Publication
Article
Science fiction stories about robots usually fall into one of two categories: good robots or bad robots. In the future, we are told, the machines will either be our obsequious servants, quietly following our orders according to an ingrained code of ethics, or our malevolent adversaries, hell-bent on eradicating humankind.
Science fiction stories about robots usually fall into one of two categories: “good robots” or “bad robots.” In the future, we’re told, the machines will either be our obsequious servants, quietly following our orders according to an ingrained code of ethics, or our malevolent adversaries, hell-bent on eradicating humankind. What most authors seem to have missed is the future we actually got. Call it “ubiquitous robots.”
Although a technophile may brag about the Roomba vacuuming the house or the Aibo playing with the kids, these novelties are new arrivals in a world already fi lled with automation. Robots build our cars, bake our bread, negotiate our stock trades, and even bomb our enemies. Increasingly, they also help us treat the sick.
Indeed, the most surprising news about medical robotics is not what’s just around the corner, but what’s already in use, often without physicians even noticing. That’s because many of the robots now infi ltrating hospitals and clinics bear a closer resemblance to power tools and appliances than latter-day C3POs. As these systems become more widespread, though, they are bringing about wholesale changes in the way many specialists practice medicine.
Robot Rounds
The medical robot that looks most like a science fi ction character is undoubtedly the RP-7, a “remote presence” system from InTouch Health of Santa Barbara, CA. Resembling a cross between Max Headroom and R2-D2, the RP-7 is a mobile videoconferencing system optimized for physicians. A flat-screen monitor at head level displays the physician controlling the machine, and a motorized base allows it to roll through hospital corridors, interacting with staff and visiting patients. Seated at the other end of a broadband Internet connection, the doctor can be across town or on the other side of the world, seeing and hearing whatever the RP-7 encounters, and talking through the unit’s builtin speaker. “What happens when a physician uses our platform is not any different than if they were there in person with their hands in their pockets,” says Michael Chan, executive vice president of sales for InTouch.
In 2005, the University of California, Los Angeles was among the first institutions to deploy a clinical remote presence system. Using the RP-7’s predecessor, the RP-6, intensive care specialists at UCLA were able to visit their patients even while attending conferences out of town. Since then, the technology’s popularity has skyrocketed. Of nearly 150 hospitals now using the RP-7, Chan says “we’ve probably contracted about 40—50% of those customers in the last 12 to 18 months.” Th e systems are especially useful for calling emergency consultations with far-fl ung specialists. In Michigan, for example, a network of 29 hospitals now uses remote presence as an integral part of the Michigan Stroke Network.
When a patient presents with stroke-like symptoms at any of the network hospitals’ emergency rooms, the attending physician can call in one of the network’s stroke specialists, who can reach the patient’s bedside—virtually—in a matter of minutes. Th at is particularly important for stroke, the nation’s number-three killer, for which diagnosis is often tricky and the window for eff ective treatment is only a few hours. Other hospitals have also adopted the platform for a wide range of specialties, typically paying about $5,000—$6,000 per month to rent the roving unit rather than buying it. The physician’s remote station, meanwhile, can be any off -the-shelf laptop or desktop computer that meets InTouch’s specifi cations, plus a joystick to control the robot’s movements. Th e company also sells turnkey systems with the necessary software already installed, and outlines the specifi cations for broadband connection speeds and fi rewall security.
With videoconferencing systems already ubiquitous in the corporate world, and many computers already equipped for video chats, some hospitals and clinics might be tempted to build a cheaper remote presence system on their own. “There are a number of people trying to do things like that,” Chan concedes, but he points out that the RP-7 off ers several important advantages over the alternatives; for one thing, the company’s software is designed around healthcare, making it easy to review paper or electronic charts, look at scans, and interact with patients and staff in diff erent rooms. Chan also cautions that maintaining the technical infrastructure for a secure, HIPAAcompliant videoconferencing system is not trivial. “With our product, that’s all taken care of,” he says.
Besides roaming the wards, the next generation of remote presence robots may also be scrubbing in for surgery, though perhaps not in the way some futurists have envisioned. In a deal announced in November 2007, InTouch and Intuitive Surgical, the makers of the famous daVinci robot, announced a partnership aimed at increasing the number of daVinci-trained surgeons. Rather than enabling true remote surgery, the two companies plan to create a platform for training, in which an experienced daVinci-qualifi ed surgeon could use remote presence to oversee and assist a less experienced colleague. Such a system could help ease the lengthy apprenticeship for robotic surgery, which currently requires mentor and student to acquire reciprocal privileges and travel to each others’ hospitals for extended stays.
Clamp... Scalpel... Robot
One reason for the extended training on the daVinci system is that the robot is so diff erent from what traditional surgeons experience. Instead of standing over the patient, applying instruments directly to tissue, the daVinci surgeon sits at a console beside the table, looking into a binocular-like interface and manipulating a complex set of controls. Meanwhile, the robotic instruments move around, often in minuscule fi elds deep inside the patient.
