A Soft Endorobot (SoftEn) for autonomous colonoscopy screening and treatment
Lead Research Organisation:
University of Dundee
Department Name: Imaging Science and Technology
Abstract
Colorectal cancer (CRC) is the third cause of cancer death worldwide. In 2018, about 1.8 million new cases were reported worldwide with a mortality of almost 900,000. This has increased more than 30% since 2012. Studies have shown that regular screening and early detection can reduce mortality by up to 70%.
The proposed project will undertake interdisciplinary research to design a low-cost, disposable, autonomous Soft Endorobot (SoftEn) to examine the lower intestine painlessly. SoftEn offers disruptive potential to replace the current optical colonoscopy (OC), including a dual capability of investigation and performance of autonomous surgical tasks. SoftEn will overcome OC's limitations (operator-dependent, patient pain/discomfort, high costs, decontamination). Crucially, the high level of autonomy will reduce the time in performing the procedure as well as shift screening process from secondary to intermediate or primary care, thus saving precious time for diagnosis and intervention.
Advantages of this approach are:
1. Efficiency of the procedure will be augmented by reducing the human-operator skills required to perform the procedure. This will increase the number of procedures performed every year.
2. Reducing the waiting list will consequently improve the CRC stage of detection and will increase the patient survivability due to early intervention and reduced cost of treatment.
3. Less pain and discomfort will increase the patient acceptability of the screening as well as patient experience. This has the potential to make sedation unnecessary, which in turn would make hospital visits shorter.
4. Autonomous robot examination of the large intestine would remove restrictions of the traditional operator-dependent OC and would enable an increased access to investigation. Optical diagnosis and tissue diagnosis could be achieved in most of the procedures, including patients requiring more advanced intervention.
5. Potentially improve screening uptake through reducing barriers to participation.
6. The use of polymers implies a reduction in production costs, allowing the device to be single-use and avoiding additional costs required for a traditional OC, including sterilisation and reprocessing. This will reduce the UK cost in CRC procedures.
7. AI software will allow the endorobot to perform procedures autonomously. A clinician can therefore supervise several devices and intervene only if needed via an external ergonomic control console, locally or remotely. This will limit the use of clinicians' time and increase the number of investigations performed per day. Additionally, more clinicians will be able to perform the procedure supporting the national endoscopy workflow.
The proposed project will undertake interdisciplinary research to design a low-cost, disposable, autonomous Soft Endorobot (SoftEn) to examine the lower intestine painlessly. SoftEn offers disruptive potential to replace the current optical colonoscopy (OC), including a dual capability of investigation and performance of autonomous surgical tasks. SoftEn will overcome OC's limitations (operator-dependent, patient pain/discomfort, high costs, decontamination). Crucially, the high level of autonomy will reduce the time in performing the procedure as well as shift screening process from secondary to intermediate or primary care, thus saving precious time for diagnosis and intervention.
Advantages of this approach are:
1. Efficiency of the procedure will be augmented by reducing the human-operator skills required to perform the procedure. This will increase the number of procedures performed every year.
2. Reducing the waiting list will consequently improve the CRC stage of detection and will increase the patient survivability due to early intervention and reduced cost of treatment.
3. Less pain and discomfort will increase the patient acceptability of the screening as well as patient experience. This has the potential to make sedation unnecessary, which in turn would make hospital visits shorter.
4. Autonomous robot examination of the large intestine would remove restrictions of the traditional operator-dependent OC and would enable an increased access to investigation. Optical diagnosis and tissue diagnosis could be achieved in most of the procedures, including patients requiring more advanced intervention.
5. Potentially improve screening uptake through reducing barriers to participation.
6. The use of polymers implies a reduction in production costs, allowing the device to be single-use and avoiding additional costs required for a traditional OC, including sterilisation and reprocessing. This will reduce the UK cost in CRC procedures.
7. AI software will allow the endorobot to perform procedures autonomously. A clinician can therefore supervise several devices and intervene only if needed via an external ergonomic control console, locally or remotely. This will limit the use of clinicians' time and increase the number of investigations performed per day. Additionally, more clinicians will be able to perform the procedure supporting the national endoscopy workflow.
