This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Beavis, A W Search for Related Content PubMed PubMed Citation Articles by Beavis, A W

Is tomotherapy the future of IMRT?

Department of Medical Physics, Hull and East Yorkshire NHS Trust and Post Graduate Medical Institute, University of Hull and Princess Royal Hospital, Saltshouse Road, Kingston Upon Hull HU8 9HE, UK

	Abstract

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
Intensity-modulated radiotherapy (IMRT) has become established in many clinics round the world and is, arguably, technically feasible in any facility. Serial tomotherapy contributed an extensive role in its introduction into the mainstream in the second half of the 1990s. In tomotherapy, literally "slice therapy", highly conformal treatments are possible because of the advantages available within the treatment planning of the IMRT process. Currently the majority of clinics implementing IMRT are doing so using conventional clinical linear accelerators (Linacs) fitted with an integrated multileaf collimator (MLC). At this point in time we may wonder if there is any scope for further dramatic changes in this new technology. As we venture from IMRT initial implementation into image guided therapy it is clear that major changes in approach are still valid and needed. If, at each treatment fraction, we can ensure that treatments are delivered accurately by integration of volumetric imaging into on-line validation, then we can attempt higher levels of conformality. A new treatment machine, the helical tomotherapy system, is available that combines the benefits of tomotherapy with on-line volumetric imaging. In this article we will review this approach and explore its features.

	Introduction

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
Intensity-modulated radiotherapy (IMRT) [1] has been introduced [2–4] into a wide spectrum of clinics worldwide. It is clear that recently its adoption has accelerated due to the availability of commercial planning and delivery software from all the major treatment planning and Linac manufacturers. A full description of the development and milestones of this novel technique is given by Webb [5] and more recently in the first article in this series [6]. In order to develop the theme for this article, we briefly revisit some of these points with some alternative reflections.

The popular introduction of IMRT started in the mid 1990s when the NOMOS Corporation (Swickley, Pennsylvania, USA) introduced the PEACOCK system [7, 8], this comprised of the MIMiC, a tertiary "bolt-on" multileaf collimator (MLC), and a dedicated inverse treatment planning system. The treatment delivered by this system is described as serial tomotherapy, since it is delivered by a number of discrete arcs or indexed arcs of finite width, between which the treatment couch is moved longitudinally [9, 10]. The first patient [8] was treated in 1994 and until around the turn of the century most patients who received IMRT in the world were treated on this system. There is no doubt that the pioneering work of the NOMOS company and their customers hastened the transition of IMRT from being unique to the hospitals supported by academic groups, into the scope of any clinical department. It was reported at the American Association of Physicists in Medicine (AAPM) meeting in 2003 that as of July of that year 141 MIMiCs had been installed globally. Interest in the system was sustained with 30 of those being installed in the previous year.

Currently, those centres commencing IMRT services are likely to be doing so with conventional Linacs equipped with MLCs. Here a finite set of fixed angle/direction beams, somewhat similar to conventional treatments are utilized, however the dose profiles delivered by each of these beams are modulated by essentially superposing many sub-beams or segments. The first treatment of this type was given in 1996 [11]. Alongside the advantage of utilizing existing treatment equipment, the provision of appropriate software by the major treatment planning vendors has ensured this methodology is currently the more popular technique. It is reasonable to assume that in the coming years most patients will be treated using some derivation of this technique and vast experience will be gained by the community. Another article in this series [12] dealt with this treatment modality in depth and as such we will not explore it further.

The major criticism of contemporary IMRT is that we cannot be certain of the geometry (relative position and shape) of the tumour or organs at risk (OARs) at each treatment episode. Careful and exacting protocols are employed to attempt to localize these in the treatment plan, however this can only be considered as describing some average positioning on a snap shot of the anatomy at the time of acquisition. The development of new systems incorporates the common theme of image guided therapy (IGT). In IGT we seek to remove uncertainties in the knowledge of absolute position of the anatomy at the time of delivery, by acquiring volumetric images on the treatment device. There are several approaches currently being introduced including the combination of a conventional CT scanner and conventional Linac served by a common treatment couch [13, 14] and cone-beam kV CT [15] apparatus orthogonally mounted to the Linac treatment head.

At this time in the history of radiotherapy we might ask the question as to whether tomotherapy has completed its role in the development of IMRT. Recently, an exciting treatment machine (see Figure 1) was made commercially available (TomoTherapy Inc., Madison, Wisconsin, USA). The helical tomotherapy machine [16] differs from the approach adopted by NOMOS, in that the treatment is delivered in a spiral fashion (see Figure 2), removing the need to manually move the patient between arcs. Furthermore, with appropriate optimization/planning considerations it ensures that all treatment arcs are "feathered" by default and junctioning between "successive" arcs is no longer a concern. This treatment device is mounted upon a CT ring gantry and combines a dedicated IMRT delivery methodology with CT imaging capability. The first patient was treated using this machine in July 2003. As of April 2004 twelve of the helical tomotherapy machines have been installed in North America, and some 200 patients will have completed treatment by its end.

View larger version (125K): [in this window] [in a new window]   Figure 1. The helical tomotherapy treatment unit.

