British Journal of Radiology (2005) 78, 548-552
© 2005 British Institute of Radiology
doi: 10.1259/bjr/53491625
Helical tomotherapy for craniospinal radiation
G Bauman, MD1,
S Yartsev, PhD2,
T Coad, BSc, CMD2,
B Fisher, MD1 and
T Kron, PhD2
Departments of 1 Radiation Oncology and 2 Clinical Physics, London Regional Cancer Centre, 790 Commissioners Road East, London, Ontario, N6A 4L6, Canada
 |
Abstract
|
|---|
Helical tomotherapy (HT) plans for craniospinal radiation were generated for the case of a 4-year-old boy with disseminated ependymoma. The HT plans demonstrated excellent target coverage, homogeneity and organ sparing compared with a conventional linear accelerator based craniospinal radiation plan. On the basis of this case study, further evaluation of HT for craniospinal radiotherapy seems justified.
|
Introduction
|
|---|
Patients with brain tumours at risk for dissemination through the cerebrospinal fluid often require craniospinal axis irradiation. Comprehensive irradiation of the craniospinal space is technically demanding [1–5]. Craniospinal radiation typically utilizes prone positioning of the patient combined with careful junctioning between opposed lateral cranial fields and a posterior spine field. In larger patients, an additional junction between upper and lower spine fields may be required [4]. Periodic junction shifts are introduced throughout the treatments to address the potential risk of underdose or overdose at the field junctions. Many of the patients requiring craniospinal radiotherapy are paediatric patients who may have difficulty cooperating with prone positioning for treatment and also may require daily anaesthesia for treatment. In these cases, supine positioning may be preferable depending on the institution preference. With either supine or prone positioning, the challenge of setting and verifying multiple field junctions and the maintenance of a homogeneous dose over a long treatment volume remains [2, 3].
Helical tomotherapy (HT) is a novel approach to the delivery of radiation for cancer treatment [6, 7]. It relies on a 6 MV linear accelerator mounted on a ring gantry that rotates around the patient as they advance slowly through the ring. During treatment delivery, the radiation fan beam is defined using a 64-leaf collimator. Each leaf projects a shadow of 0.625 mm width at the isocentre 85 cm away from the target and the fan beam width is adjustable from 0.5 cm to 5 cm. The alteration of leaf positions as a function of the gantry position while the patient advances slowly through the gantry allows great flexibility in sculpting a sophisticated target dose distribution while sparing critical normal structures. In addition, the smooth, helical delivery of the intensity-modulated fan beam allows the treatment of extended volumes in the cephalo–caudad direction in either the prone or supine position without the need for junctioning. We sought to model the potential advantages of HT delivery for craniospinal radiation for the case of a paediatric patient with disseminated ependymoma.
|
Materials and methods
|
|---|
Patient presentation and treatment
The patient was initially diagnosed with a well-differentiated ependymoma of the posterior fossa at age 18 months. The tumour recurred despite multiple surgical resections and once the child reached 4 years of age, a decision was made to treat the child with craniospinal irradiation under general anaesthesia. At the time of treatment, the child had gross tumour remaining in the 3rd and 4th ventricular regions in close proximity to brainstem and optic apparatus as well as a gross tumour deposit in the spine at L4. Given the potential risk of neurocognitive side effects in a 4-year-old child and the proximity of the gross tumour to brainstem, optic apparatus and spinal cord it was decided to limit the total gross tumour volume dose to 45 Gy with a craniospinal dose of 36 Gy. For radiation treatment planning CT images from the vertex of the skull to the ischial tuborosities was acquired with uniform slice thickness and spacing of 3 mm. For clinical treatment the patient was planned and treated with a linear accelerator (LINAC) based technique. The LINAC treatment utilized a half-beam blocked technique to match the cranial and upper spinal fields. A second, lower spinal field was placed using a fixed longitudinal displacement from the superior junction using the exact bed positioning system. Field width for the posterior spine field was 6 cm with the lateral field borders extending to the edge of the transverse processes of the vertebrae. Three junction shifts of 0.5 cm were introduced using asymmetric jaws to maintain the matchline. For the final phase of treatment, reduced size, opposed lateral fields were used to boost the gross disease within the cranium while a single posterior spine field was used to boost the cauda equina disease. For this case study the archived CT planning dataset used for the LINAC treatment was re-planned at a later date using the HT system to explore potential benefits and limitations of this form of intensity-modulated radiation delivery for craniospinal treatment.
