Ferromagnetic soft catheter robots for minimally invasive bioprinting

The study addresses the challenge of enabling in vivo bioprinting inside the body without open surgery. While 3D printing has advanced many biomedical applications, most in vivo printing remains limited to skin or superficial tissues; internal organ printing typically requires invasive procedures. Alternative approaches (e.g., NIR-induced polymerization) are limited by penetration depth (~5 mm), and conventional rigid-nozzle printers are ill-suited for tortuous internal anatomy. Soft, magnetically actuated robots present opportunities for minimally invasive, dexterous manipulation. Building on ferromagnetic soft continuum robots, the authors propose a ferromagnetic soft catheter robot (FSCR) integrating magnetic actuation with extrusion printing to achieve minimally invasive, computer-controlled in situ bioprinting on flat and curved internal surfaces.

The paper situates the work within advances in 3D printing for biomedical applications and early in vivo bioprinting efforts (e.g., skin/cartilage repair, epidermal electrodes). Prior minimally invasive approaches include subdermal photopolymerization limited by light penetration and rigid-nozzle systems impractical for internal anatomy. Soft and magnetic microrobotics have shown promise in minimally invasive navigation and intervention, including ferromagnetic soft guidewire robots capable of steering in complex vasculature. Existing magnetic actuation platforms (e.g., OctoMag, MRI fringe-field systems) demonstrate control strategies but often involve complex apparatus. The FSCR builds on hard-magnetic elastomer concepts (NdFeB in elastomers), enabling remanent magnetization and remote actuation, and aims to extend these principles to stable extrusion-based bioprinting on curved, wet tissues.

