TL;DR
Researchers have developed "Helping Hand," a two-segment continuum robotic grasper that enables simultaneous tissue traction and electrosurgical cutting through a single 3.7 mm endoscope channel. By incorporating a novel side-exit slot for the surgical knife, this device provides the "countertraction" necessary for complex polyp removal without the need for bulky external attachments or expensive dual-channel endoscopes.
What Problem Does This Paper Address?
Endoscopic Submucosal Dissection (ESD) is a gold-standard procedure for removing large, potentially cancerous polyps from the colon en bloc (in one piece). This is superior to older methods that remove tissue in fragments, as it allows for better pathological assessment and reduces the risk of cancer recurrence.
However, ESD is notoriously difficult. As a surgeon cuts the tissue, the resulting "flap" often flops over, obscuring the view of the dissection plane. To see where they are cutting, surgeons need countertraction—a way to pull the tissue flap away from the muscle wall. Currently, this requires either:
- External robotic attachments: These add bulk to the endoscope tip, making it hard to navigate the twisty colon.
- Dual-channel endoscopes: These are expensive and not widely available in most hospitals.
The "Helping Hand" aims to bring advanced robotic dexterity and countertraction to the standard, single-channel colonoscopes already found in most clinics.
Methods
The authors designed a two-degree-of-freedom continuum robot fabricated from a hollow Nitinol tube. The system is divided into two main subsystems:
1. The Instrument Subsystem
The core innovation is the notched tube design. It features two independently actuated bending segments:
- Proximal Segment: Contains a unique "knife slot." When the first notch of this segment bends, it directs an internal electrosurgical knife out through the side of the tube.
- Distal Segment: Ends in a micro-grasper.
- Routing: Internal 3D-printed "routers" (Fig. 3) manage the complex internal layout, ensuring the tendons, grasper shaft, and electrosurgical knife don't interfere with each other.
2. The Actuation Subsystem
The robot is tendon-driven via a modular platform using four lead-screw-driven linear motion units. These units control the proximal tendon, distal tendon, knife translation, and grasper activation. Crucially, the system includes inline load cells to measure tendon tension in real-time.
Datasets
As a robotics hardware paper, this work does not use traditional machine learning datasets. Instead, the system was evaluated using:
- Electromagnetic (EM) Tracking: To characterize the reachable workspace and validate kinematic models.
- Silicone Phantoms: An 18 mm "polyp" mounted on a six-axis force/torque sensor to measure retraction forces.
- Ex Vivo Tissue: A chicken thigh model was used to demonstrate a full "grasp, retract, and cut" workflow, as its layered connective tissue mimics the colonic wall.
Key Insights and Ideas
- The "S-Bend" Strategy: By orienting the distal notches 180° from the proximal ones, the robot can form an S-shape. This allows the grasper to reach "down" to a polyp while keeping the knife-exit slot positioned optimally for the subsequent cut.
- Side-Exit vs. Tip-Exit: Traditional endoscopic tools exit from the very tip. By creating a side-exit slot for the knife (Fig. 5), the authors decoupled the cutting tool from the grasping tool, allowing them to work in tandem within a tiny 3.7 mm footprint.
- Manufacturing Precision: The authors developed a repeatable two-operation CNC milling process using a custom fixture (Fig. 4) to ensure the notches on both sides of the Nitinol tube were perfectly aligned.
Key Results
- Workspace Accuracy: The kinematic model predicted the distal segment's movement with a 9.1% error (NRMSE). The proximal segment had a higher error (33.9%), likely due to unmodeled friction and tendon elongation.
- Traction Force: The device successfully applied a mean resultant force of 0.33 ± 0.02 N on the phantom. Importantly, it provided a "normal force" (lifting force) of 0.27 ± 0.04 N, proving it can effectively lift tissue flaps to expose the dissection plane.
- Functional Demonstration: In the ex vivo chicken thigh test, the Helping Hand completed a full en bloc resection in 15 minutes. This included grasping the target, applying traction, making the initial incision, and advancing the cut until the specimen was fully retrieved.
Limitations and Future Work
- Model Fidelity: The current kinematic model assumes constant curvature and ignores friction and tendon stretch. Future versions will need mechanics-based models to improve accuracy.
- Open-Loop Control: The current system is operated manually. The authors suggest that integrating visual feedback from the endoscope camera or force feedback from the load cells could enable automated countertraction, where the robot automatically adjusts its pull as the surgeon cuts.
- Clinical Integration: While successful in a benchtop model, the system still needs to be tested in a dynamic, insufflated environment typical of a real colonoscopy.
Summary updated