns is developed. We apply this methodology to generate conformal printing toolpaths for replicating the structure and directional mechanics observed in native tissue within 3D printed tissue simulants. Further, a method to embed fluid-filled capsules within the infill structure of these tissue simulants to mimic blood is also presented. The improvements in simulation quality when using 3D printed anisotropic tissue simulants over conventional tissue simulants are demonstrated via a comparative acceptability study. These advances open avenues for the manufacture of next-generation tissue simulants with high mechanical fidelity for enhanced medical simulation and training.

Fig. 1.. Design and characterization of cellular voxel models.

(A) Typical structure of a 3D printed anisotropic cricothyrotomy puck for surgical training. The skin layers contain an anisotropic infill pattern sandwiched by top and bottom cover layers. (a) A typical infill pattern that generates mechanical anisotropy. (b) The simplest voxel structure with variables labeled. (B) Image of a cricothyrotomy puck mounted on an advanced airway trainer (CREST Lab, University of Washington) for training emergency cricothyrotomies. (C) Microscope images of the cross section of the infill pattern along orthogonal directions (top). Microscope images of the infill patterns as seen from above (bottom) (scale bar, 500 μm). (D) Variation of anisotropy with underlying infill parameters. The parameter space is reduced by setting the spacing between lines across successive layers to be equal. (E) Range of mechanical anisotropies that are observed in the human body (bottom) and achievable using various material combinations (top). (F) Tensile testing plots of 3D printed anisotropic cruciform-shaped samples in the axial and transverse directions (n = 3, error bars indicate SD). These samples were printed with a single material (i.e., $E_1=E_2$ ). The plot shows that different values of mechanical anisotropy can be achieved by simply modifying the underlying voxel dimensions. (G) Plot showing the correspondence between predicted and experimental values of mechanical anisotropy of infill structures 3D printed with two materials along alternate directions (n = 3, error bars indicate SD). To reduce dimensionality, the spacing between print lines was kept constant and the ratio of print heights between the stiffer and softer materials were varied across samples.

Fig. 2.. Material selection and optimization for 3D printing cellular structures.

(A) The variation in the elastic moduli of cured mixtures of PlatSil Deadener (PD) and room temperature vulcanizing silicone (RTVS) with a fixed amount of fumed silica (FS) (n = 3, error bars indicate SD). The elastic modulus can be adjusted by two orders of magnitude. (B) The variation in the elastic moduli of cured mixtures of a fixed ratio of PD and RTVS with varying amounts of FS (n = 3, error bars indicate SD). The elastic modulus does not vary substantially. (C) The variation in the storage and loss moduli of uncured PD/RTVS/FS inks. By adjusting the amounts of PD and FS, we can control ink rheology. For instance, inks PD:RTVS = 0.0 and FS = 0, PD:RTVS = 0.5 and FS = 2, and PD:RTVS = 1.0 and FS = 3 show similar rheology. (D) Variation in print-line width of deposited material spanning gaps in the previous layer due to cohesive forces as seen through a microscope (scale bar, 500 μm). (E) Micrograph of a singular spanning print-line with measured dimensions labeled (scale bar, 250 μm). (F) The relative reduction in print-line width at different relative print heights and print-line spacings for a particular ink. The print-line width reduction is particularly sensitive to a reduction in print height. A cutting plane at $w_1/d$ = 0.5 is also shown. (G) Contour plots of $w_1/d$ = 0.5 plotted for inks with different storage moduli. Increasing storage moduli results in lowered bounds for print height and greater obtainable ranges of mechanical anisotropy. All samples printed with an 18-GA nozzle. (H) Contour plots of $w_1/d$ = 0.5 plotted at different values of d. Increasing nozzle diameter results in lowered bounds for print height and greater obtainable ranges of mechanical anisotropy. All samples printed with PD/RTVS = 0.0 and FS = 0 ink.

Fig. 3.. Incorporation of directional anisotropy in nonplanar tissue simulants.

(A) Deposition accuracy on inclined surfaces can be improved by adjusting nozzle position relative to substrate in accordance with ink rheology, angle of inclination and gradient direction. (B) Incorporation of nozzle adjustments led to improved accuracy of deposition on inclined surfaces and improved the accuracy of measured values of anisotropy to predicted values (n = 3). (C) The variation in mechanical anisotropy when the relative angle between the direction of maximum anisotropy and the measurement axes is varied (n = 3). The measured values of anisotropy closely match the values of anisotropy predicted by the mathematical model. (D) Process of 3D printing a tissue simulant in accordance with its physical structure and local variations in collagen fiber orientations. (E) The number of waypoints in each individual toolpath is plotted against the path number before and after toolpath optimization for an 8 × 8 grid. Following optimization, paths that are longer and continuous are preferentially generated to reduce the total number of start-stop extrusion events. (F) The number of individual toolpaths with and without optimization is plotted against the number of grid units. The spread indicates SD (n = 5). The toolpathing algorithm provides a reduction of ~40% in the total number of toolpaths regardless of the number of grid units defined. (G) The local variation in anisotropy of samples 3D printed with the directionality as described above is shown (n = 3). The predicted anisotropy is shown on top while the measured anisotropy is shown below it. Each sample shows a consistent reduction of measured anisotropy relative to the predicted value due to the presence of thick cover layers. Yet, the local variation of anisotropy is retained, demonstrating the utility of this process in generating structures with locally defined mechanics.

Fig. 4.. Incorporation of fluid-filled capsules to simulate bleeding.

(A) Micrograph of double-emulsion capsules immediately after production and before evaporation of shell solvent (scale bar, 500 μm). (B) Micrograph of capsules after evaporation of shell solvent showing consistent sizing (scale bar, 500 μm). (C) Capsules show minimal passive release of their core contents when stored in an osmotically balanced collecting solution (n = 3, error bars indicate SD). Inset: Pictures of 120 capsules stored in collecting solution at day 0 and day 40. The collecting solution shows minimal discoloration. (D) The capsule shell properties such as size and maximum force required for failure can be customized by adjusting the relative flow rates of the inner phase and middle phase during capsule production (n = 7, error bars indicate SD). (E) The process of embedding capsules into a sacrificial coextrusion material for incorporating them into tissue simulants. First, the coextrusion material is loaded in its liquid state into a dispensing barrel. Next, the capsules are loaded into the barrel after the coextrusion material has undergone thermal gelation. Then, the dispensing barrel is placed in a planetary centrifugal mixer to mix the capsules into the coextrusion material to prepare a suspension that can be 3D printed. (F) The rheology of the coextrusion material-capsule suspension at various loadings. A volume fraction of 20% was chosen as acceptable for extrusion 3D printing. (G) Image of the 3D printed cric-skin tissue simulant with embedded capsules to simulate bleeding. (H) An image of the cric-skin puck after an incision demonstrating bleeding.