Device design and fabrication: The FSCR is a slender, hollow soft catheter with an internal printing channel. A composite ink of PDMS (Sylgard 184, 10:1) with 15 vol.% NdFeB microparticles (~5 µm) was injected into a tubular mold containing a central steel core (inner template) and an inserted polylactide (PLA) fiber mesh reinforcement (16 PLA filaments, ~150 µm diameter, knitted into a 1 mm ID hollow tube). After vacuum degassing (1 h) and curing at 37 °C (48 h), the mold and core wire were removed. The cured body was magnetized with a ~3.85–4 T impulse field to saturate NdFeB along the axial direction, yielding remanent magnetization along the catheter axis. The PDMS+NdFeB composite at 15 vol.% provided magnetization ~100 kA/m and Young’s modulus ~1.15 MPa. FSCR dimensions are mold-tunable; the smallest achieved outer/inner diameters were 2.0/0.6 mm, respectively. Biocompatibility of the composite was validated via HCV-29 cytotoxicity assay (cell survival 98.6%). Reinforcement rationale: The embedded PLA mesh constrains lateral expansion of the printing channel under extrusion pressures (hundreds of kPa), reducing die-swell-induced channel dilation and extrusion delay. Non-reinforced PDMS catheters exhibited larger D/d expansion and longer delay times compared to reinforced versions at 120–280 kPa. At 240 kPa, reinforced designs showed ~4% lateral expansion and steady extrusion. Magnetic actuation and control system: A custom magnetic actuation platform uses four motorized cuboidal permanent magnets (50×50×30 mm, surface induction ~400 mT) arranged in a rectangle with north poles facing inward. The printing platform defines a coordinate frame; the nozzle tip is maintained near the XY plane at a fixed standoff equal to the intended linewidth (0.6–1.0 mm). Motors translate the magnet assembly along x (T_mag) and rotate about z (θ_mag), creating a superimposed, nonuniform magnetic field. The FSCR, possessing axial magnetization, experiences magnetic torque τ = M×B and body force f = (M·∇)B. The 3D magnetic field distribution was mapped with a Hall probe. Numerical modeling (ABAQUS with a user-defined element implementing magnetic Cauchy stress) used an analytical field model to simulate FSCR deformation under field gradients. Empirical kinematic relationships between magnet motion and tip motion were established: T_tip = 0.63 T_mag; U_tip = 0.00287 T_mag^2 − 0.0077 T_mag; θ_tip ≈ θ_mag. These mappings were embedded in control code to generate tip paths from target patterns. Printing materials and rheology: Biocompatible viscoelastic ink was prepared by mixing Ecoflex-0030 part A, PDMS SE-1700 base, and SE-1700 curing agent at 10:10:1 (w/w); viscosity ~339–340 Pa·s at 25 °C. Conductive silver ink comprised alginate solution with silver flakes (avg 10 µm) and ethanol in a 4:6:1 weight ratio; sheet resistivity tunable by silver loading (68.7–93% wt in dry film). Conductive hydrogel ink included hyaluronic acid, Pluronic F127, PEDOT:PSS, and polycarbophil in water (10 g total), mixed 24 h in ice bath; impedance characterized via two-probe EIS (5 mV, 10^−2 to 10^5 Hz). ASA-loaded hydrogel (for drug delivery) was formulated with hyaluronic acid, Pluronic F127, and acetylsalicylic acid. Printer hardware and extrusion: Inks were pneumatically extruded via a digital pressure controller (Nordson EFD) interfaced to a Raspberry Pi 2B (RS232). The nozzle inner diameter d down to 0.6 mm was used. Due to viscoelastic die swell, printed filament diameter equals a·d; a depends on nozzle speed, pressure, viscosity, and d. A resolution of ~0.53 mm was achieved at v = 3.3 mm/s, P = 240 kPa, d = 0.6 mm, viscosity ~339 Pa·s. Path planning and surface reconstruction: Target CAD patterns were converted to modified G-code-like commands parameterized by T_mag and θ_mag using the empirical kinematics; vertical compensation U_tip was applied to maintain standoff. For curved surfaces, tissue topography was reconstructed from point clouds using a laser line scanner (porcine tissue) or CT (rat liver). Surfaces were fitted, smoothed, and raster-sampled (0.1 mm intervals) to generate z-heights along desired 3D printing paths, which were then mapped to magnet motions. Experiments: Planar and 3D printing demonstrations included a flower pattern, square spiral, 3D five-layer tube, and 3D scaffold with PDMS/Ecoflex ink; accuracy assessed against designs. A conductive silver spiral coil was printed and connected to an LED; wireless powering via an external electromagnetic coil (1600 W) at ~10 mm distance lit the LED. Object manipulation (delivery/removal of 0.5–5 g loads) was shown in confined spaces. Drug delivery was demonstrated by printing ASA-loaded hydrogel on porcine tissue and quantifying salicylic acid release by UV–vis at 297 nm in PBS. In vitro minimally invasive printing was performed through an ~8 mm skin incision onto curved porcine tissue with conductive hydrogel along a preplanned spiral path (completion ~2 min). In vivo rat study: Adult Sprague Dawley rats were anesthetized; pneumoperitoneum generated with CO₂; CT reconstructed liver surface and a 3D Archimedean spiral path was planned. A thinner FSCR (25 mm length, 2 mm OD) was inserted through a ~3 mm incision to print conductive hydrogel on the liver surface within ~70 s under laparoscopic visualization. All animal procedures were IACUC-approved. Mechanical testing followed ASTM D412 for modulus; reinforced vs non-reinforced FSCR lateral and tensile responses were compared. Rheology was characterized on a TA HR-1 rheometer.

