By Marlene Cimons, National Science Foundation
During the past decade, the growing use of medical robotic devices to perform minimally invasive surgery--procedures done through small incisions—has made it easier for surgeons to do many things they couldn't do before. But the new technology can't do everything—yet.
"Right now, these robots are dumb," said M. Cenk Cavusoglu, associate professor in the department of electrical engineering and computer sciences at Case Western Reserve University in Cleveland. "They have no intelligence at all. They are still controlled by the surgeon. We want to make the robots smarter. We want to turn them into surgical assistants, almost like giving the surgeon a third arm."
He is collaborating with Case Western Reserve colleague Wyatt S. Newman, a professor in his department; Ken Goldberg, professor, and Pieter Abbeel, assistant professor, both in the electrical engineering and computer sciences department at the University of California at Berkeley; and Ron Alterovitz, assistant professor in the University of North Carolina department of computer sciences.
The team is developing approaches and algorithms to teach the robots new skills such as suturing, tying knots and manipulating body tissues. If a robot can learn to perform these additional tasks, surgical procedures will run faster and more efficiently, Cavusoglu said.
The project is funded over four years by a $1.3 million grant from the National Science Foundation as part of the American Recovery and Reinvestment Act of 2009.
The work has the potential to provide a further economic boost to the field of robotics, a dynamic component of the technology industry. "The United States is in the lead when it comes to medical robotics," Cavusoglu said. "We want to keep our advantage. We want to maintain our lead in this field, where engineering and computer sciences come together."
The technology also could help lower overall health care costs by making surgeries faster, safer and more effective. "During a long, tiring surgery, the surgeon may start to make errors," Cavusoglu said. "We're hoping this will save money by reducing hospital errors and patient stays, as well as reducing the time it takes to perform an individual surgery."
Minimally invasive surgeries—arthroscopy, laparoscopy, endoscopy, for example--have become preferable alternatives to traditional surgery, particularly for patients, because they produce faster recovery times, fewer complications and less trauma, as well as reduced hospital costs. But the operations can be quite demanding for the surgeons.
In traditional open surgery, the physician makes a long incision and then widens it to access the surgical target area. In traditional minimally invasive surgery, which is widely used for routine procedures, the surgeon views the anatomy on a standard video monitor and operates through small incisions using rigid, hand-operated instruments. This can be tedious, time-consuming and labor intensive. "It's like trying to do surgery using chopsticks," Cavusoglu said.
Using dexterous robots that are remotely controlled by the surgeon with a joystick has made these surgeries less tiring and more efficient. "You don't have to put your hands in; you put the tiny robots in," he said.
But, typically, the surgeon must operate alone, without human assistants, because the devices are large and space in the room is tight. "In open surgery, surgeons have assistants, but in robotic ones, there is only the primary surgeon who has to do everything," Cavusoglu said. "Nobody can help."
Ultimately, the goal of the research is to "teach" the robots to function in the operating room as human surgical assistants would.
"Right now, the way it works is that the robot doesn't do anything by itself," Cavusoglu said. "They work by remote control—the surgeon uses a joystick, and the robot replicates everything the surgeon is doing with the joystick—nothing more, nothing less. The surgeon can use the joystick to get the robot to pull tissue out of the way. We want to have the robot, by itself, push the tissue out of the way so the surgeon can work.
"You don't ask a (human) assistant to get this out of the way so I can do what I have to do. You don't grab your assistant's hand and manipulate it," he added. "The assistant does it on his or her own. We want the robot to function like an assistant."
The researchers are using computer programs and actual surgical demonstrations to teach the robots how to do it.
"We have the surgeon do the suturing and tie the knot several times, and the robot looks at the examples and learns how the surgeon is doing it—then does it faster and better than the surgeon," Cavusoglu said. "A computer program in the robot enables the robot to watch the surgeon, and learn from the surgeon's example."
The scientists also are programming the robots to manipulate soft tissue, and handle delicate organs. "Rather than learning by observing, this we do by programming," he said.
The two approaches complement each other. "Some things are easier learned by observation, some things easier by programming," Cavusoglu said. "Eventually, we want to combine them so we can do even more complicated things."
Follow U.S. News Science on Twitter.