Eversion-Based Robots Enable Safe Navigation Inside the Spinal Canal

Eversion-Based Robots Enable Safe Navigation Inside the Spinal Canal

Zicong Wu, Panagiotis Kalozoumis, S. M. Hadi Sadati, Aminul I. Ahmed, Jonathan Shapey +6 more

11 min readJul 26, 2026

Unlike conventional push-based catheters, Vine4Spine advances through distal eversion of a thin-walled, compliant sheath. Growth occurs exclusively at the tip by controlled pressurisation of the folded sheath within a chamber, causing the material to evert and extend forward. Thus, the outer surface of the grown segment remains stationary relative to the surrounding tissue. This tip-based growth mechanism reduces relative sliding between the robot body and the intrathecal environment. In the narrow, compliant, and fluid-filled spinal subarachnoid space, where lateral clearance is only a few millimetres, minimising axial shear forces and distributed friction is critical for safe navigation.

Vine4Spine Demonstrates Safe Navigation in a Human Cadaveric Spine

To facilitate fluoroscopic visualisation, the pressure-regulation fluid was mixed with a radiographic contrast agent. A clinical introducer equipped with a sealed irrigation port provided a lubricated access pathway, reducing insertion resistance while maintaining a fluid-filled environment consistent with cerebrospinal fluid and improving endoscopic visualisation during navigation. Through this conventional clinical access route, Vine4Spine was deployed using pressure-driven distal eversion while steering was achieved by the inner catheter. This represents the first demonstration of controlled eversion-based robotic growth and steering within the intact human spinal subarachnoid space.

Load-Decoupled Growth via Tip-Eversion Establishes a New Paradigm of Intrathecal Navigation

Conventional intrathecal navigation relies on proximal pushing, inevitably transmitting compressive loads and friction along the entire catheter shaft as insertion depth increases. The present study demonstrates an alternative navigation paradigm in which robot extension is generated exclusively at the distal tip through eversion, fundamentally changing how mechanical loads are transferred to the surrounding anatomy. Rather than reducing interaction through increased flexibility or compliant materials alone, Vine4Spine eliminates translational sliding of the deployed body, thereby decoupling distal advancement from proximal pushing.

Navigation within the spinal subarachnoid space presents a uniquely demanding mechanical environment: a narrow, compliant, fluid-filled corridor densely populated by fragile neural elements. Conventional push-based catheters propagate axial loads from the proximal base to the distal tip, accumulating friction along the deployed shaft and degrading distal controllability as insertion length increases. In contrast, the eversion-based architecture localises growth to the distal boundary. Because the deployed sheath remains stationary relative to surrounding anatomy, shear forces along the shaft are minimised and axial load propagation is reduced.

CAD rendering showing Vine4Spine robot assembly with steerable inner catheter and everting sheath

The cadaveric results demonstrate that this load-decoupled growth principle is not limited to benchtop or phantom environments but can be sustained in intact human anatomy under clinically realistic access conditions. The preservation of steering curvature transmission at extended intrathecal lengths indicates that steering does not degrade with deployment depth, addressing limitations of push-based systems in compliant intraluminal environments.

Together, these findings demonstrate that load-decoupled growth is not restricted to benchtop environments but can be translated to clinically relevant spinal anatomy under conventional lumbar access.

Friction-Dominated Mechanics Govern Miniature Eversion-Based Navigation

The optimisation of duty-cycle parameters revealed that eversion-based systems operating at a miniature scale are governed by friction-dominated contact mechanics rather than idealised kinematic symmetry. Under purely geometric assumptions, a retraction-to-advancement ratio of approximately 0.5 would be expected to restore the relative alignment between the inner instrument and the everting sheath after each duty cycle. However, experimental observations demonstrated persistent residual coupling even when the internal pressure approached atmospheric conditions, indicating that friction hysteresis and membrane interactions within the folded sheath substantially influence the deployment mechanics.

Reducing the retraction ratio to approximately 0.4 and introducing sub-atmospheric vacuum pressure during the depressurisation phase effectively compensated for these friction-induced asymmetries, enabling sufficient mechanical decoupling and stable incremental growth over repeated duty cycles. In this context, duty-cycle parameters function not as purely geometric quantities, but rather as control variables that compensate for configuration-dependent frictional interactions between the folded sheath and the inner instrument.

