Research project will develop intelligent robot that can screen for bowel cancer

Across the EU, national screening programmes are being rolled out to ensure that bowel cancer is detected early. However, with the healthcare system under pressure, especially in the area of cancer, the screenings can’t always be offered at the optimal time.

This problem is the starting point for the new research project “Intelligent Robotic Endoscopes for Improved Healthcare Services”, coordinated by the Department of Computer Science at the University of Copenhagen. The project’s nine participating institutions will work together to develop an intelligent robot that can navigate inside the human body.

“We will work with a whole new generation of intelligent robots that can interact with the inside of a living human being through data, simulation and learning, and at the same time communicate with a human operator. Just like we know it from electric cars that automatically navigate in traffic”, explains Professor and Head of Section Kenny Erleben from the Department of Computer Science.

The only way to find out if a patient has bowel cancer at an early stage is by means of a telescopic examination, a so-called colonoscopy, which is performed with an endoscope.

“If we succeed in extending the current endoscope technology with robotics control, the healthcare system will be able to detect bowel cancer faster and more people will be cured”, says Professor Kenny Erleben.

A more pleasant experience

Another challenge of the bowel cancer screening process is that human errors can occur, which can affect the patient’s experience of the screening.

“Because of the huge variation in human anatomy, using endoscopes is a complicated navigational task, even for experienced clinicians. Our robotic endoscope will ensure that patients experience less pain, which will increase the level and quality of the examination”, says Professor Kenny Erleben.

The development of the robot will be based on video recordings of around 2,000 colonoscopies and associated information about whether or not anything was found in the examination. The researchers also have access to MRI and CT scans of the abdominal region, which they will use to design the simulation of the human body for training the robot.

Initially, the robot will be tested on a digital copy of a human, a so-called digital twin. In addition, the robot will be tested on a phantom, which is a training model, just as we know it from the first aid training dummy. The researchers will also build this phantom themselves because they want to create a better model than the ones that already exist.

“Once the robot has been thoroughly tested and we trust it, we can start testing the robot on humans. Testing digitally and on a phantom first will allow us to avoid animal testing. In addition, the phantom will also help train doctors better. It is a soft robot, which is something completely new in this area. The phantom will be able to change shape and behaviour while you use it”, says Professor Kenny Erleben.

The project will start on 1 March 2024 and will run for a four-year period.

Updated: