Sun Microsystems helps make a difference

"You only live once, so it's great if you can use that chance to make a difference."

Making a difference is an important criterion for Dr. Randy Ellis in the work he does and in the pursuits he follows. It's also the net effect of his efforts at Queen's University and Kingston General Hospital in computer-assisted orthopedic surgery. Through his efforts and those of his colleagues in orthopedic surgery, Dr. Ellis has harnessed the power of Sun technology and helped transform orthopedic surgical planning, preparation and performance into a process that significantly enhances the outcome for patients.

Orthopedic surgery is not necessarily the first place that one would expect to find computer science. It also wasn't the first path of study that Dr. Ellis had chosen.

After completing doctoral studies at the University of Massachusetts, Dr. Ellis came to Queen's University to act as a Professor in the Department of Computing and Information Science in 1988. Dr. Ellis was continuing his research activities and in particular was fascinated by the challenge posed with enhancing the performance of functional joints in robotic technology. The discipline of kinematics helped Dr. Ellis to evaluate the performance, design and adaptation of the joints used to enhance the mobility and movement of robots.

One day in 1993, he was chatting with a doctoral student in the coffee room of the Department of Mechanical Engineering (he is a cross-appointed member of that Department, and is now also in the Department of Surgery). Ellis was describing the process he conceived for mapping joints and their functionality, and the conversation was overheard by Dr. Tim Bryant. The suggestion was made that there would be considerable value in applying this concept to orthopedic surgery.

"Dr. Bryant, who directed the Clinical Mechanics Group, overheard me talking about how we could use computing technology to detect early problems in robotic joints," said Ellis. "He asked me if I had thought of applying that approach to orthopedic surgery on joints and it all started from there."

Ellis recognized that there were already examples of how computing technology had been successfully applied to improve the diagnostic imaging procedures involved in orthopedic treatment. He did not, however, know how this could be applied to the particular needs of surgery and concluded that he needed to learn more about the specific challenges faced by orthopedic surgeons.

Following completion of his semester of teaching and research the next year, Ellis went to the Rizzoli Orthopedic Institute, a leading orthopedic clinical and research facility in Bologna, Italy where he was given the position of Visiting Research Professor. In the six months Ellis spent there, he successfully immersed himself in the world of orthopedic treatment and surgery and developed a first-hand understanding of the regular challenges faced by surgeons seeking to improve the performance of human joints through orthopedic surgery.

"It became clear to me that one of the major challenges faced by surgeons in this field was the ability to link the planning of procedures with the execution of surgery," he notes. "The use of CT scans gave surgeons valuable images of the physical problems represented in individual patients, but that image was only useful on a planning level. Once surgery began there was no direct ability to correlate that image and the planned procedure with the execution of the surgery in the operating room."

What Ellis proposed to do was to not only utilize the image data collected from a CT scan but to also use advanced computing power to initiate a process called registration. In this approach, a diagnostic surgical tool is used in the operating room by a surgeon to map out the specific surface contours of joints, such as knees or wrists, by placing the instrument on the joint surface and using that information to provide specific and detailed recommendations to the surgeon on where to apply an incision in the skin, muscle and bone mass. The goal is to minimize the scope of the surgery and limit it to the specific area of need thereby maximizing the success of the surgery and minimizing the negative side effects of an invasive procedure.

This of course requires significant computing power to dynamically assemble the extensive data collected during the registration procedure and to successfully interpret the data in a three dimensional context. Preliminary procedural plans and images must then be referenced against this data and interpreted to allow surgeons to ensure that their actions are following the plans accurately during the operation. "Unless you can correlate the reality of the operation with the previously acquired CT scan images of a joint, all you really have is nice images that don't directly benefit you in the operating room while you are performing the procedure," says Ellis. "Computer aided registration allows the images of the CT scan and the associated procedure plans to give surgeons real direction and assistance at the time and place they need it most."

Research into surgical computing resulted in Kingston General Hospital's, Dr. John Rudan performing Canada's first computer-guided total knee replacement in July 1997 and Dr. David Pichora performing the first computer-guided wrist repair in August 1998 and the results have clearly demonstrated the value of the approach. Surgery that had once required more than two hours and lengthy healing and rehabilitation sessions could now be completed in half the time with the patient returning home the same day. More importantly, only the absolutely essential removal of bone was required, meaning that screws, plates and large incisions had become a thing of the past. Patients now left the hospital later that day with nothing more than a Steri-Strip -- a high tech band aid -- to show from their operation.

This leap forward in surgery has been accomplished by orthopedic surgeons at Kingston General Hospital in part through their reliance on technology from Sun Microsystems. Initially Dr. Ellis had been utilizing an SGI workstation, placed on a rack and moved in and out of the operating room at KGH as needed. In 1997, when Dr. Ellis first connected with Doug Girvin of Stantive Solutions, Sun's Authorized Independent Marketing Organization in Eastern Ontario, the Sun alternative became real.