- FSCR design with PLA reinforcement enables stable extrusion: reinforced catheters exhibited ~4% lateral expansion at 240 kPa and markedly reduced extrusion delay compared to non-reinforced counterparts; non-reinforced designs showed larger D/d and slower flow rates. - Magnetic steering performance: reinforced FSCRs bend readily toward a cuboidal magnet with normalized deflection δ/L up to ~0.3; reinforcement minimally affects bending versus non-reinforced. - Empirical control mappings between magnet motion and tip motion: T_tip = 0.63 T_mag; U_tip = 0.00287 T_mag^2 − 0.0077 T_mag; θ_tip ≈ θ_mag, enabling accurate numerical control over tip trajectories. - Materials properties: PDMS+15 vol.% NdFeB composite exhibits magnetization ~100 kA/m (post-saturation) and Young’s modulus ~1.15 MPa; FSCR minimum dimensions achieved: 2.0 mm OD, 0.6 mm ID; biocompatibility confirmed with 98.6% cell survival (USP ISO 10993-5 threshold ≥70%). - Printing capability and resolution: Using d = 0.6 mm nozzle and viscoelastic ink (viscosity ~339–340 Pa·s), printed filament resolution ~0.53 mm at v = 3.3 mm/s and P = 240 kPa; complex 2D/3D patterns (flower, square spiral, 3D tube, 3D scaffold) matched designs with high accuracy. - Functional device printing: Conductive silver spiral coils were printed; an LED connected to the coil was wirelessly powered under an alternating magnetic field from a 1600 W electromagnetic coil at ~10 mm distance. - Object manipulation: FSCR delivered/moved/suctioned various materials (0.5–5 g) in confined environments. - Drug delivery: ASA-loaded hydrogel printed onto porcine tissue released drug detectable as salicylic acid via UV–vis at 297 nm in PBS. - Minimally invasive printing on curved tissues: In vitro, a conductive hydrogel spiral was printed on excised porcine tissue through an ~8 mm skin incision within ~2 min; in vivo, an Archimedean spiral was printed on a rat liver surface through a ~3 mm incision within ~70 s using a 2 mm OD FSCR under laparoscopic guidance. - Magnetic control system: Four-permanent-magnet setup is relatively simple versus commercial systems and provides sufficient nonuniform field for combined translational and rotational control over a large workspace.

The FSCR integrates hard-magnetic soft materials with a reinforced extrusion channel and a simple four-magnet actuation system to achieve precise, minimally invasive in situ bioprinting. The reinforcement addresses a key challenge of stable extrusion under high pressures in slender soft nozzles, preserving resolution and responsiveness. The empirical kinematic mappings between magnet motions and catheter tip motions allow digital, computer-controlled path following on planar and reconstructed curved surfaces. Demonstrations on porcine tissue and in a live rat model show feasibility of printing functional hydrogels on wet, curvilinear organ surfaces through small incisions and in short times, directly addressing the need for internal organ bioprinting without open surgery. The platform further supports functional device fabrication (conductive coils powering LEDs) and therapeutic delivery (drug-loaded hydrogels), indicating broad potential for minimally invasive interventions that combine navigation, printing, and device deployment.

This work presents a ferromagnetic soft catheter robot capable of remote, numerically controlled, minimally invasive bioprinting on flat and curved internal surfaces. Key contributions include a PLA-reinforced, hard-magnetic soft nozzle for steady extrusion; a compact four-magnet actuation and control scheme with validated kinematic mappings; and successful in vitro and in vivo demonstrations of printing functional hydrogels and conductive structures through small incisions. The approach opens avenues for intelligent, safer, minimally invasive biofabrication and device deployment inside the body. Future directions include: optimizing magnetic domain programming and miniaturizing FSCRs for complex 3D paths in confined anatomy; enhancing actuation systems (e.g., upgrading to six-pole magnets) for greater control freedom; integrating intraoperative imaging and vision-based sensing for real-time 3D surface reconstruction and closed-loop control; advancing path-coding strategies for higher-resolution, more complex 3D architectures; and developing bioinks with stronger adhesion to wet, curved tissues and faster solidification to maintain structural fidelity.

Current limitations include printing speed, achievable resolution, and pattern complexity in confined anatomical environments. Adhesion of printed materials to wet or vertical tissue surfaces can limit pattern fidelity; improved bioadhesive inks are needed. Many bioinks require evaporation-, gelation-, temperature- or solvent-induced solidification that may be hindered in minimally invasive settings, risking collapse of complex 3D structures before curing; faster solidifying inks or temporary biodegradable supports are desirable. The present four-magnet actuation offers limited degrees of freedom relative to more complex systems, and control accuracy could benefit from closed-loop feedback with real-time imaging. Further miniaturization and magnetic domain optimization are required for more intricate 3D paths and smaller anatomical targets.