Negative pressure plays a central role in this mechanism. By generating a pressure differential across the sheath wall, the depressurisation phase increases contact between the folded sheath layers and suppresses unintended sheath motion during instrument retraction. This stabilises the position of the deployed sheath while allowing the inner instrument to retract independently, thereby preserving coaxial alignment throughout successive growth cycles. Rather than serving solely as a mechanism for resetting the growth cycle, pressure regulation actively controls the mechanical coupling between the sheath and the inner instrument.

The computational model further supports this interpretation by demonstrating the dynamic evolution of interface forces throughout each duty cycle. While interaction forces increased during pressurisation and forward advancement, they progressively decreased during vacuum-assisted retraction, consistent with the experimentally implemented four-phase control strategy. These results indicate that pressure regulation serves not only as a growth actuator but also as a mechanism for controlling tissue interaction throughout the insertion cycle.

Together, these findings establish that stable eversion-based navigation in confined neural anatomy is possible with controlled internal pressure modulation of everting robots.

Endoscopic image captured from Vine4Spine showing view inside the spinal subarachnoid space

Structural Integrity Preserves Mechanical and Optical Stability

An important finding from the cadaver study is that the load-decoupled eversion architecture maintained its structural integrity throughout repeated deployment and steering cycles. No fluoroscopic evidence of contrast leakage, chamber collapse, or interruption of eversion was observed during navigation, indicating that the pressurised chamber remained effectively sealed despite repeated pressure cycling within the confined intrathecal environment. Together with the preserved coaxial alignment between the inner instrument and the everting sheath demonstrated by both fluoroscopic and endoscopic imaging, these observations indicate that the proposed architecture maintains structural stability over clinically relevant deployment lengths.

Maintaining instrument–sheath alignment is particularly important for pressure-driven growing robots because misalignment between the steering instrument and the deployed sheath may compromise both distal steering authority and optical performance. The consistent correspondence between fluoroscopic observations and endoscopic visualisation throughout deployment suggests that the depressurisation phase of the duty cycle effectively decouples motion between the inner instrument and the deployed sheath, preventing unintended sheath displacement while preserving the structural configuration required for continued navigation.

Beyond its mechanical role, this structural stability directly benefits in situ imaging. The embedded chip-on-tip endoscope relies on a stable geometric relationship between the imaging sensor, the surrounding sheath, and the cerebrospinal fluid interface. Transient sheath recoil or mechanical perturbations during repeated duty-cycle operation could introduce camera tilt, lens–sheath contact, or motion-induced blur, reducing image quality and navigation confidence. By suppressing unintended sheath motion and maintaining coaxial alignment, the proposed pressure-driven deployment strategy provides a stable optical environment that preserves a consistent field of view throughout intrathecal navigation.

Importantly, these observations suggest that pressure regulation fulfils multiple functions beyond driving distal growth. In addition to enabling eversion, the duty-cycle strategy preserves the structural integrity of the robot, stabilises the imaging platform, and maintains alignment between the steerable instrument and the deployed sheath during prolonged navigation. The ability to simultaneously maintain mechanical robustness, pressure sealing, steering capability, and endoscopic image quality represents an important systems-level advantage of the load-decoupled eversion architecture for minimally invasive intrathecal interventions.

Mechanical Considerations for Successful Entry into the Spinal Subarachnoid Space

A critical phase of intrathecal navigation occurs immediately after the robot emerges from the rigid introducer and enters the spinal subarachnoid space. Unlike navigation within the introducer, where the robot is mechanically constrained, the distal segment becomes unsupported upon exiting the introducer and must simultaneously establish its orientation while maintaining sufficient structural stability for continued growth. Consequently, the first few millimetres of free distal extension represent the most mechanically demanding stage of deployment.

Successful entry requires a careful balance between tendon-driven steering and pressure-driven distal extension. If forward growth dominates before adequate steering is achieved, the distal segment experiences predominantly axial compressive loading, increasing the likelihood of tip buckling and deviation from the desired navigation axis. Conversely, excessive or delayed steering shifts the bending moment proximally, increasing mechanical loading near the introducer exit and reducing the effectiveness of distal curvature generation. The successful navigation demonstrated in this study suggests that appropriately coordinated steering immediately after introducer exit is essential for establishing the correct intrathecal trajectory while maintaining load-decoupled growth.

Once the robot is aligned with the axis of the spinal subarachnoid space, subsequent pressure-driven eversion substantially reduces axial loading along the deployed body, allowing distal curvature control to be preserved throughout continued navigation. This transition from a mechanically demanding entry phase to stable load-decoupled growth highlights the importance of coordinating steering and deployment during the initial stages of intrathecal access.

