British Journal of Radiology (2003) 76, 254-259
© 2003 British Institute of Radiology
doi: 10.1259/bjr/59091776
Survey of tangential field planning and dose distribution in the UK: background to the introduction of the quality assurance programme for the START trial in early breast cancer
E A Winfield, DCR(T), BSc1,
A Deighton, DCR(T), BSc1,
K Venables, MSc, MIPEM1,
P J Hoskin, FCRP, FRCR1 and
E G A Aird, PhD, FIPEM2 on behalf of the START Trial Working Party
1 Marie Curie Research Wing, Mount Vernon Hospital, Rickmansworth Road, Northwood, Middlesex HA6 2RN and 2 Department of Medical Physics, Mount Vernon Hospital, Rickmansworth Road, Northwood, Middlesex HA6 2RN, UK
Correspondence: Elizabeth Winfield, Research Radiographer, Radiotherapy department, The Royal Marsden Hospital, Fulham Road, London SW3 6JJ, UK (current address)
 |
Abstract
|
|---|
A background survey of UK breast radiotherapy techniques was performed prior to the introduction of the quality assurance programme for the Standardization of Radiotherapy (START) trial in breast cancer, a UK multicentre randomized trial of different dose fractionations for breast radiotherapy. Analysis of patient treatment plans was performed at this initial stage of the quality assurance programme to ensure eventual uniformity of treatment within the randomized trial and hence ensure reliable end results.
As an integral part of this initial survey, three patient outlines of different size and shape were circulated between November 1997 and January 1998 to 56 UK radiotherapy centres. Dose distributions were produced according to the routine planning protocol of each department to provide information on treatment planning techniques. Criteria used for treatment plan production and the resultant dose distributions were analysed. The dose distributions varied between centres. Dose inhomogeneity of no more than 10% was achieved, on the central axis, for all chest wall and medium breast size plans. The number of larger breast size distributions exceeding a 10% dose gradient across the treatment volume was 54% (26). Most centres in the UK determine the breast dose distribution by planning on a two-dimensional contour taken along the central plane of the breast. Variation in the breast contour either side of this central plane is not taken into account. Care with plan optimization by selecting the most appropriate beam parameters can lead to an improvement in breast dosimetry.
|
Introduction
|
|---|
The Standardization of Radiotherapy (START) trial compares fractionation schedules used for post-operative treatment in women with early stage breast cancer. The background and details of the trial have previously been reported [1]. A quality assurance (QA) programme is vital to validate the mode of treatment used in different centres, since the reliability of trial results depends on the uniformity of treatment throughout these centres [2]. To ensure this, baseline information was required from all potential trial centres on radiotherapy technique and planning. Questionnaire responses regarding patient immobilization, technique and dose fractionation used to treat breast or chest wall, supraclavicular fossa, axilla and boost to the tumour bed have been previously published [3]. This paper examines more closely the dose distributions produced by centres in breast planning using three sample patient outlines and compares these treatment plan results with aspects of the pre-trial questionnaire.
|
Materials and methods
|
|---|
The patient outlines were circulated together with the initial questionnaire to 56 UK radiotherapy centres, which expressed an interest in the START trial, between November 1997 and January 1998. Completed questionnaires and treatment plans were returned to the quality assurance team based at Mount Vernon Hospital, Northwood, UK.
46 questionnaires were returned, an overall response rate of 85%. One centre used two different planning systems and submitted two sets of plans and another centre submitted plans but not the questionnaire, resulting in 48 sets of plan data altogether. Electrons were used for chest wall treatments in two centres so plans were not provided; another centre did not plan the large breast as only a cobalt treatment machine was available. Data from the initial questionnaire and the treatment plans have been combined to assess the treatment plan distribution from each department.
The outlines are shown in Figure 1
. They were taken from patients and comprised a right chest wall, baseline separation of 16.6 cm and maximum lung depth of 1.2 cm, a left medium sized breast of 18.6 cm separation, maximum lung depth 0.7 cm and a right larger sized breast outline with a separation of 25.3 cm and a maximum lung depth of 1.2 cm. The three outlines were planned according to each individual department's protocol to show typical isodose distributions. A second set of plans was produced, normalized to the reference point used in the START trial. It is this set of plans that have been analysed. The reference point used for the START trial is positioned at or near the centre of the target volume as stated by ICRU 50 [4], and is defined as half way between the lung surface and the skin surface on the perpendicular bisector of the posterior beam edge (Figure 2). This guaranteed that for all trial patients the reference point would remain within breast tissue, not in lung or air, ensured that the dose at this point was reasonably accurate and reduced variations in dose prescription between centres.