View larger version (39K): [in this window] [in a new window]   Figure 2. The treatment arcs are delivered in a spiral, or helical, manner in the fashion now familiar on spiral CT. This has removed the need for manual translation of the patient as in the ground-breaking serial tomotherapy (NOMOS) treatment methodology.

	Static beam delivery, arcs and tomotherapy

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
There are a handful of studies in the literature that seek to systematically compare delivery of conformal radiotherapy by static beam and arc based (such as tomotherapy) techniques. Verellen et al [17] conclude that for convex target geometries that whereas both modalities achieve treatment goals, dynamic (conformal) arcs fare better than static beam IMRT delivery with respect to treatment efficiency. As their name suggests the latter utilize arc based solutions, furthermore at each "arc position" the MLC leaf positions are adjusted to provide appropriate conformal (target) shielding. For more complicated (concave) treatment geometries the efficiency advantage was maintained though, as we might expect, OAR dose constraints were not upheld. In other words intensity modulation was required. In both cases serial tomotherapy plans provided good solutions though the treatment efficiency was compromised because of the high number of monitor units required. Pirzkall [18] noted that rotational IMRT proved to have a slight advantage over fixed field IMRT. Philosophically, the interesting aspect we might draw from these statements is that the use of arc based therapy is useful in providing highly conformal treatments and the addition of intensity modulation simply adds to this basic advantage.

Of course, this concept has been previously investigated and indeed used clinically, though on a very small scale. Intensity-modulated arc therapy (IMAT) was proposed by Yu [19, 20]. Here, in addition to the conformal shaping of the MLCs during the arc, multiple arc passes are delivered in order to lay down more than a single intensity level. Arguably, widespread clinical adoption has been slowed by the fact that the treatment planning manufacturers have not yet provided solutions for this methodology.

MacKenzie and Robinson [21] added an interesting facet to this discussion, they describe a method for "Intensity modulated arc deliveries approximated by a large number of fixed gantry position sliding window dynamic multileaf collimator fields". The premise that we should want to deliver "tomotherapy-like" distributions using conventional equipment is itself interesting.

	IMRT optimization: static beam delivery vs serial and helical tomotherapy

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
A detailed discussion of the treatment planning or optimization process was given by Webb in this series [6], however let us briefly "recap" the salient points for what follows.

We start the optimization process by defining a set of treatment beam directions, whose beams-eye-view (BEV) projections cover the planning target volume (PTV), each of these are subdivided into a regular set of beamlets. The volume to be irradiated may, of course, include OARs and we wish to minimize the doses delivered to these structures. A set of rules or constraints are defined that govern the target or maximum doses that each structure should receive, various other control parameters such as relative importance factors are included to resolve paradoxical demands. Essentially, these rules describe the idealized dose distribution wanted for that patient, collectively they form a cost function which is used in the optimization process to describe how badly treatment goals are compromised. By trying many different combinations of beamlet weighting and changing them systematically the optimization routine strives to find a minimum value of the cost function, resulting in a plan that is a close approximation to the ideal dose distribution wanted.

One of the issues that affects how successful the optimization process in the treatment planning of IMRT is the number of beamlets it can utilize. In mathematical terms we denote this as the degrees of freedom available to a particular method. We can easily appreciate this from our experience in conventional planning where the greater the number of beams used the more conformal we can make our plan. The advent of MLCs made this a very practical and widespread philosophy and the advent of computer controlled "auto-set up" functionality on the treatment machines introduced it into widespread practice. Taking a step further back in time we remember that in the pre-Linac era (non-conformal) arcs were introduced to enable more optimal treatment of deep-seated tumours and combat the restricted skin-sparing afforded by cobalt beams. The use of a large number of co-planar beams in contemporary conformal treatments basically constitute "first approximation" arcs.

In order to make some comparison between the potential of the static beam and tomotherapy methodologies we will develop a simple example. Let us consider a PTV that exists within a 10 cm by 10 cm by 10 cm volume, intruding into this volume is at least one OAR with, potentially, further OARs lying very close to the volume.

Static beam delivery
For the static beam deliveries we consider using a fixed number (N) of beam portals, typically 5 to 9, which are 10 cm by 10 cm (in the BEV plane). In general each of these can be divided into 1 cm by 1 cm beamlets, we note that some MLCs can define a better resolution, however we will not consider this to be a significant issue as we can apply the same argument when we develop the tomotherapy discussion which follows. So, each beam will have 100 beamlets and we will have 100 x N possible beamlets with which to control the dose distribution laid down.

Serial tomotherapy
For serial tomotherapy deliveries the beams available to the optimization process are arcs. In theory the arcs can describe a full 360 degree circle with "treatment beams" at each degree, in practice it is more realistic to consider them to exist at discrete angles, let us choose 51 equispaced angles. The width of the fan-beam (in the transaxial plane) at each delivery angle will be governed by the PTV dimensions in a similar manner as described above. Again, if we assume a 1 cm by 1 cm resolution then for each delivery angle there will be an arc of 10 beamlets. Assuming that 10 arcs are required at this resolution to treat the PTV, then there will be 51 x 100 possible beamlets the optimizer could use.