Helical tomotherapy planning
For generation of a comparison HT plan CT datasets and structures were transferred to the tomotherapy planning workstation (TomoTherapy Inc., Madison, WI) using the DICOM RT protocol. The tomotherapy station re-samples the CT datasets in 256 x 256 voxels with the slice thickness re-sampled to the smallest slice separation in the original CT dataset. In the present study uniform slice thickness and separations of 3 mm were used. The planning system uses an inverse treatment planning process based on iterative least squares minimization of an objective function as described by Shepard [8]. Prior to optimization, dose volume constraints, precedence, importance and penalty factors for target and critical structures were assigned. Target constraints were a minimum of 45 Gy to the gross tumour in the brain and spinal cord and 36 Gy to the remainder of the craniospinal axis. Additional critical organ constraints for eyes (Dmax 8 Gy; D5 20%), lung (Dmax 40 Gy; V10 10 Gy), liver (Dmax 25 Gy, V10 10 Gy), bowel (Dmax 25 Gy, V10 10 Gy), heart (Dmax 30 Gy, V10 10 Gy), and kidneys (Dmax 20 Gy, V20 10 Gy) were specified. As part of the optimization the dose is calculated using a superposition convolution approach [9, 10]. Parameters specified as part of the optimization/dose calculation process are pitch, fan beam thickness (FBT) and modulation factor. These factors are similar to those specified for diagnostic CT scanning with the exception of modulation factor (which determines the range of intensity levels that can be achieved). By increasing the FBT and reducing the pitch and modulation factor one can significantly reduce the treatment time. For optimum dose conformality, small FBT and pitch and high modulation factors should be employed, at the expense, however, of longer treatment times. In clinical practice, the selection of slice width, pitch and modulation factors is unique to the clinical site being treated and represents a compromise between excessively long treatment times (with highly conformal delivery) and shorter treatment times (with some loss of conformality and/or dose homogeneity). While most diagnostic CT scanners allow a mixture of slice separations to be specified over a given volume to be scanned, the current HT unit allows only one set of delivery parameters to be specified per plan. For the current case, two HT plans were generated. The first plan used a 25 mm fan beam for delivery in order to produce a treatment that could be delivered in a reasonable length of time (
15 min). The second plan used a 10 mm fan beam width for the cranial component that would allow optimum conformality and dose modulation around the eyes and cribiform plate. Dose–volume histograms for both the conventional and helical tomotherapy plans were compared.
|
Results
|
|---|
Dose–volume histograms are illustrated in Figure 1
for the LINAC based plan. The gross tumour volume with margin in the brain and spine receives the prescription dose of 45 Gy while the remainder of the craniospinal axis receives 36 Gy. Inspection of the dose–volume histogram reveals considerable heterogeneity in the gross target volume (GTV) and clinical target volume (CTV) target doses, particularly in the spine (a consequence of the use of a single posterior field with prescription at depth as well as the junctions between fields). Organ at risk doses are also quite heterogeneous with some organs receiving doses close to the CTV dose, albeit to small volumes.
View larger version (64K):
[in this window]
[in a new window]
|
Figure 1. Dose–volume histogram for LINAC craniospinal treatment. Esoph: oesophagus; Gut: small/large bowel; CTVb/CTVs: clinical target volume brain and spine; GTVb/GTVs: gross tumour volume brain and spine; PTVb/PTVs: planning target volume for GTV boost brain and spine.
|
|
Dose–volume histograms for the 25 mm fan beam HT plan are illustrated in Figure 2 with a sagittal dose distribution illustrated in Figure 3. With the 25 mm fan beam an overall estimated treatment time of approximately 15 min could be realised. Inspection of the dose–volume histogram reveals excellent conformality for the CTVcord, GTVbrain and GTVcord as well as the critical structures. In addition, excellent homogeneity in dose delivery to CTV and GTV volumes is evident and doses to the organ at risk are uniformly low. Not indicated on the dose–volume histogram but of potential concern would be dose to the teeth and dose variation across the vertebral body as this could affect growth and development of these structures. Inspection inset 1 of Figure 3 indicates point doses to the individual teeth ranging from 9 Gy to 10 Gy (over 25 fractions). With regards to the vertebral bodies, with HT delivery, the dose falls off in all directions from the target volume (in this case the spinal canal) in the axial plane. This fall off is uniform in all directions and ranges from 36 Gy to 24 Gy over the width of the vertebral body (as can be seen in the sagittal view of Figure 3). Thus, the actual dose to bone will vary depending on the distance from the spinal canal but will be relatively symmetric. In comparison, with the LINAC plan the dose to the vertebral body ranged from 40 Gy at the posterior surface to approximately 34 Gy at the anterior surface creating a non-symmetric dose gradient (and possibly a higher chance of growth plate development asymmetry as a consequence).
