Robot-assisted surgery

In the operating theatre of the future, computer-based assistance systems will play a much greater role than today. They are designed to make workflows simpler and safer. Scientists at the NCT/UCC, CeTI and EKFZ for Digital Health are developing such systems with the help of new sensor technologies, artificial intelligence and human-machine interaction.

To provide meaningful support functions for surgeons, the computer must be able to anticipate important events in the operating theatre and provide the right information at the right time. For example, Dresden scientists have developed a system that can anticipate the use of certain surgical instruments. This is a prerequisite for autonomous robotic systems to take over simple subtasks in the operating theatre, such as sucking out blood.

To develop the system, the researchers used an artificial neural network. Today, artificial neural networks are already part of our everyday lives in many areas, such as product recommendations on online sales platforms or automatic image recognition on social media. The networks are able to extract information from image data. By training with large quantities of images or videos, they learn to recognize patterns in images in order to solve a given task.

The development of intelligent assistance systems often requires a wide range of data, which can be obtained and made available in an integrated operating theatre - such as the one available for research purposes in the new NCT/UCC building. Here, sensors and devices continuously record the course of treatment. In addition, a variety of information sources are linked - planning data, images generated during the operation or information about the patient and current processes in the operating theatre.

The researchers use the data, for example, to develop navigation systems for soft tissue in the abdominal cavity. During an operation, organs can change their surface due to breathing, heartbeat or contact with instruments. These deviations must be analyzed and mapped in real time. The programs developed must therefore be able to immediately calculate surface changes from various pre- and intraoperatively acquired data as well as biomechanical models.

In the future, intelligent systems should also relieve the surgical team during keyhole surgery by robot-assisted guidance of the surgical camera (endoscope or laparoscope). In addition, they could warn of critical situations and, for example, indicate at an early stage that additional blood reserves are needed. Using machine learning methods, computer-based assistance systems should also continuously learn from the best surgeons. This could make their expertise available to all surgeons in everyday clinical practice.