"The future of surgery is not about blood and guts; the future of surgery is about bits and bytes.”
/Dr. Richard Satava/

Sunday, May 31, 2009

MR robotics at AIST

After the ICRA conference, we managed to get an exclusive tour to the AIST lab's main campus in Tsukuba, north from Tokyo.
AIST (National Institute of Advanced Industrial Science and Technology) is one of the biggest research institutes in Japan. It’s spread at 10 locations, with headquarters in Tokyo and Tsukuba. There are approximately 2500 full-time researchers, young professionals are only involved as post-docs and above. Their annual budget is around 100 billion yen (1 billion USD). AIST’s mission statement is to provide:
- Contribution to a sustainable society
- Contribution to industrial competitiveness
- Contribution to local deployment of industrial policies
- Contribution of policy-making of industrial technology
- Contribution to development of human resources skilled in technological management
AIST research portfolio is organized to seven branches from nanotechnology and materials to metrology and measurement science. There are 22 institutes, 23 research centers, and a few research initiatives, cores and centers for intellectual infrastructure.
We visited two groups (out of the 14) within the Institute for Human Science and Biomedical Engineering that aims to promote healthy longevity and a higher quality of life through development. Even though Japan has not really embraced yet surgical robotic technology, they still spend significant money on the topic, and due to their superiority in robotics, achieved great results. Currently there are only 5 da Vinci robots in Japan, all in research labs, as the public health care system does not allow for their easy financial integration. (3/4 of all procedures are payed by the government.) Rumor has it Intuitive is applying for the health authorities' approval for the robot, to be able to sell it for private clinics.
First, we were given a tour by Dr. Chinzei, leader of the Surgical Assist Technology Group. Their main focus is on improving image guided procedures, microsurgery, small scale haptic sensors and surgical performance evaluation. In the mid 1990’s they were among the firsts to develop MR compatible robots. Their first prototype was meant to fit into GE’s 3T double donut scanner. However it was still too expensive and bulky to get beyond clinical trials, and also GE discontinued that product line. (A few devices are still for sale.) The next generation of the robot used an inverted RCM mechanics with a parallel robot. Their newest version is a little smaller, 5 DOF needle insertion assisting robot (for endonasal procedures), integrated with a flat- Mitsubishi scanner. They are also developing an MR robot for micromanipulation, a force sensor for extended needle tip sensing and working together with another group on surgeons’ skill evaluation
Next, we visited the Skill Research Group, were they have developed a very detailed plastic skull phantom for endonasal training. They are experimenting with telementoring and effective teaching and evaluation through video-feedback. As for skill evaluation, they measure the applied forces to assess the possible damage caused. This could help to better understand the details of surgical procedures.
Third, we met some French people from ICRA, and they showed us their joint AIST/ISRI-CNRS/STIC lab, within the Research Division of Advanced Robotics and Cybernetics. They are working on human robots, developing hardware for the next generation of the HRP-2 robot (currently industrial secret), and improving the motion and cognitive skills of the existing version.


This year’s IEEE International Conference on Robotics and Automation (ICRA) was held May 12-17. in Kobe. Although the main focus of the conference is still localization, motion control and bio-inspired robotics, medical robotics and haptics received significant attenton, and the rooms of these sessions were full all the time. Beyond Japanese, German people were dominant; mostly form TU Münich, DLR and Freiburg. As for the USA, Hopkins (the haptics lab), Stanford and CMU had more significant presence. The invited lectures, plenary talks, industrial forum, robot challenge went in the regular way, there were no relevant topics. (The most interesting plenary was from Prof. Nayar on imaging and the future of digital photography.) The haptic sessions had most topics with human feedback and needle procedures, while medical section had a wide variety of topics. (Publications are all available through IEEE Xplore.) Here, only a few major projects presented at ICRA are highlighted:

- The robotics people at DLR under Gerd Hirzinger have long been doing world class robot design and building. Their latest system, the MIRO is based on the light-weight DLR Arm III, and should be down on the path to commercialization soon. It integrates torque sensing capabilities on the joint level to enable close interaction with humans in unstructured environments. The complex joints consist of an actuation module, position sensing module and a torque sensing module (based on aluminium parts with mounted strain gages, capable of oversampling). The robot can run in torque and impedance control mode, where the user can guide the robot to a desired position or on a desired trajectory by hand (“gravity compensation mode”). Virtual springs (virtual potential fields) are used to impose constraint forces which prevent the robot from entering predefined areas. Due to the kinematical redundancy of the manipulator, the inner joints of the robot can be controlled in compliance mode separately, i.e. reposition themselves in the case of collision or other force effects. As the force readings are integrated to the low level control loop, this gives additional safety to the system working in close cooperation with humans. The total surgical setup consists of 41 DOF.
- Greg Fisher (Hopkins graduate) has his lab now at Worcester Polytechnic Institute, and continues to design and build innovative robotic systems for MR percutaneous therapies. Their latest device is meant for accurate electrode placement in deep brain stimulation. The robot performs the insertion under real-time 3T MR image guidance. This allows for the monitoring of tissue deformation, real-time visualization of insertion, and confirmation of placement. The current prototype is mounted on a 3 DOF platform, has 2-DOF RCM motion and a yoke that gives 2 additional DOF. They will begin soon the validation experiments including the accuracy and MR safety tests.

- An interesting concept was presented from Paolo Dario’s lab: self assembling swallow-able robots for NOTES. Each segment of the robot can perform 2 DOF motion, (±90° of bending and 360° of rotation), measuring 15.4 mm in diameter and 36.5
mm in length. The modules attach to each other through magnetic connectors, allowing for many different configurations in the stomach for surgery. The current prototype does not have wireless connection yet, and the optimal strategy for disassembly and retrieval is still under consideration.
- Most of the forty-some exhibitors were hardware manufacturers or book publishers, however there was one interesting booth, representing the ROBOCAST. The ROBOCAST project is an EU FP7 funded international research that „focuses on robot assisted keyhole neurosurgery. Related pathologies are tumors, hydrocephalus, dystonia, essential tremor, Parkinson’s Disease, Tourette Syndrome, clinical depression, phantom limb pain, cluster headache and epilepsy. The ROBOCAST project outcome will be a system for the assistance of the surgeon during keyhole interventions on the brain. It will have a mechatronic part and an intelligence part. The mechatronic device will consist of a robot holding the instruments for the surgeon and inserting them in the brain with a smooth and precise controlled autonomous movement. The trajectory will be defined by the intelligence of the ROBOCAST system and will be approved by the surgeon, which is and remains the responsible of the outcome, before the insertion of the surgical instruments.” It began in 2007 and now they are getting ready to the first clinical trials with a prototype.

Friday, May 15, 2009

Workshops at IEEE ICRA 2009

These days the IEEE ICRA conference was held in Kobe, Japan, featuring many great professionals of the field from around the world. As usual, several topical workshops preceded the conference. The ICRA Workshop on Innovation in Medical Robotics, organized by Mamoru Mitsuishi (University of Tokyo), Makoto Kaneko (Osaka University) and Alois Knoll (TU Münich) featured many interesting presentations e.g. on ultrasonic diagnostics, stiffness sensing and needle insertion.
Talks included the introduction to the latest development of the neurosurgery robot at the Tokyo University, the eye-catching self-assembling robots (ARES) from the Scuola Superiore Sant’Anna, the da Vinci replacement developed at TUM, used for surgical skill evaluation.

The second day Workshop on Advanced Sensing and Sensor Integration in Medical Robotics took place, organized by Darius Burschka, (Lab for Robotics and Embedded Systems, TU Münich), Greg Hager and Allison Okamura (Johns Hopkins University) and Rainer Konietschke (DLR). Most interesting talks covered motion compensation with robotized ultrasound probe, force control for telerobotic surgery and control of flexible robots. I also presented our latest achievements on patient motion tracking.