View larger version (15K):
[in this window]
[in a new window]
|
Figure 2. Dose–volume histogram (DVH) for the 25 mm fan beam plan. Excellent coverage of the clinical target volume (CTV) (craniospinal axis) with 36 Gy and the gross tumour in the cord and brain (gross target volume (GTV) with 45 Gy. Doses to critical structures (OAR) are below specified limits. Compared with the DVH for LINAC treatment (1b), variation within the GTV and CTV for brain and spine is much less and OAR doses are overall much lower. For paired organs (Lung, Kidney) only the side with the higher (i.e. least favourable) DVH is illustrated.
|
|
View larger version (81K):
[in this window]
[in a new window]
|
Figure 3. 25 mm helical tomotherapy plan demonstrating excellent conformality in sagittal plane. The chosen fan beam width thickness results in compromises in cribiform plate coverage in order to respect eye dose volume constraints (Inset 2).
|
|
Inspection of a central axis slice through the level of the cribiform plate and eyes reveals an inability to achieve acceptable coverage with the necessary ocular sparing with the 25 mm fan beam (Figure 3
inset 2). An alternate plan using a 10 mm width fan beam is illustrated in Figure 4. Examination of the central axis dose distribution at the level of the eyes for this plan demonstrates excellent coverage of the cribiform plate with concomitant sparing of the retina and lens. Of concern with the 10 mm FBT plan however was the estimated delivery time. For the patient concerned, a total treatment time of 40 min was estimated if the whole brain and spine was going to be treated with the small FBT. This treatment time is significantly longer than that of the actual time for treatment experienced on the LINAC and use of the narrow fan beam for the entire craniospinal treatment would not be practical clinically.
View larger version (73K):
[in this window]
[in a new window]
|
Figure 4. Helical tomotherapy plan using 10 mm fan beam width for cranial treatment. Improved coverage of the cribiform plate area is evident.
|
|
|
Discussion
|
|---|
Craniospinal irradiation is often required in the treatment of common paediatric primary brain tumours such as primary central nervous system germ cell tumours, medulloblastoma and other primitive neuroectodermal tumours [11, 12]. Traditional craniospinal irradiation (CSI) techniques on linear accelerators have employed the prone position with multiple photon fields with junction matches. Supine positioning may be used as an alternative to the prone position at the expense of additional complexity and uncertainty in the junctioning of the photon fields [2, 3]. The HT technique examined in this study offers several advantages over LINAC based craniospinal techniques:- Helical delivery allows the treatment of extended treatment volumes without the need for field junctioning.
- A localization pre-treatment megavoltage CT scans (MVCT) with correction for any detected shifts can be acquired prior to treatment in order to precisely align the patient each day [13].
- Highly conformal treatments can be efficiently delivered without the need for discontinuous couch and gantry movements as employed by non-coplanar LINAC IMRT techniques.
- The absence of beam modifiers such as blocks and attenuators reduces the chance for treatment error and speeds up the treatment process.
Clinical implementation of an optimal HT plan that realises all these advantages is limited by constraints inherent in the current HT unit. The use of a narrower fan beam and pitch as well as an increased modulation factor would facilitate improved conformality, especially around the cribiform plate, but the clinical use of these parameters would result in estimated treatment times in excess of 40 min. For children requiring general anaesthesia, remote monitoring without access to the patient for this length of time may be of concern.
Use of a wider treatment fan beam of 25 mm allows for treatment of the craniospinal volume in this case with an estimated treatment time of 15 min, similar to the treatment time on a linear accelerator. However, one consequence of this fan beam width selection, is less "resolution" for dose shaping, especially around the cribiform plate and reduction in the sharpness of the superior inferior dose gradients. The simplest way to address this concern could include treatment with variable fan beam width between the cranial and spinal fields, however the current tomotherapy unit does not provide this option. Alternative strategies include treatment of separate cranial and spinal fields with different fan beam widths but with introduction of a junction, however this obviates some of the benefits of tomotherapy delivery. Alternatively, treatment of the whole craniospinal axis with the wider fan beam width with a subsequent "dose patch" with a fine fan beam to top up dose within the cribiform plate could be considered [5]. The use of a "dose patch" with tomotherapy has been proposed for dealing with sub-optimal brachytherapy implants [14]. In the case of the craniospinal treatment the large fan beam thickness could be used to achieve a short overall delivery time to the craniospinal axis, accepting under dose in some parts of the target. The areas of underdose in the cribiform plate region would then subsequently be boosted or "patched" during the same treatment session using the fine resolution afforded by the smaller fan beam thickness and a small pitch factor. In this approach the patient position is not altered and different dose levels can be given in highly non-uniform fashion to make up exactly for the imperfection of the initial delivery. As well the overall treatment time would be kept short by restricting the use of the narrow beam to the cribiform region only. Currently, the dose patch approach cannot be performed with the current version of planning software as it is neither possible to plan on previous isodose surfaces as structures nor to add plans but may be available as a therapeutic strategy in the future.