For a simple comparison a seven field equispaced static beam delivery will be able to utilize 700 beamlets and a tomotherapy delivery as described above will have 5100 available beamlets. The comparison is not as simple as stated, since many of the tomotherapy beamlets may become redundant. In selecting directions in static beam delivery, we may not have the most optimal set of beamlets, we can improve this somewhat by selecting directions based on experience, nevertheless even a few degrees of annular change can improve a resulting dose distribution. The point being that although many of the tomotherapy beamlets may be eventually redundant they are initially available to the optimization process and are discarded by it and not a priori. So, even if we conservatively consider over half to be redundant we still have a good illustration of the much greater degrees of freedom, or potential to create highly conformal distributions, that tomotherapy offers.

Helical tomotherapy
For helical tomotherapy deliveries we have another parameter to include! As the treatment beam rotates about the gantry, the couch supporting the patient is translated through it. This can be parameterized by a quantity known as "pitch" as in spiral CT, where a pitch of 1.5 is typical, resulting in the patient being longitudinally translated 1.5 times the width of the imaging beam for each full rotation. However in the treatment scenario a pitch of less than unity is beneficial, typically values of 0.35 or 0.5 are used, in order to irradiate voxels with several gantry rotations or arcs. If we consider a single voxel of unit length and utilization of a pitch of 1/P (P>1) then it will receive dose from beamlets from any/all appropriate beam directions over P arcs. If we consider the delivery from each arc to deliver one "intensity level" then the pitch control allows P intensity levels to be laid down in our voxel of interest. Furthermore, bearing in mind that the voxel is translating through the gantry the position of the "edges" of those intensity levels are also altered and controllable. This latter effect can be thought of in the same way as the "step size" or spatial discretisation [22] of intensity profiles in leaf sequencing for fixed field IMRT beams. Treatments planned on this system use 51 equispaced beam directions per gantry rotation, if we assume a beam width of 1 cm and a helical pitch of 1/5, then we now have 25 500 beamlets available, i.e. the serial tomotherapy answer scaled by the pitch.

Again, this argument is very simple but effectively illustrates the issue at hand, the inclusion of the helical pitch increases the degree of freedom further than in the serial tomotherapy case and means that the optimizer potentially has tens of thousands of possible beamlets at its disposal. If we consider very difficult problems such as pelvic nodal irradiation [23] and certain lung treatments [24] then the potential of the helical tomotherapy approach becomes very clear. In these complex geometries we wish to conformally treat the "rind" of the lung whilst minimizing the dose to the lung itself.

Delivery of the intensity-modulated beams
In most current commercial implementations, the optimization of the treatment is separate from any considerations as to its deliverability. In other words, as far as it can achieve, a perfect treatment is proposed by the optimizer, and if the idealized radiation/fluence profiles it needed to use could be delivered, then it would be achieved. Following the optimization the leaf sequencing algorithm divides [22] the ideal beams into a series of deliverable sub-beams or segments. Compromises, in relation to the realization of the ideal treatment beam, are required because of physical constraints determined by the specific delivery system (typically the MLC) in use. This results in the perfect dose distribution not being achieved. Various investigators are considering the inclusion of these "deliverability constraints" into the optimization process itself [25–28] thereby only proposing beams (and therefore dose distributions) that are truly deliverable. This would benefit all IMRT delivery schemes and will certainly make its way into all commercial planning systems eventually.

	Helical tomotherapy system

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
The helical tomotherapy device was conceived and developed by the research group headed by Mackie [29] at the University of Wisconsin, USA. Latterly a commercial product has been realised, and marketed, by a spin-out company Tomotherapy Inc., Madison, USA.

The machine was designed to be a purpose built IGT machine, to deliver highly modulated IMRT whilst also being able to verify the integrity of the delivery of such treatment. It is best described by the rather romantic reflection that it is the marriage of a treatment Linac with a CT scanner. A 6 MV Linear accelerator waveguide and a CT (MV) detector subsystem (see Figure 3) are mounted upon a rotating gantry assembly, they are provided with power via slip-ring technology which also allows the transmission of data. The capability for continuous rotation, coupled with translation of the patient through the gantry, allows helical treatment arcs in an identical fashion to that now familiar on helical or spiral diagnostic CT scanners.

View larger version (18K): [in this window] [in a new window]   Figure 3. Schematic of the beam-line, illustrating the integration of the therapy Linac with the MVCT detectors. MLC, multileaf collimator.

Independent jaws, integrated with the primary collimation, form the fan-beam, both are manufactured from an alloy containing 95% tungsten. The "clam-shell" design of the jaws provides maximum shielding whilst maintaining compactness, the latter is important in the design of this machine where it is undesirable to have bulky components resident on the rotating gantry. The primary/secondary collimation has been designed to ensure any "head" leakage for IMRT deliveries is comparable with that obtained for conventional treatments. The jaws can be continuously moved and completely opened or closed to form a 5 cm wide fan beam; typical treatment beams widths are between 1 cm and 5 cm.

