BY MITCH LESLIE
In a Boston hospital, a surgeon-in-training picks up a scalpel and tentatively begins his first incision. Though the "patient" is a computer creation, the student is still a bit nervous and pushes the blade too far into the cyber-flesh -- on a person, such a blunder might slash an underlying nerve or blood vessel. From her office in San Francisco, the supervising surgeon feels the deviation and advises the student to lighten up.
That's right -- she feels his actions from across the country. This could be the future of surgical teaching: would-be doctors honing their skills on computer simulations so accurate and vivid that they look, sound and feel like the real thing. A superfast network, the successor to the Internet, would whisk the simulations to any computer, allowing students and teachers across the globe to share and learn from the same experience.
Right now, this vision is still in the realm of science fiction. But the Stanford University Medical Media and Information Technologies (Summit) recently received a $4 million grant from the National Library of Medicine to build a small-scale prototype that would link classrooms, teaching labs and computers in the medical school. The project will focus on using high-quality graphics and high-powered computers for remote, real-time teaching of surgery and human anatomy.
According to Parvati Dev, PhD, Summit's director, the impetus for this effort comes from the federal government, which is planning to build the Next Generation Internet (NGI), a network that would be 100 times faster than the current Internet and that would incorporate a lot of "smart" technology. The Summit project is one of a handful of demonstration projects selected to assess the potential of the NGI and to help determine what specifications the new network would require.
To show how the NGI could improve medical education, Summit's team is developing two programs -- one for surgery and the other for human anatomy. The aim of the "Surgical Workbench" is to teach surgery's "hidden curriculum -- the manipulative skills and maneuvers that comprise surgery," said LeRoy Heinrichs, MD, PhD, emeritus professor of gynecology and obstetrics.
Students are expected to pick up these skills by observing surgeons and their peers during operations, said Heinrichs, who came out of retirement to work on this project. The motto of surgical teaching, Heinrichs said, has been "watch one, do one, teach one." To allow students more practice before they pick up a real scalpel, the Summit team plans to create high-resolution, three-dimensional surgical simulators that also have a "haptic" capability -- essentially, a sense of touch for the user. By donning special gloves or manipulating a computerized tool, a student or surgeon could experience (as closely as possible) what it feels like to wield a scalpel or suture a wound or palpate a tumor.
"What we've developed is a visual language for surgery -- for those active tense verbs that we use like cut, slide, move, pull, push," said Heinrichs. Specifically, using the simulators students would be able to practice eight key surgical skills: incision, excision, evacuation, injection, aspiration, scarification, closure (including suturing, stapling, gluing), transplantation and probing.
Moreover, all the data would be shared over a network, so that as one student "operates," other students and teachers can sit at their own (suitably equipped) computers and follow along, seeing, hearing and feeling the same things. For example, by picking up her version of the scalpel simulator, a surgeon would know just how deeply a med student is cutting and could offer corrections or guidance if necessary.
The technology allows such precise measurements of each surgical maneuver that, besides instructing new doctors, it could be used to evaluate surgeons who have already completed their training, Dev said. Summit is collaborating with Thomas Krummel, MD, professor and chair of surgery, to develop a program to assess surgical skills.
The other teaching tool, "Anatomy Workbench," will offer a wealth of graphics for learning human anatomy, including images from dissections and 3-D models that students can take apart and reassemble. These models will be built from digital photos taken from multiple angles and then assembled by computer, allowing students to rotate and scrutinize them from any angle.
Some parts of the project, such as the 3-D anatomical models, will be integrated into classrooms in the medical school by this fall. Other parts, such as the haptic interface, will require more development time, said Dev. Some of Summit's other key collaborators include Robert Chase, MD, emeritus professor of surgery; Dale Harris, PhD, professor of electrical engineering; Jean Heegaard, PhD, assistant professor of mechanical engineering; and Jean-Claude Latombe, PhD, professor and chair of computer sciences.
The real promise of this kind of technology lies in so-called remote teaching -- connecting students and teachers no matter where they are, on or off campus. This is where the Next Generation Internet enters the story, because the current Internet isn't up to that task.
Almost every Net user gripes about the current network's pokiness. But few of us know how unsophisticated the Internet is, at least compared to its potential successors. Some of the network's biggest problems, such as security, arise because it was designed 30 years ago to promote collegial conversation and exchange of data among far-flung academics, said Ramani Pichumani, PhD, Summit's technical manager.
And security isn't the only problem. Unlike the mail, the Internet doesn't have different classes so that high-priority messages can travel faster. As a result, the streaming video widely available on the web typically has a several second delay between the transmission and reception of the images -- a gap that is noticeable and maddening if you're trying to, say, stitch up a wound. To allow remote surgical teaching, that delay will have to be reduced to only 1 millisecond, Pichumani said.
The NGI would not only be faster, delivering 100 times more bits per second, but it would be smarter. High-priority transmissions might get more bandwidth, for instance, and packets of data could be tagged with security codes that would prevent hackers from intercepting them.
A consortium of federal agencies -- including NASA, NIH, the Defense Department and the Commerce Department -- is committed to building the network. But first they are trying to settle on its specifications and standards.
By pushing the limits of the NGI,
Summit's real-time surgery project will help define those
specifications. "If we can make the distributed visual and haptic
system work over the NGI, almost anything else would work,"
Pichumani said. SR