In terms of the boost field for treatment, HT is ideally suited for the conformal irradiation of gross tumour residual or the post-operative bed. Reduced volume irradiation of the posterior fossa with conformal techniques for patients with medulloblastoma has been described [15, 16]. A potential benefit of the HT plan is the ability to easily deliver a synchronous, field-in-field boost to a target sub-volume. A synchronous in-field boost to the tumour bed may offer an opportunity to improve tumour control without an increase in normal tissue side effects. In addition, with a synchronous boost, overall treatment times can be shortened, for improved disease control [17]. For those children requiring sedation for treatment a shorter overall treatment time could also mean fewer general anaesthesia sessions. The use of a synchronous boost has implications for surrounding normal tissues in terms of fraction size in adjacent structures. Modelling of synchronous versus sequential boost in the context of head and neck treatments suggests radiobiological advantages to this approach [18]. The application of this approach in paediatric patients requires careful evaluation in the context of clinical trials where normal tissue toxicity is tracked.
In addition to technical issues, biological questions with regards to HT delivery for craniospinal radiation also exist. There may be concern regarding the potential of tumour cells within the cerebrospinal fluid to circulate in and out of a moving fan treatment beam. However, good results have been achieved with scanning electron field radiation of the posterior spine field suggesting this may not be clinically important [19]. An additional biological concern is the integral dose to the body resulting from helical delivery of the craniospinal treatments. While individual organ tolerances are incorporated in the delivery, a significant volume of tissue receives dose in the 3–10 Gy range. While these doses are unlikely to cause late tissue effects there remains a concern regarding potential for second malignancies or effects on normal tissue/organ development with larger treatment volumes in paediatric patients. Despite this concern, the paediatric radiotherapy community is starting to embrace technologies (radiosurgery, 3DCRT and IMRT) for the precision irradiation of intracranial targets that deliver relatively large low dose regions in the brain [16, 20–22]. Finally, many paediatric brain tumour patients are now treated with concurrent and/or adjuvant chemotherapy. The toxicity of a rotational vs a fixed field radiation technique when combined with chemotherapy for craniospinal radiation is not known. Evaluation of IMRT treatments for paediatric patients clearly will require long term follow-up to assess the possible long-term consequences on growth and development as well as second malignancy, particularly in organs that may not have been exposed to significant radiation with previous techniques.
|
Conclusions
|
|---|
A sample case of a 4-year-old boy planned for HT treatment of the craniospinal axis illustrates potential advantages and some practical implementation issues. Further development of HT techniques for craniospinal radiation seems justified and ideally, the implementation of this technique should be in the context of clinical protocols where toxicity and outcome for patients can be carefully tracked.
Received for publication June 1, 2004.
Revision received July 26, 2004.
Accepted for publication January 25, 2005.
|
References
|
|---|
- Mah K, Danjoux CE, Manship S, Makhani N, Cardoso M, Sixel KE. Computed tomographic simulation of craniospinal fields in pediatric patients: improved treatment accuracy and patient comfort. Int J Radiat Oncol Biol Phys 1998;41:997–1000.[CrossRef][Medline]
- Michalski JM, Klein EE, Gerber R. Method to plan, administer, and verify supine craniospinal irradiation. J Appl Clin Med Phys 2002;3:310–6.[CrossRef][Medline]
- Thomadsen B, Mehta M, Howard S, Das R. Craniospinal treatment with the patient supine. Med Dosim 2003;28:35–8.[CrossRef][Medline]
- Van Dyk JJD, Leung PMK, Cunningham JR. Medulloblastoma: treatment technique and radiation dosimetry. Int J Radiat Oncol Biol Phys 1977;2:993–1005.[Medline]
- Halperin EC. Impact of radiation technique upon the outcome of treatment for medulloblastoma. Int J Radiat Oncol Biol Phys 1996;26:233–9.
