音声ブラウザご使用の方向け: SKIP NAVI GOTO NAVI

VIRTUAL REALITY AND HEALTHCARE TECHNOLOGY: THE SHEFFIELD EXPERIENCE

Robin Hollands
Department of Automatic Control and Systems Engineering
University of Sheffield
Sheffield S1 3JD, UK
Voice: +44 114 2225619
FAX: +44 114 2731729
Email: r.hollands@sheffield.ac.uk

Tony Trowbridge
Department of Medical Physics and Clinical Engineering
I Floor, Royal Hallamshire Hospital
Sheffield S10 2JF, UK
Voice: +44 114 2713647
FAX: +44 114 2713403
Email: e.a.trowbridge@sheffield.ac.uk

Justin Penrose
Department of Medical Physics and Clinical Engineering
I Floor, Royal Hallamshire Hospital
Sheffield S10 2JF, UK
Voice: +44 114 2713647
FAX: +44 114 2713403
Email: j.m.penrose@sheffield.ac.uk

Web Posted on: December 2, 1997


1. INTRODUCTION

Virtual reality (VR) is a technology which simulates complex three-dimensional environments in real-time and allows users to interact with them in a intuitive manner. In addition to software which allows the fast manipulation of three-dimensional graphics, a variety of input devices have been developed to monitor anatomical position and physiological status in order to provide a natural interface to the virtual world.

The human body is a complex three-dimensional structure, and medical practitioners have a long history of using three-dimensional visualisation aids to understand it, from simple skeletons to complex 3D computer reconstructions of CT scans. Virtual reality technology allows many computer visualisations to be experienced in a more 'realistic' manner, in full 3D and with the capability to manipulate the data interactively. More importantly, however, the user is no longer restricted to passive visualisations, but instead can use the power of computer simulation to give the objects realistic dynamics, and reproduce 'life-like' behaviours within the computer models.

The Virtual Reality in Medicine and Biology Group (VRMBG) was formed in Sheffield, UK, to exploit the many advantages of VR technology within the fields of medical research and practice. The interdisciplinary group has a wide remit, covering everything from enhancing existing visualisation techniques, to developing virtual reality simulations of biological processes, and using virtual reality interface technology for appraisal or support of physical disabilities. This paper describes a sample of the projects currently underway, and more information can be found on the World Wide Web at http://www.shef.ac.uk/~vrmbg/.


2 KNEE ARTHROSCOPY SIMULATOR

Although extremely common, knee injuries can cause a great deal of distress and debilitation. In the case of professional sports people, an uncorrected knee injury can mean loss of livelihood for a time or the end to a career. In the case of ruptured ligaments, leading to knee instability, physiotherapy can often be employed to train the surrounding muscle structures to support the knee. However, for maximum recovery of knee performance, the damaged ligament must be reconstructed. In other cases, it may be necessary to inspect the interior structures of the knee to ascertain the exact nature of the injury, and there will often be no alternative but to correct the fault by surgery.

Examination and repair of knee disabilities can be performed using open surgery. However, there are many benefits to using the minimally invasive approach of arthroscopy. In addition to the benefits to both patient and health care provider of significantly reduced recovery time, arthroscopic techniques minimise loss of synovial fluid and allow many of the knee structures to be examined in a relatively natural state. However, as with other 'key-hole' type operations, arthroscopy can demand skills in excess of their open-surgery counterparts.

To inspect the inside of the knee, the surgeon uses a miniature fibre optic camera inserted through a small incision in the skin. Miniature, long handled probes and other tools are inserted through another small incision. The picture from the camera is viewed on a monitor, requiring the surgeon to look elsewhere other than at his hands and tools, as in open surgery. In addition, the surgeon must be able to orient himself correctly using the camera view and navigate the tools through the anatomy of the knee to the operation site, without necessarily being able to see them. The camera and tools themselves are only free to move around the fulcrum of the entry point, removing two of the degrees of freedom a surgeon would have with tools in open surgery, and making the tools more sensitive to slight movements, whilst also reducing the sense of touch to the surgeon.

Because arthroscopic knee surgery is considerably more difficult than its open surgery counterpart, the risk of missing an important pathology or actively doing damage due to inexperience is increased. The trainee must be intimately familiar with the topology of the joint, and able to navigate through it without damaging the sensitive articular surfaces. The trainee must also learn to perform the various procedures required whilst removing the minimum of material and ensuring that no debris or foreign objects remain in the joint cavity. Failure to do so can result in the need for further corrective surgery or may even lead to permanent disability in later life.

