One key player in this arena is Elekta, the Swedish medical technology vendor, which unveiled its VMAT capabilities at the ASTRO annual meeting in Los Angeles, CA, last October. Elekta's VMAT technology, which builds on the company's work in volumetric image-guided radiation therapy, is now in use at several European hospitals and recently received 510(k) clearance from the US Food and Drug Administration.

The introduction of volumetric image-guidance and VMAT technologies onto Elekta's treatment systems was led by Kevin Brown, the company's research director. Tami Freeman met with Brown to discuss what's state-of-the-art in IMRT, and what the future has in store for this evolving technique.

How does VMAT technology work?
Elekta's VMAT offers simultaneous dynamic control of the multileaf collimator leaves, the collimator jaws, the gantry and the dose rate during radiation treatment. The ability to move the jaws dynamically is unique to Elekta's VMAT and minimizes the leakage to the patient, as the radiation transmission through the jaws is incredibly low. The system can also perform dynamic collimator rotation, to give the leaves the best fit to the shape that you want to produce at each gantry angle. This flexible delivery technology is combined with powerful treatment-planning optimization.

How does this translate into a clinical benefit?
Take the example of a simple prostate case. Delivering a single-arc VMAT treatment takes about two minutes, compared with six to eight minutes for conventional step-and-shoot IMRT. In this type of simple case, there's no obvious dosimetric advantage of using VMAT - the dose-volume histograms are pretty similar - but it is much quicker.

For this example of a complex head-and neck treatment we'd use two radiation arcs. While the dose to the target is comparable to that delivered by standard IMRT, two-arc VMAT delivers a lower dose than IMRT to every critical structure. The two-arc treatment takes about four minutes, compared with 15 to 20 minutes for IMRT. So not only does VMAT offer a better dose distribution here, it is also still quicker. Shorter treatment time is a great benefit to patients, because they don't have to hold themselves still for as long.

Why use more than one radiation arc?
Our system can treat using single or multiple arcs. Many people that I've spoken to say that in the head-and-neck region, for example, one arc will not be enough to achieve high-quality treatment. Elekta's policy is to maximize the quality of the plan using as many arcs as needed. We think that it will still be as quick as, or faster than, the conventional technique.

How does Elekta's VMAT system compare with tomotherapy?
One study I've seen compared tomotherapy with VMAT for a cranial case, using multiple coplanar and non-coplanar arcs. The techniques offered comparable target coverage, although tomotherapy delivered more uniform target coverage. The dose to the brain stem, however, was greatly reduced using coplanar VMAT. And with non-coplanar arcs - arcs that aren't in the axial plane - VMAT was dramatically better.

Non-coplanar VMAT is achieved by putting an angle on the table about the isocentre. This is something that tomotherapy physically cannot do because it's restricted by the bore. I believe that the use of non-coplanar arcs will be particularly beneficial for intracranial treatments.

So how do these advancements in IMRT relate to image guidance?
I think that volumetric image guidance was a real breakthrough; it brought precision and insight to radiation treatment. Elekta started working on this in 1999, with the William Beaumont Hospital [Royal Oak, MI] and we shipped our first cone-beam CT guidance system back in 2003. We call this functionality VolumeView and it's available on Elekta Synergy and Axesse. A lot of people just think it's having cone-beam CT on the machine. But the point I like to make is that it's not just volumetric imaging; we've developed tools to make volumetric image guidance easy. We were creating cone-beam CT images back in 1999 - what we provide to actually use them in the clinic is much more than this.

What differentiates VolumeView from existing 2D image-guidance schemes?
The objective is to use clinically relevant anatomy for image guidance. VolumeView can use either high- or low-contrast objects depending on the clinical task, whereas 2D imaging can only use high-contrast objects. Critical organs, which limit the dose that can be given to the target and must not be accidentally irradiated, are usually soft tissue. And 2D images simply cannot show these critical structures. We also introduced the idea of using a registration volume, where the clinician defines which part of the volumetric image is used for the guidance process. The 2D imaging uses an area, but we use a volume, which can be right in the middle of the patient if needed.