Publications
Mathew A
(2023)
Self-supervised monocular depth estimation for high field of view colonoscopy cameras
in Frontiers in Robotics and AI
Asgari M
(2023)
A review on model-based and model-free approaches to control soft actuators and their potentials in colonoscopy
in Frontiers in Robotics and AI
Description | 1. The use of a double-balloon soft endorobot to inspect the large bowel and to deploy a full HD camera and onboard instruments inside the colon has been successfully demonstrated in a "colonoscopy training model". These preliminary experiments have shown lack of loop formation stretching of the colon due to the self-propelled locomotion, and lack of external pushing force. The internal locomotion force was able to overcome the tether drag with no need for an operator to push the tether. 2. The two balloons are connected by a novel 3 degrees of freedom (DOFs) soft pneumatic actuator (SPA) with a cavity of 8 mm. In current SPA technology, the cavity is squeezed during activation of pneumatic chambers, causing friction of cables hosted inside, with limitation in the controllability and manoeuvrability, due to the stick-slip effect. SoftEn's novel SPA design overcomes this limitation, keeping a constant diameter of the internal cavity, performing a precise control of the camera and instrumentation. 3. A new artificial intelligence technique for video processing in real time, enabling a small autonomous robot to move in a human colon. The technique is based on supervised learning, (learning from labelled examples, i.e., for which the answer is known) but can also make use of unlabelled data. 4. Exploration of real-time lumen detection on video feed from the robot using a semi-supervised deep learning-based framework with a student-mentor scheme training to leverage unlabelled data. The developed system achieved detection at 16 frames per second, with little improvement resulting from the student-mentor scheme. 5. Development of a depth estimation system based on deep learning to provide a local 3D reconstruction of the internal colon wall from the video feed. This is based on improving recent work with self-supervised single-camera depth prediction adapted to the context of colonoscopy, such as considering the large field of view of the camera and masking out problematic areas for the automatic vision system. The system was trained mostly from simulated data, but a dataset was also acquired using a colonoscopy training phantom in order to conduct more realistic experiments. This dataset showed limitations due to the lack of texture and limited camera motion within the phantom however, allowing us to identify the specifications for a better simulator including realistic deformations of the colon. The prototype developed was further extended to embed the lumen detection module, creating a network capable of estimating distances and segment video frames into meaningful regions (w.g., colon wall, lumen). This was achieved by designing two networks working together by exchanging information, resulting into large improvements. This system runs at high framerate and with high accuracy. 6. Experiments in Thiel cadavers were performed. However, the use of standard equipment to inflate and expand the colon was not effective, limiting the locomotion of the device. This is most likely due to the much higher internal pressure needed to expand the colon in a cadaver compared to the in vivo procedure. |
Exploitation Route | 1. The use of a soft endorobot can be translated into clinical practice, increasing the efficiency of the procedure by reducing the human-operator skills. In turn, this will increase the number of procedures performed every year. 2. The double balloon technology can reduce pain for patient and increase acceptability. 3. The use of soft and recyclable materials can reduce production costs, making the device single-use. This can impact on cost for health centres while avoiding disease cross contamination. 4. The vision system capabilities can be built upon to achieve automous analysis of the colon, for instance adding lesion segmentation to the symbiotic network. This would enable automatic estimation of lesion volume and their severity. The vision system could also be trained to be used with other endoscopic imagery, such as from the upper tract. 5. The vision algorithm developed for real-time 3D reconstruction makes the use of 3D cameras unnecessary, thus reducing size and costs of the medical device. 6. Finding of this project can be used for commercialisation by the within the excellent entrepreneurial environment at UoD. Dr Manfredi & Prof Trucco are already involved in spinouts creation. |
Sectors | Healthcare |
URL | https://miccai2022-isgie.github.io/#contact |
Description | Workshop Hamlyn Symposium, 2023 |
Organisation | University College London |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | Organisation of a workshop. |
Collaborator Contribution | Contribute to organise a workshop at the Hamlyn symposium. |
Impact | Workshop at the Hamlyn Symposium, 2023 |
Start Year | 2022 |
Description | Workshop MICCAI, 2022 |
Organisation | Johns Hopkins University |
Country | United States |
Sector | Academic/University |
PI Contribution | Collaboration to organise a Workshop |
Collaborator Contribution | Collaboration to organise a Workshop at MICCAI conference, Singapore, 2022 |
Impact | Organisation of a Workshop |
Start Year | 2022 |
Description | Workshop MICCAI, 2022 |
Organisation | National University of Singapore |
Country | Singapore |
Sector | Academic/University |
PI Contribution | Organisation of a Workshop |
Collaborator Contribution | Collaboration to organise MICCAI Workshop. |
Impact | Organisation of workshop at MICCAI conference bringing medical expertice in the pannel of the chairs. |
Start Year | 2022 |
Description | Workshop MICCAI, 2022 |
Organisation | Olympus |
Country | Global |
Sector | Private |
PI Contribution | Sponson by Olympus Corporation |
Collaborator Contribution | Olympus has sponsored $1000 for the best paper award. |
Impact | Organisation of a workshop. |
Start Year | 2022 |
Description | Invited Talk, EAES, Barcelona, 27th November 2021 |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Other audiences |
Results and Impact | Endorobots for colonoscopy: design challenges and available technologies, Invited Talk, EAES, Barcelona, 27th November 2021 |
Year(s) Of Engagement Activity | 2021 |
Description | Invited Talk, Grand Rounds, University of Dundee, 2 June, 2022 |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | Regional |
Primary Audience | Other audiences |
Results and Impact | Soft robotics for colonoscopy: design challenges and commercial opportunities, invited talk at Grand Rounds, University of Dundee, 2 June, 2022 |
Year(s) Of Engagement Activity | 2022 |
Description | Invited Talk, Hamlyn Symposium, London, 29 June, 2022 |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Other audiences |
Results and Impact | Requirements for Flexible and Intraluminal Robotics, Hamlyn Symposium, London, 29 June, 2022 |
Year(s) Of Engagement Activity | 2022 |
Description | Invited talk @ UCL, February, 2023 |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | Local |
Primary Audience | Postgraduate students |
Results and Impact | Soft robotics for colonoscopy, invited talk at UCL, February, 2023 |
Year(s) Of Engagement Activity | 2023 |
Description | Invited talk, Exeter Symposium, 21 June, 2022 |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Other audiences |
Results and Impact | A Cost-Effective Smart Soft-Endorobot for Colonoscopy, invited talk at the Exeter Symposium, 21 June, 2022 |
Year(s) Of Engagement Activity | 2022 |
Description | Keynote Lecture, Spanish Association of Laparoscopic Surgery and Robotics, Spain, 18 October, 2022 |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | Smart materials to improve dexterity in Robotics Flexible Endoscopy, invited talk at the Spanish Association of Laparoscopic Surgery and Robotics, Spain, 18 October, 2022 |
Year(s) Of Engagement Activity | 2022 |