Robot practices breast interventions on phantom at University of Minnesota

Remote control works inside cylindrical MRs

Engineers at the University of Minnesota and Washington University in St. Louis are developing a robot to biopsy breast lesions. Operators control the robot remotely using real-time MR guidance.

"Ideally, when you are doing an intervention, you would want an open magnet. Unfortunately, there are no widely available high-field open (1T) magnets with good homogeneity," said Nikolaos Tsekos, Ph.D., an assistant professor of radiology and biomedical engineering at the Mallinckrodt Institute of Radiology. "But there are thousands of cylindrical magnets installed in this country and a similar number elsewhere in the world. A remotely controlled robotic system would make interventions in these possible."

Tsekos' robot is built into a customized MR breast coil designed to work in conventional cylindrical scanners. MR-compatible material is used to eliminate artifacts and distortion of the local magnetic field.

The robot is composed of two plates that encircle the breast. These plates "condition" the breast for biopsy by compressing and orienting it. They also move the interventional probe into the correct position, defining the trajectory by setting the height and angulation using a probe guide that is fastened to and rotates with the plates. Once in position, the probe penetrates under remote control and real-time MR guidance to the depth of the lesion and takes the biopsy.

"With this system, we believe we will be able to exchange the probes and do multiple interventions while the patient is in the magnet," Tsekos said.

Two prototypes have been constructed at the University of Minnesota, where Tsekos developed his concept with funding from the National Cancer Institute. The prototypes are remotely controlled using ultrasonic actuators and a graphical user interface, which determine the computer-controlled 5° of freedom needed to deliver and monitor interventions, he said. These are designed to support transcannula or subcutaneous minimally invasive procedures.

The robot has been tested successfully on a breast phantom by researchers at Minnesota, who have continued to collaborate with Tsekos since he left UM to join the Mallinckrodt Institute. Arthur G. Erdman, Ph.D., UM professor of mechanical engineering, is in charge of developing the robot. J. Thomas Vaughan Jr., Ph.D., UM associate professor of Roentgen diagnosis, runs the MR side of the project.

Together, they are planning continued and more challenging phantom tests leading eventually to clinical trials. The next step will be to address simulated lesions in a phantom positioned in places that are difficult to reach.

"We don't anticipate any problems that cannot be overcome, but the additional challenges will probably put more load on the system," Erdman said.

Erdman and Tsekos believe the robot will have to be smaller to meet these challenges.

"Making it smaller will make it more flexible," Tsekos said.

But miniaturization will come at a price. MR-compatible materials do not have the same tolerances as more robust materials, according to Erdman.

"When you start to make it smaller, you will eventually reach a point where some components will fail," he said.

While one goal is to determine the limits material science will place on the design, the larger goal is to see how far this idea can go.