image of a medical robot

Robot with medical training wanted!


Kevin Coorens


Robotics (EN)

What seemed to be science fiction up until a few years ago can nowadays be seen almost everywhere around us. More and more hospitals now make use of robots and advanced technology to perform surgeries more accurately and faster than ever before. Although this (quite) new technology is only being used as an aid for now, we're very curious what the future may hold for these medically trained robots.

What exists?

Many still see robots as something very futuristic, but the implementation of this cutting-edge technology, like robots in the medical sector, isn't new at all. The first robot assisted surgery already took place in 1983 by a robot called Arthrobot. This robot was developed in Canada for orthopaedic applications. Newer and better robots have been developed ever since and can be subdivided in three categories.

The first category consists of pre-programmed robots that perform the surgery fully independently. Despite the fact that this robot performs the operation fully independently, a surgeon is still indispensable. The surgeon is present during the whole operation to supervise and intervene if necessary. Furthermore, the surgeon needs to plan every step of the surgery and program it into the robot itself.

A Da Vinci 'remote surgery' robot, leader in its field

The second category refers to the remote surgery robots. These are robots which are controlled by a doctor during the operation itself, nothing is pre-programmed. The doctor controls all of the instruments via the robot through a console while overseeing the condition of the patient using several video streams and sensors with the necessary data. With this type of robots, the surgeon doesn't need to be in the same room as the robot.

The third and last category consists of the so called shared-control robots. These aid the surgeon during an operation, but the surgeon still needs to take the largest part of the operation for his account. The robot guarantees a stable hand but the surgeon still holds all of the medical instruments himself. Robots of this category are mostly used in smaller operations where bigger more expensive robots aren't cost effective.

A logical move

Leaving these very complicated medical procedures to robots isn't applauded by everyone. Lots of people think it's unethical to make use of robots when human lives are at stake. However, it is not such a weird move to make. Robots give surgeons a better visualization of the problem area and increase the skill of surgeons during a surgery. Thanks to these medically trained robots, surgeries which were much too dangerous to perform just a few years ago, are now being performed without a single hitch. Doctors are obviously still indispensable as they are still very much needed to supervise, program or even control these medical robots. A robot can never be smarter than the surgeon who controls it. It is not supposed to replace a doctor but should exclusively be considered as an extension or aid which enables surgeons to operate more precisely and quicker in places of the human body which are hard to reach.


One of the key aspects of these medical robots consists in sufficiently visualising the specific area where the intervention will take place. In most cases the surgeon can only rely on screen-images (remote surgery) or virtual three dimensional models to pre-program the intervention. An appropriate number of images with the necessary information about the patient is a very important aspect to ensure the correct functioning of the robots. To achieve this a lot of techniques can be used from the field of image-guided surgery. This can be done because they also extensively use three dimensional representations of the patient and its internal anatomy. Based on these models a doctor can already start studying the internal structures of the patient without making any incision at all.

Building a model

Today X-ray technology is still the most commonly used technique to represent internal structures of the human body. When these X-rays penetrate our body, they get absorbed by tissues and or bones. The degree of absorption depends on the type of tissue or thickness of the bones it's trying to get through. As a result darker and lighter spots are generated. This X-ray is however nothing more than a two-dimensional representation of our three-dimensional body and as a consequence provides too scarce information for the doctors and robots.

A Siemens MRI scanner

This is why we make the transition to MRI scanners. This method uses hundreds very thin slices, as if we would cut small slices of our body. These slices are then combined into one three-dimensional model. This alternative does not only provide a far more detailed model, but is also a lot healthier for the patient on the ground that an MRI scanner does not make use of radiation, which is harmful for the human body. MRI uses magnetism instead.

The MRI scanner produces a continuous magnetic field around the patient which will react, by releasing energy, with its underlying tissues or bones. The amount of energy released will determine the brightness of the spots. The more energy released the brighter it will be shown on the model. Once the three-dimensional model is drafted, a radiologist can start to analysing and labelling the different tissues in the model.

Completing the model

The process of labelling all of the tissues and bones on a three-dimensional model is established through a collaboration between doctor and computer. For every type of tissue the doctor will indicate a level of brightness on the model which clearly belongs to this kind of tissue. Subsequently the computer will scan the rest of the model and will try to label all the remaining tissues using the data entered by the doctor. It may happen that a tissue will be impossible to label because its brightness level is too similar to that of other tissues. When this happens, the computer will check whether every candidate tissue is likely to appear in this area of the human body. For example, white matter and muscle tissue commonly have a very similar level of brightness. When the computer is now analysing a tissue in the brains which appears to match with the brightness levels of both white matter and muscle tissue, the computer can automatically label this tissue as white matter as there is no muscle tissue in the human brain. A computer knows this because it can consult a pre-programmed digital anatomical atlas of the human body. Thanks to this technique a computer can, in most cases, label a tissue correctly even though there are multiple possibilities.

Using a three-dimensional model

The described three-dimensional models provide more information than a doctor could do, such as small tumors and their exact location. A weak spot of an MRI scan on the other hand can be the representation of blood vessels. In order to solve this problem a second MRI scan is made with different settings of the magnetic field, specifically to improve the visualisation of these blood vessels. In a further stage both scans will be combined into one model which will eventually be used to plan the surgery and program the robots.

Another effective use of these models is the alignment of the model and the patient during an operation. By doing this it enables the surgeon to observe the patients' internal anatomy during the operation itself. This alignment also allows the surgeon to determine the exact location of i.e. a tumor and lets him check whether it has been completely removed.

Technology as an assistant

It is clear that computers or robots are not ready to function without a doctor controlling them. However, technology already plays a key role in the medical world, may it only be as an assistant (for now). Every year the number of robot aided and image-guided surgeries rises significantly. However, it is unlikely that robots will be operating by themselves anytime in the near future.


  1. Jochanan Eynikel, Robot aan het stuur: over de ethiek van techniek, Lannoo Campus, Leuven, eerste editie, 2017.

  2. Grimson, W. Eric L., Kikinis, Ron, Jolesz, Ferenc A., Black en Peter McL. Image-Guided Surgery, Scientific American Vol. 280, Issue 6 1999.

Download this article as a PDF.

Recente Artikels