Introduction.
It is widely recognized that the Oncology community have a problem with training enough students to fill our demanding professional roles, and to keep pace with increasingly complex technology. Recently, the NRAG committee have advocated an increased provision of fractions of Radiotherapy by 91% by 2016; to achieve this, it is evident that we will need to 1) further increase the number of staff we are training and 2) increase efficiency of Patient throughput in an increasingly complex world. Indeed the report states that over the next five years it will be necessary to increase the number of trained Radiographers by a further 30%

The VERT team.
This project is an excellent example of a successful collaboration between different sciences and professions. Andy Beavis is a NHS clinical scientist (Radiation Physicist) at the Princess Royal Hospital in Hull. With 15 years experience in the clinic he has been centrally involved with the introduction of conformal therapy, IMRT, image based treatment planning and been responsible for much of the clinical development. He has also been involved in teaching of all professions on a variety of National and International courses, including a Visiting Professorship at Sheffield-Hallam University where he has taught on the BSc and MSc Therapy Radiographer courses. Roger Phillips is the Research Professor at the Department of Computer Science at the University of Hull. He has been involved in medical research projects for approximately 15 years, including an image guided system for orthopaedic surgery. James Ward is a Research Lecturer in Computer Science, with a large portfolio of research projects, including medical training systems and a Virtual Paraglider simulator. Phillips and Ward have been the driving force and creators of the Hull Immersive Visualisation Environment (HIVE) which has been the home of our work.

In mid 2007 the three inventors founded a company to manage the increasing interest in this work. VERTual Ltd. (www.vertual.co.uk) is a spin-out company from the University of Hull and Hull and East Yorkshire (NHS) Hospitals Trust.

The concept.

In Radiotherapy, we continually improve techniques and technology as we strive to improve treatment. However, lack of clinical resources, the cost of machines and the relatively fast development cycle of technology make it is difficult to provide training in a clinical setting. Pressures from waiting lists and the need to manage work flow can even prevent short periods of time being ‘carved out’ for training.

Through Virtual Reality (VR) simulation we can provide additional training outside the clinic, freeing clinical resources for  treatment. This allows students to learn important concepts, and to practice techniques in a safe environment, where no harm can come to student, machine or patient. Therefore, we allow the luxury of ‘learning by mistakes’ which is not a learning methodology we can afford in the clinic. There are other advantages with our approach: All commercially available Linacs could be placed in a single room and, when these machines are updated or new versions released, the VR Linacs can immediately be updated.

We do not propose to replace clinical experience. Indeed, if students are familiar with the technology in the VR system, they can concentrate on patient care when they enter the clinical treatment bunkers; Learn the ‘bed-side manner’.

The Virtual Environment.

VERT has two main components: software and hardware. The hardware basically comprises of a stereoscopic projection system, 3D glasses, and interaction devices. These are available in a variety of forms and scale – we advocate the use of two basic systems and possible ‘variations’ between.

The full auditorium system uses a back projected screen, large enough to simulate a ‘life sized’ Linac and patient. The Seminar room version is a smaller, forward projected system that can be installed into any existing room. The experience is very close to that of using a real Linac, including the use of an authentic hand held control pendant to control the machine. This ideal for providing practical experience of controlling the Linac, without the associated risks.

The auditorium system includes a motion tracking system to provide an enhanced 3D projection, which correctly accounts for the users natural viewpoint. In summary, this means that if you wish to view the patient from the left, you simply walk around to the left of the patient. This provides a strong sense of presence in the treatment room. Furthermore, the tracking system provides the ability to directly point to and ‘activate’ features within the VR scene, using a hand held pointer.

The VERT software has been written by the development team and is totally independent of other requirements. The basic version of VERT provides a Varian 2100C Linac and treatment room. This is a faithful reproduction of one of the treatment bunkers at Princess Royal Hospital in Hull and can be considered representative of a typical facility. Patients can be loaded from the users treatment planning system, using the built in DICOM RT interface. To date we have tested and successfully loaded plans from three of the major planning companies.