MLC
The system incorporates a binary MLC, this is very similar to that used by the serial tomotherapy MIMiC device. The binary reference indicates that in operation the leaves are either open or they block the beam, hence turning individual beamlets on or off. This approach requires very fast leaf transition times and this is achieved by using a compressed air system enabling them to close or open in approximately 20 ms. There is also some "latency" or lag time due to associated mechanical delays/response times, however these can be ignored in the delivery since they are allowed for in the implementation of a "direct aperture optimization" algorithm. As discussed in the optimization section this leads to deliverable beams and the achievement of the dose distributions assumed by the optimizer.

There are 64 interleaved leaves, each has a nominal width of 6.25 mm at the isocentre, thus providing a fan beam length of 40 cm. Again they are constructed from a (95%) tungsten alloy. The leaves have a tongue and groove design and are 10 cm thick, these characteristics lend to limiting transmission to approximately 0.5% "in-field" and 0.2% outside. Mackie et al [16] have presented a detailed description and justification for the choice of MLC leaf width. Shepard et al have shown [30] that whereas a 6 mm leaf is seen to provide better dose homogeneity and normal tissue avoidance than a 10 mm leaf, no benefit is seen if the leaf width is reduced to 2 mm.

Recalling that the entire beam-line/collimation system is fixed to the gantry and rotates about the patient this provides an imaging field-of-view (FOV) of a 40 cm diameter. The treatment FOV is wider. Let us consider a PTV which is 41 cm wide at its widest aspect. There will be two directions from which the 40 cm wide beam cannot irradiate the entire PTV, other beam directions will (obliquely) irradiate the entire volume. Developing this argument further we estimate that a cylindrical PTV with diameter of up to 60 cm could be treated and with some degree of non-uniformity (due to increasing number of excluded beam directions) a diameter of 70 cm could be achieved.

Couch
The patient support couch is similar to that found on spiral CT scanners, though naturally it has a flat top. The patient is positioned on it with the aid of conventional localization devices which can be indexed to the treatment couch. Automatic couch motion in the vertically and longitudinal directions are possible to position the patient correctly within the gantry and accurate longitudinal movement allows the patient to the translated through the gantry whilst the gantry rotates to enable the helical treatment approach. Manual lateral translation is allowed. The available couch translation allows unrestricted treatment volume lengths for the large majority of deliveries. The maximum treatment length is 160 cm length for a single spiral with the couch at isocentre height. However it is dependent upon actual couch height, due to the "Cobra-action" of the raising mechanism. Upto this length, long volumes can be treated without the need to match or feather sets of treatment beams. It may be necessary to junction helical fields for tall patients undergoing whole body exposure or total body irradiation, this is possible.

In the discussion of the benefits that the helical tomotherapy delivery afforded the treatment planning/optimization process we touched upon the concept of pitch. The pitch ratio is that between the distance the couch travels during a gantry rotation to the fan beam width. Typically, in the treatment scenario we want to use a fractional pitch meaning that each voxel of the patient is irradiated multiple times before the beam "moves on".

MV imaging
The helical tomotherapy machine has an integrated (MV) CT imaging capability. A conventional xenon ion chamber CT detector system is mounted on the gantry directly opposite to the beam-line. This offers two benefits for IGT: the obvious one being the acquisition of volumetric CT imaging of the patient at the time of treatment delivery. Secondly, during treatment the intensity of the photons that exit the patient can be collected and used in a back-projection computation to assess the dose distribution that is delivered to the patient.

The imaging beams are produced at a lower quality/energy (3.5 MV) than the treatment beams and the output of the guide is reduced. This results in the acquisition of volumetric images at very acceptable doses, typically between 0.5 cGy and 3 cGy, which are comparable with doses required to obtain planar images on contemporary MV electronic portal imaging devices (EPIDs). Though an "off-the-shelf" system, the detector used has an inherent co-incidental design benefit that provides a quantum efficiency of approximately 25%, which is much greater than for EPIDs. The detectors have thin tungsten septa that separate the ionization chambers, a significant number of electrons released at photon scattering events in the septa are conducted into the gas chambers and detected. So, whereas the septa were included to reduce cross-talk "noise" between detectors for kV imaging, they result in an increase in efficiency for MV CT imaging.

The system is designed to produce imaging for verification purposes, hence the obtainable contrast should be of an appropriate standard. A nice demonstration of its capability has been shown by Ruchala et al [31]. In this work they use a familiar CT contrast-resolution phantom and demonstrate the ability to detect a 3% contrast object with dimension of approximately 3 cm. This indicates that the contrast difference between fat and muscle for structures of reasonable dimensions (for example, prostate) will be identifiable. Furthermore the same work demonstrates that for higher contrast differences (air to solid water) objects of 1.2 mm are able to be resolved. It is this authors anecdotal experience that in discussions with Radiologist colleagues the MV CTs are remarked to be "much better than early kV CT images".

Though the CT imaging of the helical tomotherapy machine is not intended to be used to collect diagnostic images, it may have a useful role in the treatment planning of radiotherapy patients. Any experienced treatment planner can testify to the problem and annoyance introduced by metal artefacts in CT planning scans. The non-linearity of attenuation, introduced by the photoelectric effect, results in severe streaking aberrations are seen in the images. As seen in Figure 4 these artefacts are simply not present in the MV CT images because of the different physical processes contributing to the creation of the image. For patients with artificial hips or dental fillings proximal to the PTV a MV CT could be obtained and registered/fused with the standard planning CT. In this way it could be used in parallel to the latter similar to the way an MRI is now commonly used in routine planning. In cases where the kV CT is potentially "worthless" due to severe imaging artefacts we might consider the use of the MV CT alone as MRI has been used [32, 33] in the past.