Biomechanical Mechanisms Underlying Reduced Tissue Interaction

Although the phantom experiments demonstrated substantial reductions in interaction forces during navigation, force measurements alone cannot explain how the eversion mechanism alters tissue loading. The comparison among continuous insertion, duty-cycle insertion and Vine4Spine showed that, while the duty-cycle strategy reduced the overall interaction force, it retained transient force peaks comparable to conventional insertion. In contrast, Vine4Spine simultaneously reduced both the mean and peak interaction forces and produced the smoothest force profile throughout navigation. These observations suggest that the reduced interaction cannot be explained solely by differences in insertion kinematics, but instead reflects a fundamentally different mode of mechanical interaction with the surrounding tissue.

Importantly, the finite element model reproduced the experimentally observed reduction in interaction force with good agreement for Vine4Spine, supporting its use as a tool for investigating tissue-level biomechanical phenomena beyond the capability of load-cell measurements alone. Although discrepancies remained in the predicted peak force during conventional catheter insertion, the model consistently captured the relative biomechanical advantages of eversion-based navigation. Together, the computational and experimental findings indicate that the principal benefit of Vine4Spine extends beyond reducing insertion force alone. By localising motion to the growing tip while the deployed sheath remains stationary relative to the surrounding anatomy, the system redistributes tissue loading, limits frictional shear accumulation along the robot body, and may therefore reduce the risk of localised mechanical trauma during navigation.

Translation Towards Clinically Relevant Intrathecal Procedures

Initial cadaveric investigations highlighted that the access interface between the robotic system and the spinal canal is a critical determinant of translational feasibility. Early attempts to deploy the system directly through a standard introducer for lumbar puncture revealed substantial friction at the entry interface, exacerbated by the viscosity of contrast-containing pressurisation media and the tight dimensional tolerances of the introducer lumen. These factors increased resistance to sheath eversion and limited the reliability of initial deployment.

To address this limitation, a clinical introducer (Cordis AVANTI®) was adopted as the access interface. The introducer provided a large, smooth internal lumen and a sealed side port, enabling controlled saline flushing throughout deployment. Continuous or intermittent saline irrigation maintained a fluid-filled pathway analogous to the CSF environment, reducing friction at the access corridor and facilitating eversion without risk of rupture. This configuration preserved compatibility with standard lumbar puncture workflow and required no additional surgical exposure.

Together, our findings illustrate that intrathecal navigation using tip-growing robots requires coordinated regulation of extension length, steering actuation, and access interface conditions. While the eversion-based growth mechanism mitigates friction accumulation along the deployed shaft, effective clinical deployment still depends on maintaining a low-resistance entry interface and carefully managing the mechanical transition from introducer-guided insertion to free intrathecal navigation.

Safety Implications Across Computational, Phantom and Cadaver Validation

The present study evaluates navigation safety using three complementary levels of evidence. Computational modelling demonstrated substantial reductions in local tissue stress and interfacial shear stress; phantom experiments quantified significantly lower global interaction forces during navigation; and cadaveric studies revealed no observable macroscopic disruption of the dura mater or spinal nerve roots during or following robotic deployment. Collectively, these findings provide converging evidence that the load-decoupled eversion architecture offers a safety profile distinct from conventional push-based catheterisation, potentially reducing the risk of iatrogenic injury during intrathecal navigation.

Frequently Asked Questions

How does Vine4Spine differ from conventional catheters for spinal navigation? Vine4Spine grows by everting a sheath at its tip rather than being pushed from the base, so the outer surface of the deployed body stays stationary relative to tissue, dramatically reducing friction and sliding forces.

What are the key technical challenges for entering the spinal subarachnoid space? The most demanding phase is the first few millimetres after exiting the introducer, where the robot must balance steering and forward growth to avoid buckling while establishing the correct intrathecal trajectory.

How is endoscopic image quality maintained during navigation? The pressure-driven duty cycle preserves coaxial alignment between the inner instrument and everting sheath, stabilising the imaging platform and preventing camera tilt or motion blur throughout deployment.

What safety evidence supports the eversion-based approach? Three complementary levels were validated: computational modelling showed reduced tissue stress, phantom experiments measured lower interaction forces, and cadaveric studies confirmed no observable damage to the dura mater or spinal nerves.

🍪 Cookie preferences

We use cookies to measure performance. Privacy Policy