Summary of VERT

The virtual Linac is designed to achieve everything that its real world counterpart can achieve, with the exception of producing radiation! The radiation beams themselves are displayed and therefore a treatment can be visualized. The user can operate the Linac as they would in the clinic. For example, the room lights are dimmed by pressing the room light button and the targeting lasers switched on by pressing the laser button on the pendant. These features are demonstrated in video 1. Note that, through the use of a back projected screen, the user does not cast a shadow which would detract from the 3D effect.

In video 2 we illustrate the ‘collision detection’ feature, which can help students to understand the physical limitations of the treatment machine, and the safety constraints when a patient is lying on the bed. As the gantry approaches the couch the Linac glows red to warn that a collision is imminent.

Video 3 demonstrates the delivery of a simple IMRT plan. The treatment plan was read from the CMS treatment planning system used on our clinic. Individual segments are sequentially shown in order. We have found this visualization useful, outside of basic training, in the teaching of IMRT.

Video 4 shows that in VR we can provide experiences not available in the real world. In presentations the author usually claims that “it is difficult to get a patient to agree to become transparent”! The serious point is that CT and anything else stored within the treatment plan can be displayed in the VR world.

Figure 1 illustrates how dose can be displayed in VERT. It is always clearly understood that 95% isodose clouds cover the PTV (or similar, depending on local prescription rules) however, experience from teaching early students tells us that it is not always obvious that low isodose clouds cover a significant portion of the body. This sequence reminds us, and is useful to prompt discussion of potential long term cancer inducing effects of low level irradiation.

Figure 2 demonstrates the SSD projection tool, including the projections of the field and cross hairs on to the skin. As with a real machine, the light field, SSD and even the collimator jaws can all be interactively controlled via the hand pendant.

Teaching capability examples.

In this section we cover two applications and pilot studies The first serves to illustrate how VERT can be used to teach basic Linac and treatment planning concepts. We are currently working with the Society (College) of Radiographers to develop comprehensive teaching plans for use in the teaching Universities.

Basic concepts: Isocentre, accuracy requirements and margins.

The use of a virtual front pointer, familiar to all physicists who perform QA on their machines, and some virtual Cartesian axes at the isocentre make it very clear that the gantry rotates about a single point in space . Video 5 shows this sequence. Rotation of the collimator and couch can be used to similar effect. We have added side crosses on the patient here to indicate the isocentre or set-up points on the skin surface. The student can move the couch until these points are aligned with the set-up lasers video 6, thus placing the patient in the correct position for treatment.

Once the patient is set up for treatment we could run through the treatment as seen before in video 3 or raise questions about positional or treatment accuracy. By ‘zooming’ into the patient (as seen in video 7) we can enhance the typical learning experience by showing the relationship. In the example shown here we see the relationship between the isocentre (the axes), the patients correct set up position (the projection of the lasers inside the skin surface) and the PTV (green volume), the GTV (red volume) and the proximal kidneys and spinal cord. This is an excellent tool for leading a debate on the role of set-up accuracy on gaining the intended treatment outcome. Furthermore the role of laser calibration/QA can be added.

To facilitate this we have implemented a ‘set-up error’ tool in which the user can offset the patient by a few millimeters in any direction (sup-inf, ant-post, lft-rt) or by rotation. The plan dose is not recalculated, but remains registered to the machine isocentre. In this way we can demonstrate that movements within the allowed CTV-PTV margin result in the latter still being covered by the prescription isodose surface but larger set-up errors result in geographical miss. This is shown in figure 3.

Use as a ‘flight simulator’: Direct field apposition.

This work was done in conjunction with colleagues from Sheffield-Hallam University and has been published. We wanted to evaluate how VERT could be used to learn and practice techniques which typically cause problems for students. To do this we chose the ‘Direct electron field set-up’ or ‘skin-apposition’ problem. Here we try to find a beam direction that is perpendicular to the skin surface, to achieve a uniform irradiation of the patients skin surface with a single electron beam. We see in video 8 shown here how the gantry and couch can be manipulated to place the patient in the appropriate position. There are a variety of skin ‘mark ups’ for practice, and an automated scoring feature measures the accuracy of the student.

Current developments.

More information.