Whereas the daVinci system is highly adaptable, already performing an impressive range of fi ne-scale operations worldwide, some surgical robotics experts see the fi eld moving toward more narrowly tailored platforms. “One trend that has developed is that instead of taking these large industrial robots and basically training them to do surgery, over the last fi ve years, there has been the development of specialized surgical robots... for one or several tasks,” says Branislav Jaramaz, PhD, director, Institute for Computer Assisted Orthopaedic Surgery, Western Pennsylvania Hospital, Pittsburgh, PA. Th e trend is easy to see in orthopedics, which often requires forceful manipulations and specialized tools that a delicate, general-purpose surgical robot can’t handle. At the same time, many orthopedic procedures call for precise three-dimensional sculpting, which can be tough to visualize. To address the visualization problem, orthopedic surgeons now turn to three-dimensional modeling systems, which process the input as a surgeon touches a pointer to key anatomical points, then uses those points to project a map of the patient’s skeletal anatomy.
Such systems can, for example, highlight the mechanical axis of the leg during a total knee replacement. “Once you know the mechanical axis of the bone, you place your implant perpendicular to that axis,” says Jaramaz, adding that “you derive all of your information just based on palpation or manipulation” of the leg. Using similar visualization systems for navigation, Jaramaz and his colleagues have developed handheld robots that resemble powered surgical instruments. One tool carries a rotating burr that extends or retracts based on its current position in the operating fi eld. “It works within the navigational environment, so it needs position tracking, but then it behaves intelligently and reacts based on the surgical plan,” says Jaramaz. He and his colleagues have also developed a smaller robot that mounts directly on the bone and moves itself around the fi eld. Although the Western Pennsylvania group has focused on orthopedics, their general design approach has some advantages that could make it more popular in other types of robotic surgery as well. “With large robots, you have to store them, they are intruding into the operating room space, and you have to work around them,” says Jaramaz. “With [smaller] devices, you bring them in, you use them ... and then you take them out.”
Portable, specialized robots are also easier and less expensive to introduce than large, general-purpose systems; experienced surgeons can often learn how to use them in a short time and incorporate them into well-established procedures. Indeed, one of the major challenges for Jaramaz’s team is to make their highly precise robots compatible with often less-precise positioning systems that are already in use. If the system requires operating rooms to retool, they’ll be less likely to adopt it. “Most of the tracking that is currently done in surgical navigation is done with... lower-priced cameras,” says Jaramaz, adding that “when you introduce a robotic tool that is supposed to perform at even higher accuracy, you also need a tracking device that is of a similar level of performance.”
Where No Robot Has Gone Before
For robotics developers in Johns Hopkins University’s urology department, the accuracy of the positioning system is no problem at all; they’re building devices that operate inside patients during MRI scans. But their accurate, full-color view of the patient’s anatomy is off set by an obvious—and enormous—technical challenge: how can a robot function inside a multi-Tesla electromagnetic field?
“It’s quite a challenging engineering task,” says Dan Stoianovici, PhD, associate professor of urology and mechanical engineering at Hopkins and director of the Urology Robotics Program. Stoianovici and his colleagues have focused on building robots that can be inserted into the patient before the MRI, then navigated by remote control to take biopsies or inject localized
therapies. For some components of the tiny robots, the team could simply substitute different materials. “If you have a needle, rather than making it of steel, you make it of titanium, and perhaps it works,” says Stoianovici.
One problem that couldn’t be solved that way was the power source. Electric motors, the motive force of most robots, are fundamentally electromagnetic machines. Other researchers had tried to solve the problem with piezoelectric devices, which use an electric current to move a thin crystal, but that only allows small movements and requires high voltages. Instead, Stoianovici’s team turned to pneumatics, pressure-actuated systems usually used for fast, imprecise movements.
The scientists built a tiny air-powered engine, with cylinders and rods positioned to turn a shaft as puff s of air are distributed through the system. Because its power source is pressurized air rather than electricity, the engine can be built entirely from nonmagnetic parts. The Hopkins team is not the only group working on pneumatic robots for MRI fields. Innomedic (Herxheim, Germany) is currently the commercial leader in this area, with a system already approved for use in Europe. “From the technical standpoint, I believe that we have the better technology, because we have made this new motor, which is specifi cally designed for MRIs, whereas Innomedic tried to cope with the problems that [standard] pneumatic pistons have,” says Stoianovici. In particular, the pneumatic engine design means that the Hopkins robot makes slow, precisely controlled movements at the mechanical level, so the needle can’t go far astray if a failure somewhere in the system causes a sudden burst of air or vacuum.
In the Innomedic approach, a more traditional pneumatic cylinder is controlled by sophisticated electronics, a design some engineers might consider less safe. Nonetheless, Stoianovici concedes that the Hopkins device still has several stages of clinical trials to clear before it can be used in patients, so for now, Innomedic has the edge. With either system, the technology should allow physicians to look at a patient’s MRI scan and—while the patient is still inside the scanner—take biopsies of any suspicious areas through a pointand-click interface. A robot right in the MRI fi eld could also allow delivery of treatments with unprecedented precision. “You can go in and stick a needle and somehow kill that tumor with radiofrequency ablation, or cryoablation, or you put a few radioactive seeds right at that location,” says Stoianovici.
Like other medical robots, the pneumatic systems are tied tightly to specific applications. As robots have extended their reach in the clinic and the clinical lab (see Sidebar), they’ve evolved to look less and less like science fiction’s visions of robots.
Hopefully, they’ve also lost their urge to rule the world.
Alan Dove, PhD, is a freelance science writer who writes extensively on biotechnology, molecular biology, and public health.