- Olivera GH, Shepard D, Ruchala K, Aldridge KS, Kapatoes J, Fitchard EE, et al. Tomotherapy. In: Van Dyk J, editor. The modern technology of radiation oncology. Madison, WI: Medical Physics Publishing, 1999:521–87.
- Mackie TR, Balog J, Ruchala K, Shepard D, Aldridge S, Fitchard E, et al. Tomotherapy. Sem Radiat Onc 1999;9:108–17.
- Shepard DM, Olivera GH, Reckwerdt PJ, Mackie TR. Iterative approaches to dose optimization in tomotherapy. Phys Med Biol 2000;45:69–90.[CrossRef][Medline]
- Papanikolaou N, Mackie TR, Meger-Wells C, Gehring M, Reckwerdt P. Investigation of the convolution method for polyenergetic spectra. Med Phys 1993;20:1327–36.[CrossRef][Medline]
- Mackie TR, Olivera GH, Reckwerdt PJ, Shepard DM. Convolution/superposition photon dose calculation. In: Shui AS, Mellenberg DE, editors. General practice of radiation oncology physics in the 21st century. Madison, WI: Medical Physics Publishing, 2000:39–56.
- Packer RJ VG. Pediatric glial neoplasms including brain-stem gliomas. Semin Oncol 1994;21:260–72.[Medline]
- Pollack IF. Brain tumors in children. New Engl J Med 1994;331:1500–7.[Free Full Text]
- Ruchala KJ, Olivera GH, Kapatoes JM, Schloesser EA, Reckwerdt PJ, Mackie TR. Megavoltage CT image reconstruction during tomotherapy treatments. Phys Med Biol 2000;45:3545–62.[CrossRef][Medline]
- Lewis C, Kron T, Chen J, D'Souza D. Helical tomotherapy or IMRT to boost a "suboptimal" prostate brachytherapy implant: a feasibility study. AAPM 2002;29:1941.
- Merchent TE, Happersett L, Finlay JL, Leibel SA. Preliminary results of conformal radiation therapy for medulloblastoma. Neuro-oncol 1999;1:177–87.[Abstract]
- Wolden S, Dunkel IJ, Souweidane MM, PHappersett L, Khakoo Y, Schupak K, et al. Patterns of failure using a conformal radiation therapy tumor bed boost for medulloblastoma. J Clin Oncol 2003;21:3079–83.[Abstract/Free Full Text]
- Taylor RE, Bailey CC, Robinson K, Weston CL, Ellison D, Ironside J, et al. Results of a randomized study of preradiation chemotherapy versus radiotherapy alone for nonmetastatic medulloblastoma: The International Society of Paediatric Oncology/United Kingdom Children's Cancer Study Group PNET-3 Study. J Clin Oncol 2003;21:1581–91.[Abstract/Free Full Text]
- Dogan N, King S, Emami B, Mohideen N, Mirkovic N, Leybovich LB, et al. Assessment of different IMRT boost delivery methods on target coverage and normal-tissue sparing. Int J Radiat Oncol Biol Phys 2003;57:1480–91.[CrossRef][Medline]
- Gaspar L, Dawson DJ, Tilley-Gulliford SA, Benerjee P. Medulloblastoma: long-term follow-up of patients treated with electron irradiation of the spinal field. Radiology 1991;180:867–70.[Abstract/Free Full Text]
- St Clair WH, Adams JA, Bues M, Fullerton BC, Shell S, Kooy HM, et al. Advantage of protons compared to conventional X-ray or IMRT in the treatment of a pediatric patient with medulloblastoma. Int J Radiat Oncol Biol Phys 2004;58:727–34.[CrossRef][Medline]
- Merchant TE, Mulhern RK, Krasin MJ, Kun LE, Williams T, Li C, et al. Preliminary results from a phase II trial of conformal radiation therapy and evaluation of radiation-related CNS effects for pediatric patients with localized ependymoma. J Clin Oncol 2004;22:3156–62.[Abstract/Free Full Text]
- Saran FH, Baumert BG KV, Adams EJ, Garre ML, Warrington AP, Brada M. Stereotactically guided conformal radiotherapy for progressive low-grade gliomas of childhood. Int J Radiat Oncol Biol Phys 2002;53:43–51.[CrossRef][Medline]
Top
Abstract
Introduction
