Many surgical procedures now use long, thin devices – such as “steerable needles” – that can be inserted into a patient’s body through a small incision and then steered to a target location. These “minimally invasive” procedures allow doctors to perform surgeries without having to make major incisions, which decreases the risk of infection and shortens the patient’s recovery time.
This graphic illustrates a surgical tool in a human lung. The blue curve corresponds to what we expect the device to do. The green curve represents what would happen in a real procedure were some perturbations introduced. The red-dots represent the estimated shape based on where the new x-ray algorithm says the surgical tool actually is. (Click to enlarge. Image credit: Edgar Lobaton.)
However, these techniques pose a challenge to surgeons, because it is difficult for them to determine precisely where the surgical device is in the patient’s body.
One solution to the problem is to use X-rays to track the progress of the surgical device in the patient. But doctors want to minimize the number of X-rays taken, in order to limit the patient’s exposure to radiation.
“We have now developed an algorithm to determine the fewest number of X-rays that need to be taken, as well as what angles they need to be taken from, in order to give surgeons the information they need on a surgical device’s location in the body,” says Dr. Edgar Lobaton, an assistant professor of electrical and computer engineering at NC State and lead author of a paper on the research.
The new tool is a computer program that allows surgeons to enter what type of procedure they’ll be performing and how precise they need the location data to be. Those variables are then plugged into the algorithm developed by the research team, which tells the surgeon how many X-rays will be needed – and from which angles – to produce the necessary location details.
For example, if a surgeon needs only a fairly general idea of where a device is located, only two or three X-rays may be needed – whereas more X-rays would be required if the surgeon needs extremely precise location data.
The paper, “Continuous Shape Estimation of Continuum Robots Using X-ray Images,” will be presented at the IEEE International Conference on Robotics and Automation, being held in Karlsruhe, Germany, May 6-10. The paper was co-authored by Jingua Fu, a former graduate student at UNC; Luis Torres, a Ph.D. student at UNC; and Dr. Ron Alterovitz, an assistant professor of computer science at UNC. The research was supported by the National Science Foundation and the National Institutes of Health.
Note to Editors: The study abstract follows.
“Continuous Shape Estimation of Continuum Robots Using X-ray Images”
Authors: Edgar J. Lobaton, North Carolina State University; Jingua Fu, Luis G. Torres and Ron Alterovitz, University of North Carolina at Chapel Hill
Presented: May 6-10, 2013, IEEE International Conference on Robotics and Automation, Karlsruhe, Germany
Abstract: We present a new method for estimating the shape of a continuum robot continuously during a medical procedure using a small number of X-ray projection images. Continuum robots have curvilinear structure, enabling them to maneuver through constrained spaces in a snake-like manner. An accurate estimate of the robot’s shape is crucial for the success of procedures that require avoidance of anatomical obstacles and sensitive tissues. Online shape estimation of a continuum robot is complicated by uncertainty in its kinematic model, movement of the robot during the procedure, noise in X-ray images, and the clinical need to minimize the number of X-ray images acquired. Our new method integrates kinematics models of the robot with data extracted from an optimally selected set of X-ray projection images. Our method represents the shape of the continuum robot over time as a deformable surface which can be described as a linear combination of time and space bases. We take advantage of probabilistic priors and numeric optimization to select optimal camera configurations, thus minimizing the expected shape estimation error. We evaluate our method using simulated concentric tube robot procedures and demonstrate that obtaining 3 images from viewpoints selected by our method achieves shape estimation errors significantly lower than using the kinematic model alone or using uniformly spaced viewpoints.