Is there a problem with increased dose?
There has been a lot of scepticism in the market about using cone-beam CT for image guidance: is it over the top, too complicated, too expensive, too high a dose? There's still a lot of confusion about this in the community. But with our technology, you can adjust the dose patient-by-patient, according to the imaging task.

For example, one study on intracranial registration used a dose of 0.05 cGy and the results were still spot on. In the intracranial region you have a lot of high-contrast objects so the registration task is easy and you need hardly any dose. If you want to see soft tissue then you've got to use more dose, say up to 2 cGy. But that higher dose is associated with the task itself, it's not inherent to volumetric imaging.

How is Elekta tackling the issue of respiratory motion during treatment?
Respiratory motion affects the treatment process in many ways, causing distortion, blurring, artefacts and so on. We thought that it would be a big problem with cone-beam CT, but actually it turns out to be a real feature. Elekta's forthcoming 4D VolumeView image guidance enhances the existing 3D VolumeView to handle respiratory motion.

VolumeView samples the complete volume-of-interest continuously throughout the rotation. So during a long scan, you're sampling multiple breathing phases and effectively imaging the patient's breathing while you do the acquisition. You can then use that data to create a motion-averaged 3D image or use the 4D data to actually look at how the tumour is moving.

How does this help the treatment process?
A couple of years ago, the concept of the internal target volume was introduced to deal with breathing motion. This was effectively the clinical target volume propagated over the breathing cycle. A lot of people presented this and said that the obvious thing to do was gating. Our approach is different - we treat the tumour at the time-weighted average position, so you apply the dose symmetrically to the tumour. The tumour will still be moving, but the dose is applied centrally on the tumour's motion on that particular day.

How is this achieved?
VolumeView works by acquiring a sequence of projection images and using them to determine the breathing phase. This is different from all other systems on the market, which use external surrogates such as a belt or RPM [real-time position management] system. The problem with these surrogates is that they don't necessarily have a stable relationship with the internal anatomy. In contrast, VolumeView sorts the projections into their breathing phases according to the internal anatomy itself. It reconstructs each phase while the gantry's still going round, so when the rotation ends, a 4D image is produced almost instantaneously.

The next part of the process is the automatic 4D registration. Having acquired the 4D image, the system does a bone-match registration to find where patient is on the couch, then it does an automatic soft-tissue match using the region-of-interest for the tumour, for every phase of the 4D CT scan. You end up with a registration result for each breathing phase and can then calculate the mean tumour position and breathing amplitude. Many lung patients exhibit significant baseline shifts in tumour position during their course of treatment; these are detected and managed by this process.

Finally, the system delivers a conventional treatment, including a margin to accommodate the breathing motion. There's no gating or tracking needed. By using this integrated soft-tissue image guidance process to remove errors you can substantially reduce the margins.

But why not just use gating?
For the vast majority of patients the benefit of gating is marginal at best. The problem with gating is if you gate on the wrong phase that's worse than no gating at all. It also extends the treatment time, because the beam is gated off for a significant period during the treatment. The method described above is a very robust technique, because it uses means and averages, and the margin you need to accommodate the breathing motion is in fact very small.

Is the 4D VolumeView software commercially available yet?
It's pre-clinical at the moment, and is being used by our development partner [the Netherlands Cancer Institute/Antoni van Leeuwenhoek Hospital] in Amsterdam. We expect to be releasing it later this year.

Finally, what do you see as the next big developments in radiotherapy?
I think that there are two major developments. One is born out of the information derived from volumetric image guidance. This information will allow us to personalize patients' treatments to a much higher degree. We anticipate that this will result in better patient outcomes and more patients receiving radiation therapy.

The other could be the use of MR imaging during radiation therapy. There are three groups working on this right now. In Utrecht [the University Medical Center Utrecht in the Netherlands] they're exploring the idea of performing the radiotherapy through a standard 1.5 T MR magnet. This will enable live 3D imaging during the radiation delivery.