A Perfect Surgeon?
For the first time ever, a robot has carried out surgery sufficiently accurately to out-compete human surgeons. The breakthrough makes doctors believe that robots are going to take over surgery completely, so patients can be operated faster and with much
With cool, mechanical motions, the robot moves needle and thread back and forth through the shiny tissue of the pig bowel. Named STAR (Smart Tissue Autonomous Robot), the robot is the first of its kind to carry out a difficult type of suture joining of two bowel segments, which has only been carried out by human surgeons so far. In the operating room, scientists from the Johns Hopkins University in the US and others, are pushing the technological limits well into the future. In a few years, robotic surgeons in hospitals throughout the world will enter the operating rooms and take over the tasks of human surgeons.
SUTURE IS DIFFICULT
The STAR robot’s achievements were published in an article in the Science Translational Medicine journal in 2016. And sutures are not by far the only medical discipline in which robots will be able to beat humans over time. Robots with extremely accurate vision and motor functions allow complex eye surgery, and artificial intelligences such as IBM’s Watson can now diagnose cancer patients much faster than doctors, allowing that the right diagnose be
made in a matter of minutes instead of days or months. The surgery is much more complex than mending a superficial cut in a finger. The bowels behave like two wobbling, unmanageable rubber tubes, and the robot must find a way to sew along the edges of the two separate, tubular organ segments without making any erroneous stitches. Last, but not least, the robot must constantly use its cameras to observe and continuously update its suture plan, if the tissue moves. When STAR's algorithms have completed the joining of the pig bowels, the scientists conclude that the robot has solved the task more accurately than a surgeon could have.
The STAR robot’s suture capabilities have attracted attention in scientific circles, as suture is one of the fields of surgery, in which there is room for improvement. Undesirable complications can easily occur after suture inside a patient’s body, as this type of surgery requires much experience. Human tissue gives way, is deformed, or contracts, when the doctor impinges it, makes a suture, or cuts. Because the tissue is so “alive”, it can be difficult for human surgeons to make a perfect suture. The robot can do the job with a margin of error of 0.05 mm almost indefinitely, whereas a human being, disregarding his expertise, cannot do the same. Accurate suture that does not ”leak” is vital for wound healing and so for the patient’s fast recovery. The distance between individual stitches cannot be so long that the suture ”leaks”, but on the other hand, the distance must be so long that the blood can flow unimpeded through the tissue to help the healing process. The STAR robot uses two cameras – a near infrared camera and a plenoptic 3D camera – to overlook the area in which the suture is made. With the information from the cameras, the robot can determine the thickness of the bowel walls and the shape of the tissue surface in 3D. The planning algorithm uses the information to prepare a plan of the needle’s and the thread’s most optimum route through the bowels. The algorithm has also been fed surgical knowledge about what an optimum suture looks like. Moreover, the software will automatically find out, in case the tissue moves during surgery.
SECOND GENERATION
Until now, it has been impossible for a machine to suture internal organs – until the STAR robot’s performance.
Yet robot surgeons have been used in operating rooms since 2000, assisting during prostate cancer surgery. Instead of working directly above the patient on the operating table, the surgeon sits down by a console, where an enlarged image of the point of surgery of the patient is shown. Via console joysticks, the surgeon can remote-control several robotic arms above the patient. The arms hold the instruments that the surgeon normally uses during surgery. The idea is that instead of using the scalpel and needle himself, the surgeon can control the robotic
arms with much stabler motions than any human is able to offer. The company behind this type of robot claims to have carried out more than three million robotic surgery sessions globally.
STAR can't complete the bowel surgery without human help. Today, using dissecting forceps, an assistant sees to it that the thread does not get entangled in the robot’s sewing instrument. But there is every indication that still faster computer chips, improved cameras, and more sophisticated and intelligent algorithms will one day make robots as capable as humans.
In order to make the new generation of robots more versatile, scientists from, Berkeley will make them learn from surgeons by having the robots observe the surgeons when they work. The motions of a human surgeon who controls one of the robots during different types of surgery will be saved as digital knowledge, which scientists’ fully automatic robotic prototype can analyse and use to refine its algorithms. In practice, the robot becomes a medical student who looks over his teacher’s shoulder to improve his abilities. But to be able to learn from surgeons, robots must be able to think.
SMARTER THAN THE HUMAN BRAIN
Artificial intelligence is one of the most important fields of research in scientists’ efforts to improve the treatment of diseases in humans. IBM’s artificial intelligence, Watson, became world famous in 2011 by beating two Jeopardy champions. The victory demonstrated Watson’s ability to read lots of text material and understand the meaning. Cancer researchers from the Institute for Cancer Research in Oslo, Norway, will now benefit from Watson’s analytical capabilities. The artificial intelligence can search through large quantities of research articles about cancer much faster than a human and quickly identify new contexts regarding the disease, which it would have taken scientists months or years to find.
Google researchers have used the deep learning method to create an algorithm that can identify the diabetic retinopathy eye disease, which affects diabetics and could lead to blindness. To identify early symptoms of the disease, images of the patient’s retina must be studied carefully for any evidence of haemorrhage. Google used 128,000 retina photos to train an artificial intelligence algorithm to recognize evidence of diabetic retinopathy. After the training, tests of the algorithm showed that it was just as good as doctors at spotting evidence of the eye disease. In the future, this type of software can become an invaluable help for doctors, not least in poor regions of the world, where lack of specialists often means that eye diseases such as diabetic retinopathy are not identified and treated in time.
SURGEON ROBOTS GO TO MARS
Earth is not the only place in which artificial intelligence and robots could revolutionize medicine. On future missions to Mars and beyond, astronauts will be away from Earth for years, and a robotic surgeon would be able to save lives.
The American company Virtual Incision has developed a robotic surgeon that weighs 0.4 kg and is so compact that it can be introduced via a small incision by the navel and use small robotic arms in the stomach of the patient to suture an ulcer, etc. In a state of weightlessness, it would be a great advantage to be able to avoid large openings in the body during surgery, as blood and other bodily fluids could easily escape. So far, the robot must be controlled by a surgeon, but future versions could become partly or fully i ndependent of human help and consequently save lives hundreds of millions of km from our own planet.
For surgeons, robots will not be synonymous with unemployment right away. Even when the robots have become better than their human competitors at making sutures on their own or reacting to unexpected complications, doctors and surgeons will be able to contribute with their experience and make decisions, when robots turn out to be inadequate. Hence, scientists are expecting a future, in which the roles of the operating room will be reversed. The surgeon and the nurses can ”circulate”, participating in several operations to be of assistance, when necessary, while robots carry out the rough tasks. The future perspective will mean better and cheaper surgery and faster patient recovery.