View larger version (78K): [in this window] [in a new window]   Figure 4. A mega voltage CT image obtained on the tomotherapy is shown on the left and a kV CT planning image on the right. The former does not suffer with the dental-filling artefacts present in the planning data set. Though the MVCT is intended as a verification tool for image guidance it is clear that other benefits may be apparent for planning. We also note that sufficient soft tissue contrast is present to demonstrate the position of the parotid glands in the MVCT.

View larger version (53K): [in this window] [in a new window]   Figure 5. (a) An oesophagus treatment, the isodose levels given in the key or in units of Gy. This treatment took 12 min 24 s to deliver. Courtesy of Dr D Cho of the Southeast Regional cancer Centre in Tallahassee, Florida, USA. (b) The cumulative dose–volume histogram obtained for the treatment whose isodose distribution is shown in (a). The colours of the curves relate to the structures shown in the images.

Once the data sets have been registered they may be fused for the user to inspect. Normal useful features in such software tools are available such as image integration using "checker-board" displays (see Figure 6) enabling an easy comparison of the two data sets.

View larger version (81K): [in this window] [in a new window]   Figure 6. A sagittal reconstruction from the MVCT verification data set is shown in the top left box and the corresponding image from the planning data set underneath. The tinted blocks in the "checker-board" show the MVCT data. Once again the aberrations in the kV image due to prosthesis are not evident in the MV image. This figure was kindly provided by Dr Chester Ramsey III of the Thompson Cancer Survival Centre, Knoxville, Tennessee, USA.

	Image-guided therapy and the adaptive therapy process

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
The helical tomotherapy was designed not just as a dedicated IMRT delivery system but as an IGT delivery system. IGT is the next logical step in the evolution of IMRT to realise highly conformal treatments, here on-line (volumetric) imaging is used in order to localize the targets and conformal avoidance structures at each treatment fraction. This leads to the question: what do we do when we detect that the treatment geometry has changed? In standard conformal therapy we establish our volume delineation and safeguard against organ motion/distortion or set-up errors by applying margins to the treatment volumes. In IGT we have the opportunity to reduce these margins and consequently give higher target doses and make fewer compromises to lower normal tissue doses. If we decide to follow this course of action then it demands we have a strategy for adapting the progression of the treatments when deviations from that planned are detected. Figure 7 demonstrates the "adaptive therapy" process.

View larger version (44K): [in this window] [in a new window]   Figure 7. The adaptive radiotherapy process map as defined by Mackie. The importance of an integrated volumetric imaging system in the treatment system is apparent from this diagram.

It is most likely that in the first instance a protocol that makes one correction for the last week of treatment would be deemed most appropriate. At this time we might actually consider this quite ambitious as (in general) a change in volume will result in a very different IMRT delivery and therefore necessitate the same checks as any new treatment. However, departments that have been using IMRT for some time no longer perform patient specific QA, but perform secondary or independent dose calculations in its place. They are satisfied, from 5 to 10 years experience, that the intensity modulation is delivered accurately and in the way the plan demanded.

Taken to an extreme the daily delivery may be adapted based on the MVCT anatomy presentation, though faster and more powerful computers would be required to make the necessary calculations, however these are always just around the corner! In fact this correction scheme may not be so outlandish: not only will we eventually become comfortable with the ability of the treatment machine to deliver what is demanded, but with an IGT machine we will be able to easily measure the dose exiting the patient using its imaging system. The planning system could compute the expected exit dose patterns and compare them in real time with those measured, if a significant deviation were detected then an interlock could be applied. Any minor deviations detected would be simply fed back into the "adaptation process" described in Figure 7. In other words we replace pre-treatment QA with during-delivery QA. Would this be accurate enough? Given that the anatomy presentation is "measured" before delivery of the treatment this comparison is more likely to be relevant than in the contemporary setting. The latter suffers from having to assume the patient always presents as in the planning scan and so the "attenuation maps" obtained at each fraction may not have the same anatomy baseline as those pre-calculated. Furthermore, even the most carefully executed pre-treatment measurement protocols are assuming that the dose distributions will be produced in "the planning scan" patient. It is our belief that given the resources available on a purposely designed IGT machine that daily correction strategies or an ultimate adaptive therapy will be feasible in the near future and the time scales for its consideration are more dependent on the computer industry than the oncology manufacturers.

The last comment here should consider how these daily treatment plans will be generated. Current methodologies, that demand we contour structures as part of the IMRT prescription, may raise questions as to whether the contrast available in the MVCT images are sufficient. The achievable contrast was discussed earlier in this article. Work is underway, within various groups [37, 38] to automatically produce contours for MV portal (2D) images. This work should be adaptable towards the volumetric MVCT and enable us to produce such prescriptions without the need for manually contouring.