Current training techniques use a mixture of cadavers and physical models for early training, with the majority of real learning done on live patients. A computer based arthroscopic trainer for knee surgery is being developed at Sheffield as an alternative to traditional training methods in an attempt to ensure that trainee surgeons perform their first surgery on live patients higher on the learning curve than they do at present. The computer contains a geometric model of the anatomy of the knee and a numerical model of the knee kinematics. Connected to the computer is a simple artificial leg with jointed knee, a replica arthroscope and replica tools. The trainee surgeon is free to manipulate the artificial knee in order to expose the various joint cavities, as in real surgery, and the computer monitors the position and orientation of the arthroscope within the knee joint in order to provide a simulated view on the computer display. The replica tools are similarly tracked to allow the surgeon to interact with components inside the joint in an intuitive manner.

Although other surgical simulators have been developed elsewhere, particularly for laporascopic surgery, one of the significant features about the Sheffield simulator is that it based on a standard PC. Unlike solutions that need high cost graphics workstations, the use of a PC should mean that the final system is financially competitive with existing training modalities. The current system is suitable only for basic orientation, navigation and triangulation training, but work is underway to model an increasing number of pathologies and procedures.

Development of the simulator has been supported by Smith and Nephew Surgical who plan to incorporate the final system into their training centres. Medical appraisal of the system is performed regularly by surgeons based at the Northern General Hospital and Royal Hallamshire Hospital in Sheffield. Additionally, the simulator has been presented to many of the UK's leading arthroscopic consultants at a number of medical workshops, and the response has been overwhelmingly positive.


3 HAND DISABILITY APPRAISAL AND MODELLING

An example of virtual reality technology recommissioned for medical use is the SIGMA (Sheffield Instrumental Glove for Manual Assessment) glove being developed at Sheffield. Data gloves were once a common feature of virtual reality systems, monitoring hand shape and gesture to provide 'natural' input to the controlling computer. The need for users to have to learn a system-specific sign language and the general inconvenience of having to 'suit-up' at every session has seen the glove fall out of favour with the general VR community. However, the sensors developed for tracking hand gestures also provide an ideal opportunity for instrumenting the hand for medical purposes.

The sensors used in the SIGMA glove are based on carbon-ink technology used in gloves for low-cost entertainment systems, but have been combined together and processed in such a way that they can accurately monitor extension-flexion and abduction-adduction of fingers, thumb and wrist with 21 degrees of freedom. The SIGMA glove provides a unique opportunity to obtain real-time goniometric measurements for the entire hand during the performance of a series of tasks. The results of these tests could be used, for example, to determine the success of corrective or reconstructive surgery, or to provide assessment of the severity of hand disability for compensation purposes.

Although the current version of the glove is still undergoing medical trials, it's functionality is being extended by developing additional integral sensors to combine the finger position measurements with applied pressure across the surface of the hand and along the length of the fingers. Furthermore, studies are also underway to enhance current hand data with correlated dynamic measurement of EMG signals from extrinsic and intrinsic muscles of the hand.

In parallel with the glove design is the development of an accurate numerical model of the underlying hand anatomy. The movement of the individual segments of the hand is complex, however most models to date represent the joints quite simply, usually as a hinge. Anatomical studies for the development of the model have actually shown that the knuckle joint has more in common with the knee joint that a simple hinge. One of the first benefits to come from this new hand model is the analysis of knuckle prostheses. Incorporation of a finite element model of two knuckle prostheses with the hand model has shown weak points in the prosthetic design. The high stress concentrations in these areas are probably responsible for the prosthetic's limited life span of 3-4 years.

The finite element model and the SIGMA glove can be combined together to analyse the underlying hand function for a given range of tasks, and the results can be visualised in real-time 3D using virtual reality visualisation techniques.


4 GAMMA CAMERA ROOM VISUALISATION

A project which is probably more familiar as 'traditional' virtual reality is the use of the technology for room layout design. Modern hospitals use a large amount of high technology equipment, much of which is large, heavy, or has special requirements. Although a certain amount of planning can be performed using floor plans, the results can never be fully examined until the equipment is actually installed, by which time it's too late for any changes.

At Sheffield, a virtual reality model of the Gamma Camera of the Royal Hallamshire hospital has been created. The room had been chosen previously, and therefore the space available was fixed. Along with the Gamma Camera itself is a large amount of support equipment: workstations, monitors, carts etc. By modelling all of these objects in a virtual environments, various layout options could be examined before the 2-tonne Gamma Camera was installed. The ability to visualise the options as a real-time three- dimensional environment meant that all potential users could be consulted; most of the clinical and technical staff would not normally be able to visualise a two-dimensional floor plan.

