Medical Robotics

Laser Tumor Ablation

Laser scalpels offer the benefit of fast, accurate, and gentle tissue removal via photoablation, or vaporization, with inherent cauterization effects that reduce bleeding and collateral mechanical damage that would otherwise occur with standard metal scalpels. These types of systems form the basis of LASIK eye surgery and similar systems have been demonstrated in a variety of surgical applications. However, in each of these applications only small ‘unit’ ablations or simple straight cuts are performed with little or no closed-loop feedback for control. We’ve been working with Dr Patrick Codd, MD and the Brain Tool Lab to develop the closed-loop surgical planning methodology that is needed to use laser scalpels for larger, volumetric resections (e.g. resection of an entire tumorous volume) without a surgeon performing teleoperation.

Image
The figure of the initial tissue boundary before laser cutting.
Initial tumor boundary, objective, and constraint surfaces before laser ablation.
Image
The figure of the final tissue boundary after laser cutting.
Final tumor boundary, objective, and constraint surfaces after laser ablation.

Autonomous Ultrasound

Image
The figure of the experiment setup for robotic ultrasound imaging
Robotic ultrasound imaging experiment setup.

In collaboration with Dr Dan Buckland, MD and Dr Siobhan Oca, we’ve been researching how to automate ultrasound scans. Medical roboticists and sci-fi fans alike have fantasized about devices that can autonomously perform numerous medical procedures. A reasonable first step towards this goal would be to automate processes that are both key to numerous procedures, and tedious for humans to perform. Ultrasound (US) scanning is in the `sweet spot' of processes that require high precision and low creativity: collecting useful scans requires practice, demanding skilled technicians, but the task is well-defined and repeated. Our goal is to make autonomous US (AUS) a viable tool, but most robotic US systems rely on teleoperation. The few existing robotic US systems that do not rely on teleoperation have so far only scanned predefined trajectories, putting them at a low-level of autonomy. Moreover, they have only scanned nearly-flat, relatively hard surfaces, meaning they fail on muscular people, fat people, and anyone’s abdomen or chest -key diagnostic regions. We aim to enhance the state-of-the-art in AUS by automating trajectory selection for interactions with soft tissue.


Bass Connections: Reimagining Surgery for Rural Needs: Robotics Teleoperation 

For years, people have claimed that medical robotics would transform care in rural and remote settings, but medical care is often limited or hard to reach, even in rural settings that are hardly remote; in North Carolina, 92% of counties currently have a shortage of health professionals. Through Duke’s Bass Connections program, which involves undergraduate students in community-based research, we’re investigating how to make surgical robotics helpful in rural care settings, and prototyping robots to fulfill those needs.