The final step, deformable dose registration, described in Figure 7 is concerned with the method that combines daily dose distributions to produce realistic treated distributions. Though it is potentially a subtle point in the first instance it really is quite an obvious strategy to want to use. Consider a prostate treatment and recall the potential problem introduced by differential rectal filling at each fraction. Let us consider the IMRT beams to be delivered to isocentre at the centre of the prostate, which remains essentially fixed in space, however the rectum/prostate interface is quite unstable throughout treatment moving in the anterior to posterior direction and therefore causing the anterior rectal wall to be appropriately spared on most occasions but treated with high doses on others. We are not really interested in what doses particular voxels or theoretical locations in the treatment plan receive. However, we are interested in knowing what dose the anterior rectal wall received. In the deformable registration [39–41] work developed by Ruchala [42] the daily MVCT data is deformed elastically in order to topographically match the planning CT images, once this elastic transformation has been computed it can be applied to the dose distribution associated with that fraction. Having performed this process to the data for all treatment fractions the "deformed" dose distributions can be added together and the true doses to each organ can be assessed. This approach is likely to be very powerful for contemporary treatment outcomes analysis.

	Discussion

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
As stated elsewhere in this review, the provision of IGT does not require the helical tomotherapy delivery system, however as discussed it certainly provides several advantages in this process. Nevertheless, there are a number of reasons why it would be advantageous to use conventional equipment. For example (and not exclusively): familiarity — operationally and service considerations; availability — existing Linacs could be upgraded to include the required imaging platforms for IGT. If one of these new devices were installed in a centre then the issue of redundancy and guarding against unplanned treatment gaps (due to breakdown) in treatments is an important consideration. Under current popular IMRT acceptance/QA protocols it is not possible to create an alternative Linac-based IMRT solution at short notice. However, if we do believe that the strengths of this new technology lie in the fact that it can deliver (and verify) highly optimal treatment solutions then we should question whether un-optimal treatments should always be preferred in order to guard against such issues. There are certainly many clinics worldwide that have a single high energy Linac with electron treatment capability, so fear of lack of redundancy has not negated the use of Electron therapy!

In the section on optimization we presented an argument as to why the tomotherapy approach has merits over what is currently considered the more standard alternatives. In Figure 8 we present a clinical example for the treatment of a tumour in the base of tongue. These images were provided by Dr Cho of the Southeast Regional Cancer Centre in Tallahassee, Florida, USA. The colours of the curves in the DVH correspond to those of the contoured structures in the (inset) transaxial image. We note the excellent sparing of the each of the parotids and the cord, the majority of former getting less than 15 Gy and the latter less than 8–10 Gy.

View larger version (75K): [in this window] [in a new window]   Figure 8. High dose gradients that can be formed at structure interfaces due to the high number of beamlets etc. This treatment took 4 min 47 s to deliver. Courtesy of Dr D Cho of the Southeast Regional Cancer Centre in Tallahassee, Florida, USA.

View larger version (71K): [in this window] [in a new window]   Figure 9. A planning study of a total marrow irradiation protocol. In this simulation the bones are the planning target volumes and the soft tissues of the body are designated as the organs at risk. Courtesy of Dr T Schultheiss of the City of Hope Comprehensive Cancer Centre, California, USA.

In summary, the ability to provide highly conformal treatments and on-line verification in an integrated unit is a highly desirable system for IGT, helical tomotherapy has this ability.

	Conclusion

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
In this paper we have reviewed an exciting new technology which recently became commercially available. The helical tomotherapy system is a fully integrated purpose designed IMRT or IGT delivery system. It is beyond doubt that serial tomotherapy played a very important part in the history of IMRT to date, we look forward with interest to see whether helical tomotherapy becomes a significant part of the rest of our subjects history.

	Comments

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 
This article has been deliberately pitched to reach the majority of the intended (multidisciplinary) readership. I beg the indulgence of my mentors, whom will undoubtedly find some over-simplifications amusing, but hopefully not too irritating! We, the multidisciplinary UK community, often view trans-Atlantic developments with some awe, this article was intended to introduce and familiarize the reader with one such development and provide background understanding. It is a pleasure to acknowledge the financial support of Computerized Medical Systems, Hull and East Yorkshire Trust, the University of Hull (Clinical BioSciences Institute) and Yorkshire Cancer Research enabling the research within our group. We have no commercial sponsorship from Tomotherapy Inc. Finally, I would like to thank Dr Chris Nutting for his honouring me with the invitation to write about this subject and addressing the question posed in the title.

Received for publication February 25, 2004. Accepted for publication March 2, 2004.