Friday, May 8, 2009

Surgery in space IV

The final part on the series of post on the feasibility of extreme telesurgery deals with a provisionary concept of scaled telemedicine paradigm supporting long haul human space flights.
Based on the physical conditions, the difficulties listed in the previous post and on the system requirements, three-layered mission architecture is proposed to achieve the highest degree of performance possible, by combining robotic and human surgery. (See picture for more details.)
By adaptively switching, different levels of surgical service can be provided throughout the mission. Mainly within the range of 380 000 km (app. the Earth-Moon average distance), regular telesurgery techniques can be used in space to provide medical support in case of emergency. Leaving the orbit, special control strategies have to be applied, to extend the feasibility of telesurgery up to 2 s of delay. With robot assisted surgery, a shared control approach should be followed, integrating high-fidelity automated functions into the robot, to extend the capabilities of the human surgeon. This concept could be most beneficial for long duration on-orbit missions, primarily on board of the ISS. Presently, there is no other option than the immediate evacuation of the affected astronaut, which poses bigger health risk and costs a lot more. If losing the signal, the integrated robotic system should stop immediately, and the crew has to be prepared to take over the control of the robot, and finish the procedure, if the connection cannot be reestablished. To reduce the frequency of failure, network redundancy is essential as showed by the NEEMO projects.
Flying further from the Earth and having reached the limits of pseudo real-time communication, the procedures should be performed by the flight surgeon, or by any other trained astronaut, under the telementoring guidance of the master surgeons on the ground. As showed by the NASA undersea experiments, telementoring can be an effective alternative to direct teleoperation, allowing the controller to perform the tasks based on the visual and voice commands of the ground centre. With adequate training and practice, the astronauts with a basic surgical training might be able to successfully accomplish complete procedures. Telementoring may extend the boundaries of telepresence, as it can still be effective with a 50-70 s delay (within the range of app. 10 000 000 km). Upon this phase, the built-in semi-automatic functions of the surgical robot may have a significant role to improve the overall quality of the surgery. On one hand, motion scaling, adaptive tremor filtering, the automated following of the organ’s movement, automated suturing could significantly improve the less practiced crew members’ performance, while special security measures could also be applied on the other hand. The setting of virtual boundaries for the robot, tool limitations and speed constraints may reduce the risk of malpractice. Astronauts should also benefit from advanced imaging technologies, as accurately matched anatomic atlases could help their navigation around the organs. With the use of augmented reality systems, real and virtual images can be merged in real time to make the operation even smoother.
There is no sharp limit between telementoring and consultancy telemedicine. Above a certain time delay, the terrestrial medical support crew will not be able to react on time to unforeseeable events during the procedure. By the time they receive the video signal from the spacecraft, the operating environment might have drastically changed, therefore the astronauts should be able to perform the procedure on their own, after having consulted the ground centre. Above approximately one minute of delay, it is inconvenient and impractical for the crew to wait for the guidance of the ground after every step accomplished, and in some cases, it would endanger the success of the operation. The flight surgeon must be trained to conduct the operation and make decisions on its own.
If there is no real-time connection between the spacecraft and the ground control, the terrestrial surgical centre could still run complete surgical simulations. Given the astronauts’ precise 3D model gained for extensive MRI, CT and PET scanning prior to the mission, a variety of operations and possible outcomes could be simulated and analyzed on the ground. Complete risk assessment, identification of bottlenecks and personalized best-practice methods could be evaluated. The condition updates of the ill or injured crew member could be gained from Ultra Sound imaging and other scanning equipment on board, along with the data of biosensor-networks. These are to be merged with the recorded model before to the real operation; therefore the surgeons on Earth could provide a priori results and recommendations in the form of consultancy, prior to the actual in-space surgery.
It was shown during the NEEMO missions that the general performance of the telesurgery is higher than of the telemedicine, and a team of experts may do better than the flight surgeon. Therefore depending on the feasibility, telesurgery should be preferred on telementoring.