"I was fascinated by the work that Dr. Ellis was doing. I went directly to Sun Microsystems and applied for an Academic Equipment Grant for a Sun Ultra workstation," said Girvin. "That donation was the foundation for a long term relationship that has truly helped to advance the important work that Dr. Ellis is doing."

Starting with the adoption of the Sun platform, Ellis began to recognize an improved performance in graphics generation and platform stability. In addition, Ellis recognized that both Sun Microsystems and Stantive Solutions had taken a direct interest in supporting the realization of his research goals. As Ellis and others successfully negotiated grants from the Canadian Foundation for Innovation and the Ontario Research and Development Challenge Fund, his relationship with Sun expanded and so did the composition of his Sun environment which now includes 3 Sun Blade 1000 workstations, 7 Ultra 10 workstations, plus a Sun Enterprise 220R server with 2 Sun StorEdge A1000 disk arrays.

This collaboration of scientists is part of a larger research initiative called the Human Mobility Research Centre. The Centre focuses on musculoskeletal research and is an example of the synergism that exists between KGH, the Queen's University Faculty of Health Sciences and other sectors of the university, says Dr. Sam Ludwin, Associate Dean (Research) Queen's and Vice President Research Development at KGH.

In addition, the work being conducted at Kingston General Hospital is successfully leveraging the computing resources of the High Performance Computing Virtual Laboratory (HPCVL) formed by a consortium of four universities in Eastern Ontario. With its main facility at Queen's University, this Sun Center of Excellence in Secure Grid and Portal Computing serves the research and educational needs of a variety of disciplines at Queen's, The Royal Military College of Canada, Carleton University and The University of Ottawa. By having access to HPCVL, Dr. Ellis' project at Kingston General Hospital has significantly expanded its progress in the pre-operative planning and software research and development components of his work.

"In the operating room I may need to quickly regenerate the CT images as part of the operating procedure," says Ellis. "Doing that within computing capacity available in the OR could take hours. By connecting into the Sun equipment at the HPCVL, the same tasks can be accomplished in seconds."

"One thing you need to be certain of in computer assisted surgery is the reliability of your computing power. You can't afford to have a failure or delay," says Ellis "Sun delivers. Robust and reliable are terms we take very seriously in the operating room and Sun Microsystems technology has proven itself on both counts."

The computer-aided operating room at KGH is unique in Canada and puts the facility in some pretty rare company in North America. MIT, Harvard, Johns Hopkins and Stanford University are the only other facilities that have so far adopted similar approaches. Even within that group, Dr. Ellis perceives that KGH has some advantages.

"The great thing about the relationship between Queen's and Kingston General Hospital is how well it works," says Ellis "I know that the relatively close knit nature of the community we have here in Kingston works in our favour and has helped in strengthening the collaboration between the computer scientists at Queen's and the orthopedic surgeons at the hospital."

In fact, the adoption rate has been so successful that seven of the nine orthopedic surgeons at KGH used the facility in 2001. The wider use of the technology promises to broaden the benefit to patients in a significant way.

It is the desire for broader adoption that now occupies Dr. Ellis. Because the present process requires individual CT scans for each procedure, the availability for surgery in this approach can serve to minimize its impact in the patient population. Dr. Ellis is, however, evaluating an alternative that would have standard models of human joints, such as knees and wrists, that could be created by amalgamating the data from a large number of CT scans. That information could then reside on the Sun servers and be used as a common reference point to use in individual patient registrations in the operating room.

"The true potential of computer aided surgery could best be realized if the need for CT scans were reduced to exceptional cases," said Dr. Ellis "The truth is that most people have remarkably similar joint shapes, so a standard model that accounted for variables in each registration would allow for more hospitals to adopt this approach and eliminate the need for CT scans except in unusual circumstances."

Ellis calls this technique morphing because it represents an adaptation of similar processes used in computer animation with powerful servers. In this example, however, the digital image of the joint is morphed from the standard model and adapted to the specific information gathered during the registration process.

Since November 2001, his team has used morphing to replace knee joints in arthritic patients. No CT scans were needed and orthopedic surgeon performed the operations using the smallest possible incisions. The morphing method represents a major achievement, as the computer-aided operating room can now become a reality in almost any facility where orthopedic surgery is performed. Given the dynamic improvement that has been shown in patient outcomes through computer-aided surgery, broader adoption would represent a significant benefit to patients across North America and the world.

The name that Ellis and others have adopted for his project is "OR/2010", in reference to the concept that this facility should represent the standard of orthopedic surgical facilities within ten years. Based on the progress made today, Dr. Ellis is optimistic about the realization of this goal.

"We are making significant progress at Kingston and have charted a path which promises to make a real difference globally in the practice of orthopedic surgery," says Dr. Ellis. "Broader adoption will ultimately make a real difference to patients, and making a difference is what all of this is about."