	References

Top Abstract Introduction Static beam delivery, arcs... IMRT optimization: static beam... Helical tomotherapy system Image-guided therapy and the... Discussion Conclusion Comments References

 

1. Intensity Modulated Radiation Therapy Collaborative Working Group. Intensity-modulated Radiotherapy: current status and issues of interest. Int J Radiat Oncol Biol Phys 2001;51:880–914.[CrossRef][Medline]
2. James HV, Scrase CD, Poynter AJ. Practical experience with intensity-modulated radiotherapy. Br J Radiol 2004;77:3–14.[Abstract/Free Full Text]
3. McNair HA, Adams EJ, Clark CH, Miles EA, Nutting CM. Implementation of IMRT in the radiotherapy department. Br J Radiol 2003;76:850–6.[Abstract/Free Full Text]
4. Beavis A. IMRT - in our real (NHS) world? RAD magazine. 2003 September:33–34.
5. Webb S. Intensity modulated radiotherapy. Bristol: Institute of Physics Publishing, 2000.
6. Webb S. The physical basis of IMRT and inverse planning. Br J Radiol 2003;76:678–89.[Abstract/Free Full Text]
7. Carol M. Peacock: a system for planning and rotational delivery of intensity-modulated fields. Int J Imag Syst Technol 1995;6:56–61.
8. Woo SY, Butler B, Grant III W. Clinical experience: benign tumours of the CNS and head and neck tumours. In: Sternick ES, editor. Intensity modulated radiation therapy. Madison, WI: Advanced Medical Publishing; 1997:195–8.
9. Verellen D, Linthouta N, Van Den Bergea D, Bela A, Storme G. Initial experience with intensity-modulated conformal radiation therapy for treatment of the head and neck region. Int J Radiat Oncol Biol Phys 1997;39:99–114.[CrossRef][Medline]
10. Low DA, Chao CKA, Mutic S, Gerber RL, Perez CA, Purdy JA. Quality assurance of serial tomotherapy for head and neck patient treatments. Int J Radiat Oncol Biol Phys 1998;42:681–92.[Medline]
11. Ling CC, Burman C, Chui CS, LoSasso T, Mohan R, Spirou S, et al. Implementation of photon IMRT with dynamic MLC for the treatment of prostate cancer. In: Sternick ES, editor. Intensity modulated radiation therapy. Madison, WI: Advanced Medical Publishing, 1997:219–28.
12. Williams PCW. IMRT: delivery techniques and quality assurance. Br J Radiol 2003;76:766–76.[Abstract/Free Full Text]
13. Cheng CW, Wong J, Grimm L, Chow M, Uematsu M, Fung A. Commissioning and clinical implementation of a sliding gantry CT scanner installed in an existing treatment room and early clinical experience for precise tumor localization. Am J Clin Oncol 2003;26:e28–36.[CrossRef][Medline]
14. Court L, Rosen I, Mohan, Dong L. Evaluation of mechanical precision and alignment uncertainties for an integrated CT/LINAC system. Med Phys 2003;30:1198–210.[CrossRef][Medline]
15. Jaffray DA, Siewerdsen JH, Wong JW, Martinez AA. Flat-panel Cone-beam computed tomography for image-guided radiation therapy. Int J Radiat Oncol Biol Phys 2002;53:1337–49.[CrossRef][Medline]
16. Mackie TR, Olivera GH, Kapatoes JM, Ruchala KJ, Balog JP, Tome WA, et al. Helical tomotherapy. In: Palta P, Mackie TR, editors. Intensity-modulated radiation therapy: the state of the art. College Park MD: American Association of Physicists in Medicine, 2003:247–84.
17. Verellen D, Linthout N, Soete G, Van Acker S, De Roover P, Storme G. Considerations on treatment efficiency of different conformal radiation therapy techniques for prostate cancer. Radiother Oncol 2002;63:27–36.[CrossRef][Medline]
18. Pirzkall A, Carol M, Lohr F, Hoss H, Wannenmacher M, Debus J. Comparison of intensity-modulated radiotherapy with conventional conformal radiotherapy for complex-shaped tumors. Int J Radiat Oncol Biol Phys 2000;48:1371–80.[Medline]
19. Yu C. Intensity-modulated arc therapy with dynamic multileaf collimation: an alternative to tomotherapy. Phys Med Biol 1995;40:1435–49.[CrossRef][Medline]
20. Yu C, Allen L. Clinical implementation of intensity modulated arc therapy. Int J Radiat Oncol Biol Phys 2002;53:453–63.[Medline]
21. MacKenzie MA, Robinson DM. Intensity modulated arc deliveries approximated by a large number of fixed gantry position sliding window dynamic multileaf collimator fields. Med Phys 2002;29:2359–65.[Medline]
22. Beavis A, Ganney P, Whitton V, Xing L. Optimisation of the step-and-shoot leaf sequence for delivery of intensity modulated radiation therapy using a variable division scheme. Phys Med Biol 2001;46:2457–65.[Medline]
23. Guerrero Urbano MT, Nutting CM. Clinical use of intensity modulated radiotherapy: part II. Br J Radiol 2004;77:177–82.[Abstract/Free Full Text]
24. Guerrero Urbano MT, Nutting CM. Clinical use of intensity modulated radiotherapy: part I. Br J Radiol 2004;77:88–96.[Abstract/Free Full Text]
