Robotic revolution: medical robotics operate with precision
At the cutting edge of this research is Kit Hang Lee (Croucher Foundation Scholarship 2015) who is a PhD student of Mechanical Engineering at the University of Hong Kong. Lee’s interested in robotics stems from his younger self’s fascination of the possibility to bring machines to life - from simple toys that he enjoyed as a child, to the sophisticated robotics that he encounters today. For his research, Lee uses high-performance computing to process medical images and robotic control in real time.
Much of the advances in medicine and inventions in modern surgery have been achieved with reducing human pain. Today, image-guided techniques allow early diagnosis and area-specific treatment. Scientists around the world are exploring ways to reduce the pain caused by invasive surgery and complicated treatment. The use of three-dimensional biomedical intra-operative images with the assistance of robots is an emerging field for minimally invasive surgery.
Image-guided surgery and medical robots
Common image-guided intervention includes X-ray, computerised tomography, MRI (Magnetic Resonance Imaging) and ultrasound. Medical imaging is a technique where doctors are aided by visual representation of internal organs to perform diagnosis, biopsies, radiation treatment and delivery of a drug. While the image provides physiological information of the anatomy and the location of surgical site, access often requires complicated incisions, that if inaccurately performed, could inflict long-lasting damage on patients.
“Robots can minimise the invasiveness of surgeries. For example, we are able to make robots significantly smaller than a human hand, enabling us to operate with much smaller incisions,” explains Lee. “If we can get access to, and perform the required manipulation with a smaller incision, risks to the procedure are reduced, and patients are able to recuperate faster.”
Motion scaling is another advantage of robotic intervention, which reduces the potential for errors during surgery, by allowing higher magnification of the targeted area. At the same time, surgical robots also perform tremor filtration, which reduces the shaking of a surgeon’s hand, thus enhancing performance.
Robotic systems for Magnetic Resonance Imaging
Lee’s work mainly focuses on designing robotic systems for MRI guided intervention. One of the most powerful imaging techniques, Magnetic Resonance Imaging (MRI) makes detailed 3D pictures of the anatomy and physiological processes of the body, crucial in the diagnosis of a disease or a condition, without using radiation or X-ray. MRI machines produce a strong magnetic field, and the scanner is restricted in area, which is why its use is still limited for diagnostic purposes.
“I’m looking to design a new kind of robotic device which can be dedicated to an MRI environment, in order to utilise its imaging advantages to the fullest,” said Lee, whose team have made a hydraulic system for robots, to allow high precision and performance inside the MI scanner. The outcome of this is an MRI robot which is able to accurately treat cardiac arrhythmia.
The conventional treatment for cardiac arrhythmia, or an irregular heartbeat, includes the delivery of drug or heat energy to the affected tissue through a catheter inserted in the atrium. The new surgery, aided by MR conditional robot, instead uses a needle to provide fast and high-contrast soft tissue images, without emitting radiation, and during ablation, physiological changes in the tissue are detected by intra-operative MRI with which physicians can easily monitor progress. This reduces the chances of arrhythmia reoccurrence. The robot won the best live demonstration prize in the Surgical Robot Challenge 2016.
Lee explains that the first step is image processing: MR images usually have low resolution, since they are obtained online during surgeries, making them unusable for surgeons. Patients also undergo a preoperative scan. As these scans are done over a longer period of time, the resulting preoperative pictures are higher resolution, however, they cannot be used either because typically pre-operative scans are carried out one day before the surgery, so there is a risk that the tissues might have already deformed or changed location.
One of Lee’s focus areas is intra-operative MRI, which gives exact and updated locations during surgery, after employing a high performance computing technique. This allows for better surgical maps and qualitative measurement for accuracy during the procedure. Accuracy at the millimetric level is particularly important and challenging in treatment of conditions like cardiac arrhythmia, where a catheter is inserted inside a beating heart.
Humanoid robot
After his undergraduate studies, Lee and his team joined a robotic challenge competition organised by the US Defence Advanced Research Projects Agency, which drew inspiration from the Fukushima disaster. In the aftermath of the 2011 nuclear accident, technicians were required to go inside the building and fix the settings on power plants to prevent further destruction. However, the radiation was too strong and harmful for humans to enter. This has led governments to explore robotic options in disaster response.
Soft robotics
While robots do not cease to amaze, a typical robot structure is made up of rigid parts and motors made up of metals, which is largely unsafe and can easily crush humans. Scientists are now starting to search for soft materials to replace the heavy and dangerous structures.
Using compliance control, Lee’s team was able to demonstrate how an ATLAS robot could clamp an egg without breaking it. The team input exact calculation and compliance on the robot’s arm that limits its turning force so that the motion was smoother and did not break the egg. Compliant control limits the contact force that is applied on the egg, so that even powerful robots can handle delicate items.
Like in any other sector, soft robotics has its useful applications in medical science. Soft structures would allow for higher compliance, are squeezable, stretchable and nearly unbreakable, all while being much safer, says Lee, who is also researching on soft robotics. One of the main materials used in soft robots is silicon rubber, which is much cheaper than metals and can be used for the production of disposable robots. “Disposable robots are particularly useful for medical applications. Having plenty of disposable robots mean we don’t risk cross-infection,” points out Lee. Endoscopes are usually made from metal and are expensive. It has been proven that there is no way to fully sterilise the scope, making cross-infection a risk.
However, there are limitations. While soft materials are stretchable, scientists have yet to discover a sensor that can be stretched as much in order to learn about robot deformation. At the same time, given its flexibility and inherent complaint nature, it is also difficult to predict the motion of soft robots, making the computation more complicated. “My study on the soft robotics is on the control side. Since its mathematical representation is difficult, it’s hard to predict how it deforms. So I propose machine learning to study about its deformation and motion mapping,” said Lee.
Challenges
Lee explains that sometimes researchers are prone to over-engineering and making devices with applications not practical for doctors. “We make medical devices, but we do not use them on patients, only surgeons know best what they need and the actual challenges they face.” At the same time, there have been instances where necessary technique and devices for novel treatment were easily accessible, but because surgeons didn’t know, they resorted to traditional equipment. “By working with surgeons and doctors, we can come up with ideas that can change and revolutionise the whole field, so we need more collaboration and communication between the clinicians and engineers,” concluded Lee.
Kit Hang Lee is a PhD student of Mechanical Engineering at the University of Hong Kong (HKU). He obtained his bachelor’s degree in Mechanical and Automation engineering in 2011 from the Chinese University of Hong Kong. During his undergraduate study, he participated in a one-year exchange program hosted by Technical University of Denmark. Lee was awarded with a Croucher Foundation Research Scholarship in 2015 for his study in the field of image-guided robotic interventional systems. He was previously a research assistant at Advanced Robotic Laboratory at HKU and at Department of Mechanical and Automation Engineering at Chinese University of Hong Kong in 2014 and 2013 respectively.
To view Lee’s Croucher profile, please click here.