Early stage diagnosis of colorectal cancer increases the patient’s survival chance. Colonoscopy, a process used to examine the rectum and colon, helps to delve deep inside the human body. Conventional colonoscopy is a painful, highly complex, expensive procedure in which a semi-flexible tube (colonoscope) fitted with a tiny video camera at the tip is inserted through the anus for screening of diseases such as colorectal cancer. This process of pushing the colonoscope through the colon causes such extreme pain to the patient due to stretching of tissues that anaesthesia is required. This complex process requires highly trained staff thus there is a shortage of expert endoscopists with respect to the demand. Due to these limitations there has been a low early screening of this disease which has reduced the patient’s survival rate.
In order to address the limitations of conventional endoscopy robotic intervention has been developed in the past but the discomfort and pain caused by colonoscopy has not been addressed. The task of using robotic arms manually is difficult and complex which again requires experienced technicians and may often lead to unsatisfactory results.
An international research team, led by researchers from the University of Leeds, has used machine vision to develop intelligent and autonomous control of a magnetic colonoscope which can be easily used by non trained staff to effectively perform colonoscopy. Their work has been published in the journal Nature Machine Intelligence on 12th October, 2020. The research team has developed a semi-autonomous colonoscopy system which is easier to use and is less painful. The team has developed a user friendly system in which the complex manual aspect of endoscope manipulation has been automatized. This would reduce the procedure time and the need for trained operators. The difficult task of manoeuvring the device being handled autonomously by the robot making it easier for doctors or nurses to operate the system means more number of individuals can perform the procedure making colonoscopy widely accessible to patients. Thus ensuring early stage detection of colorectal cancer.
After 12 years of extensive research work the team has developed a magnetic flexible colonoscope which is capable of autonomous navigation and can be handled by non specialized staff. They have exploited different levels of computer assistance in order to improve this complex procedure and reduce the burden on the operator so that they can focus mainly on the clinical aspect of the procedure.
The magnetic flexible colonoscope (MFC) is a capsule shaped, small device attached to a narrow cable which is inserted into the anus. This device is then navigated autonomously by a magnet on a robotic arm hovering over the patient. The magnet on the robotic arm positioned over the patient interacts with the magnets in the small capsule within the patient’s body manoeuvring it through the colon. The research team has claimed that this is less painful compared to conventional colonoscopy.
Manoeuvring the robotic arm manually is a difficult task hence the team has developed different levels of robotic assistance so that non specialized staff can easily use the device. They have developed three levels of autonomy (depicted in the image below).
- Direct robotic control/ no autonomy – The operator has to manually control the robot with the help of a joystick to guide the magnetic flexible colonoscope (MFC) through the colon. In this case there is no assistance from the robot.
- Intelligent endoscope teleoperation- In this level there is robotic assistance with respect to the magnetic manipulation of the colonoscope in order to get the required colonoscope motion. The operator focuses on the location of colonoscope in the colon while the robotic system calculates the movements of the robotic arm, generating a suitable magnetic control, to ensure the colonoscope reaches the desired location in the colon.
- Semi-autonomous navigation- With the help of computer vision the robotic system detects the colon lumen and autonomously navigates the colonoscope through the colon. The operator can override the system if required.
Link to a video depicting the three levels of assistance:
The researchers tested the success rates of the three different levels of assistance in navigating the colonoscope in a laboratory simulation. They asked 10 non specialized staff to get the magnetic flexible colonoscope to a point in a simulated colon within a time span of 20 minutes. They were asked to repeat the task five times using the different levels of assistance. The researchers observed the following success rates:
- Using direct robotic control/ no autonomy the individuals had 58% success rate.
- Using intelligent endoscope teleoperation the individuals had 96% success rate
- Using semi-autonomous navigation the individuals had 100% success rate
After establishing, in the above experiment, that with increased level of assistance there was a higher success rate (improved ease of use and performance) the researchers moved to an in vivo study to test the system. Researchers asked two participants to navigate a conventional colonoscope into the colon of two anaesthetised pigs (animal model). Then they were asked to repeat the activity with the magnet controlled robotic system with the help of the three levels of assistance. The researchers observed that the participants found it easier to navigate the colonoscope with increased robotic assistance. The research team has also claimed that the participants found it challenging and frustrating to use the conventional colonoscope and the direct robotic control/ no autonomy system.
According to the research team in this ingenious work they have been able to simplify and reduce the complexity and challenges of colonoscopy. As claimed by the researchers they have automatized the manual aspects of navigation of colonoscope thereby enabling the operator to fully focus on the clinical aspect of the procedure. The robotic assistance can now help non trained staff to perform this procedure. The research team wants to begin the clinical trial of this novel system in 2021 or in early 2022.
Using intelligent and autonomous navigation system to manoeuvre the colonoscope in a complex environment, like colon, the team has revolutionized colonoscopy. The technology developed for colonoscopy by the team can also be applied to other endoscopic devices for navigation through complex and unstructured environment such as lungs (bronchoscopy), digestive tracts (gastroscopy), pancreas (pancreatic endoscopy) and even catheters.
Using robotics, computer vision and advanced control the team has developed a painless colonoscopy technology that can be easily used for the crucial early detection of colorectal cancer in order to save lives. This advanced technology would also reduce dependency on manual expertise. By fusing magnetic and visual feedback system the team has developed a user friendly system compared to the conventional colonoscope. The team is optimistic that this innovation can make colonoscopy widely available as it can be easily carried out in clinics and healthcare centres and not just hospitals. Colorectal cancer, the third most common malignancy worldwide, is diagnosed late due to the demand for colonoscopy surpassing its availability. This scientific breakthrough of developing a painless colonoscopy which can be widely available could be the dawn of a new era for early diagnosis of the malignant disease.
James W. Martin, Bruno Scaglioni, Joseph C. Norton, Venkataraman Subramanian, Alberto Arezzo, Keith L. Obstein, Pietro Valdastri. Enabling the future of colonoscopy with intelligent and autonomous magnetic manipulation. Nature Machine Intelligence, 2020; DOI: 10.1038/s42256-020-00231-9
Available from: https://www.nature.com/articles/s42256-020-00231-9