25. Markman J, Low D, Beavis A, Deasy J. Beyond bixels: generalizing the optimization parameters for intensity modulated radiation therapy. Med Phys 2002;29:2298–304.[Medline]
26. Alber M, Nusslin F. Optimization of intensity modulated radiotherapy under constraints for static and dynamic MLC delivery. Phys Med Biol 2001;46:3229–39.[CrossRef][Medline]
27. De Gersem W, Claus F, De Wagter C, De Neve W. An anatomy-based beam segmentation tool for intensity-modulated radiation therapy and its application to head-and-neck cancer. Int J Radiat Oncol Biol Phys 2001;51:849–59.[CrossRef][Medline]
28. Shepard DM, Earl MA, Li XA, Naqvi S, Yu C. Direct aperture optimization: a turnkey solution for step-and-shoot IMRT. Med Phys 2002;29:1007–18.[Medline]
29. Mackie TR, Holmes T, Swerdloff S, Reckwerdt P, Deasy JO, Yang J, et al. Tomotherapy: a new concept for the delivery of dynamic conformal radiotherapy. Med Phys 1993;20:1709–19.[CrossRef][Medline]
30. Shepard DM, Olivera G, Angelos L, Sauer O, Reckwerdt P, Mackie TR. A simple model for examining issues in radiotherapy optimization. Med Phys 1999;26:1212–21.[Medline]
31. Ruchala KJ, Olivera GH, Kapatoes JM, Reckwerdt PJ, Mackie TR. Methods for improving limited field-of-view radiotherapy reconstructions using imperfect a priori images. Med Phys 2002;29:2590–605.[Medline]
32. Beavis AW, Gibbs P, Dealey RA, Whitton VJ. Radiotherapy treatment planning of brain tumours using MR images alone. Br J Radiol 1998;71:544–8.[Abstract]
33. Lee YK, Bollet M, Charles-Edwards G, Flower MA, Leach MO, McNair HA, et al. Radiotherapy treatment planning of prostate cancer using magnetic resonance imaging alone. Radiother Oncol 2003;66:203–16.[CrossRef][Medline]
34. Kapatoes JM, Olivera GH, Balog JP, Keller H, Reckwerdt PJ, Mackie TR. On the accuracy and effectiveness of dose reconstruction for tomotherapy. Phys Med Biol 2001;46:943–66.[CrossRef][Medline]
35. McNutt TR, Mackie TR, Paliwal BR. Analysis and convergence of the iterative convolution/superposition dose reconstruction technique for multiple beams and tomotherapy. Med Phys 1997;24:1465–76.[CrossRef][Medline]
36. Van Esch A, Depuydt T, Huyskens DP. The use of an aSi-based EPID for routine absolute dosimetric pre-treatment verification of dynamic IMRT fields. Radiother Oncol (In press).
37. Dekker N, Ploeger LS, van Herk M. Evaluation of cost functions for gray value matching of two-dimensional images in radiotherapy. Med Phys 2003;30:778–84.[Medline]
38. Kim J, Fessler JA, Lam KL, Balter JM, Ten Haken RK. A feasibility study of mutual information based setup error estimation for radiotherapy. Med Phys 2001;28:2507–17.[Medline]
39. Bharatha A, Hirose M, Hata N, Warfield SK, Ferrant M, Zou KH, et al. Evaluation of three-dimensional finite element-based deformable registration of pre- and intraoperative prostate imaging. Med Phys 2001;28:2551–60.[CrossRef][Medline]
40. Liang J, Yana D. Reducing uncertainties in volumetric image based deformable organ registration. Med Phys 2003;30:2116–22.[Medline]
41. Brock KM, Balter JM, Dawson LA, Kessler ML, Meyer CR. Automated generation of a four-dimensional model of the liver using warping and mutual information. Med Phys 2003;30:1128–33.[Medline]
42. Ruchala KJ, Olivera GH, Kapatoes JM. Limited-data image registration for radiotherapy positioning and verification. Int J Radiat Oncol Biol Phys 2002;54:592–605.[Medline]
43. Chang SD, Main W, Martin DP, Gibbs IC, Heilbrun MP. An analysis of the accuracy of the CyberKnife: a robotic frameless stereotactic radiosurgical system. Neurosurgery 2003;52:140–6.[Medline]
44. Ryu SI, Chang SD, Kim DH, Murphy MJ, Le QT, Martin DP, et al. Image-guided hypo-fractionated stereotactic radiosurgery to spinal lesions. Neurosurgery 2001;49:838–46.[Medline]
45. King CR, Lehmann J, Adler JR, Hai J. CyberKnife radiotherapy for localized prostate cancer: rationale and technical feasibility. Technol Cancer Res Treat 2003;2:25–30.[Medline]

This article has been cited by other articles:

				N. Han and S. Galandiuk Induction Chemoradiation for Rectal Cancer Arch Surg, December 1, 2006; 141(12): 1246 - 1252. [Abstract] [Full Text] [PDF]

				S Webb Intensity-modulated radiation therapy (IMRT): a clinical reality for cancer treatment, "any fool can understand this": The 2004 Silvanus Thompson Memorial Lecture Br. J. Radiol., October 1, 2005; 78(Special_Issue_2): S64 - S72. [Full Text] [PDF]

				BJR Review of the Year - 2004 Br. J. Radiol., March 1, 2005; 78(927): 181 - 185. [Full Text] [PDF]

				P. M. Harari Promising new advances in head and neck radiotherapy Ann. Onc., January 1, 2005; 16(suppl_6): vi13 - vi19. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Beavis, A W Search for Related Content PubMed PubMed Citation Articles by Beavis, A W