In addition to equipment layout, the system was extended to incorporate a human system modelling package to put virtual humans into the virtual room designs. The virtual human package allowed different sizes of humans to be introduced to the room, and could show the view from the virtual human's viewpoint and model the strain the human undergoes when carrying out required tasks.

The advantages of the system were numerous. For the equipment operators, the layout could be designed to optimise day-to-day tasks. The human systems modelling software allowed furniture to be chosen and procedures to be drawn up to minimise the chance of work related disabilities. The same techniques could be used to experience the room from the patient's point of view. Some interesting changes to the proposed room design that occurred as the result of the virtual reality model was to move the Gamma Camera to one side so that it was not ominously silhouetted against the window on the patient's arrival. Similarly, various monitors were moved when the virtual environment showed that the patient might see something upsetting on the display!


5 HEART VALVE SUBSTITUTE VISUALISATION

The wide variety of substitute heart valves available show that the design of the perfect valve has yet to be achieved. Mechanical valve substitutes have a long fatigue life but the central flow occluders often induce blood cell trauma. Tissue valve substitutes have an unimpeded central orifice when open, cause minimal cell damage but have a relatively short fatigue life, especially in children where calcification may be a major problem. In the design stage and before clinical implementation, heart valve substitutes have been subjected to rigorous fluid mechanics testing in the laboratory by using a physical model of the heart called a pulse duplicator.

One disadvantage of the pulse duplicator system is the need for a physical prototype of the substitute valve, and the data obtained from the device can only be interpreted by experts. By combining computer modelling of fluid flow along with a three-dimensional model of the heart valve substitute, it is possible to simulate the effectiveness of the valve in a variety of blood flow conditions. Combination of fluid dynamic modelling of the blood and finite element analysis of the substitute can indicate any sites of likely failure and help design a more durable valve. Coupling the numerical modelling of the blood flow by using computational fluid dynamics with a CAD model of the substitute heart valve also provides a very rapid means of assessing a number of design options, and even provides the opportunity to modify the design 'on-the-fly', which would be impossible with the physical system.

Although the results of the simulation could be analysed using flow diagrams etc., a system developed at Sheffield uses virtual reality techniques to provide the unique perspective of watching the path of simulated blood vessels through the valve, and even provide the opportunity to 'travel' on the back of a blood cell. The intuitive graphical nature of this technique allows for rapid initial inspection of a candidate for more rigorous numerical testing, and the novelty factor provides an ideal opportunity to use the system as an educational aid for the prospective patient.


6 MOTORISED WHEELCHAIR SIMULATOR

Like any powered vehicle, a motorised wheelchair can be difficult to control and potential dangerous to both the user and those around if not controlled correctly. Since a number of wheelchair users may be children or have communication disabilities it is essential to have a simple but safe method of training the user to control the device and assessing their control after the period of training. Traditional methods have relied on using wheelchairs with restricted controls, or remotely controlled chairs equipped with TV cameras. Whilst these methods are fairly effective, they still risk physical damage to the expensive equipment, and need to be used in closely controlled environments. Another problem during training, especially where children are concerned, is the need to keep the subject interested and motivated.

A low cost system being developed at Sheffield models the performance of a wheelchair entirely within a virtual environment. The environment consists of a number of rooms designed to be visually engaging, in which simple tasks of navigation are set and automatically scored. Depending on the equipment available, the virtual world can be experienced in a number of different formats: on a normal computer screen, in full 3D on a large screen or fully immersed with a head mounted display.

Because all action takes place in the virtual world, no person or property is at risk, and loss of control need not be traumatic. Unlike the expensive physical set ups, the minimum equipment required is a home PC and joystick, making the tool widely accessible.

The system is still under development and currently only uses a relatively simple model of wheelchair movement. However, it has proved sufficient for initial control studies and is being further developed for use as a test bed for different wheelchair control devices.


7 CONCLUSIONS

This paper has shown a brief overview of the activities of the Virtual Reality in Medicine and Biology Group in Sheffield, UK Although virtual reality has traditionally been used primarily for entertainment purposes the projects presented above show how many of the body-centred and intuitive features of VR technology can be exploited to improve the well being of both user and supplier of healthcare technology.


8 ACKNOWLEDGEMENTS

The projects described in this paper have been supported by the following: Smith and Nephew Surgical, Virtual Presence Ltd, The Engineering and Physical Sciences Research Council, The National Health Service, Trent Health Authority Research Scheme, and The University of Sheffield Research Fund.