View larger version (9K):
[in this window]
[in a new window]
|
Figure 1. Sample patient outlines of varying sizes sent to departments, together with the initial START trial quality assurance questionnaire, for planning according to each individual department's protocol. (a) Outline for chest wall, (b) outline of medium breast size and (c) outline of larger breast size.
|
|
View larger version (11K):
[in this window]
[in a new window]
|
Figure 2. Positioning of the START trial reference point; all plans analysed were normalized to the reference point, R. a, posterior beam edge; b, perpendicular distance between lung surface and skin surface.
|
|
Standard criteria for planning were not stipulated at this stage and no restrictions were set. The exercise was to observe each department's individual criteria for the production of a treatment plan, looking specifically at tangential field orientation, use of wedges, inhomogeneity corrections and dose homogeneity. The isodose distributions for the three outlines from each centre were assessed independently by two members of the QA team.
To allow comparison of wedge angles used at different centres, an equivalent angle (
0) was calculated for universal wedge data [5], using the equation:
where w1 and w2 are the relative contributions to the dose for wedged and non-wedged components, respectively.
Dose gradients were used to assess plan homogeneity. Maximum and minimum dose values were recorded and the dose gradient taken as the difference between these two values. Centres were asked to state their usual dose variation across the breast by defining a maximum and a minimum dose. These limits have been used to give a single dose gradient value for breast planning. It was not specified in the questionnaire whether point doses or dose areas should be given for minimum and maximum values. However, for the analysis of the plans the maximum value was defined as the isodose encompassing a 2 cm2 area and the minimum dose as a point dose excluding build up and penumbra regions. This difference should be considered when the questionnaire and treatment plan dose gradient data are compared.
|
Results
|
|---|
Questionnaire
Isodose distribution and dose gradient
Criteria for the production of a treatment plan varied between centres, as did definitions of an acceptable dose distribution. Table 1 shows a summary of the accepted values for dose gradient across the breast for all centres.
View this table:
[in this window]
[in a new window]
|
Table 1. Dose gradient across the breast as stated on the questionnaire (defined by the difference between maximum and minimum dose values)
|
|
Some centres aimed to cover a defined treatment volume with a specific isodose and others attempted to produce a typical isodose distribution that positioned areas of maximum dose evenly or unevenly throughout the treated area.
Lung
An indication of the extent of lung within the tangential fields was given by 29 (63%) centres. 10 of these centres gained their information from CT imaging, the remaining 19 either indicated maximum lung depth only, defined from the simulator film or image (12 centres), or estimated the shape of the lung contour using other methods. These other methods included the use of standard lung templates in two centres, the patient's inferior chest wall outline in a further two centres and additional screening angles to approximate the lung shape in three centres.
Beam energy
A variety of treatment energies were used for breast irradiation, ranging from cobalt-60 to 15 MV photons. The most common energy used was 6 MV. Machines with higher energy photons (above 6 MV) were available for breast treatments in 22 (48%) departments. In all but one of these the choice of energy for treatment was influenced by breast size; 8 MV or 10 MV would be used to treat patients with very large separations. It was found that the initial choice of using higher beam energy was made prior to the production of a treatment plan and based solely on patient size or separation. The isodose distribution was stated as a primary consideration in only 3 (14%) centres. Two departments, in order to achieve the optimum isodose distribution, used a combination of 6 MV and 10 MV, treating with both energies on the same day. Other field parameters are shown in Table 2.
Plans
The beam energies used by departments to plan the three outlines are shown in Figure 3.
The majority of centres produced isodose distributions that aimed to position areas of maximum dose evenly through the breast. Plans were also seen resulting in distributions with the maximum dose towards either the apex or the medial and lateral aspects of the breast. The results are shown in Table 3; also indicated is the use of lung density correction factors.
The percentage value for each dose gradient measured from the treatment plans was compared with the acceptable dose gradient values in the questionnaire (where information was provided by centres) to assess whether practical limits were being set.
For the chest wall plans all measured gradients were within the questionnaire stated values for each centre, the majority having less than 3% gradient. Slightly higher gradients were seen in three centres (gradients of 5–8%) where either no wedge or small wedge angles were used to plan this outline.
When planning the medium breast size, three centres were outside their specified dose gradient. In these cases the target values stated on the questionnaire were low compared with values stated by other centres (5%, 6% and 8%). All kept the dose gradient within 10%. Of the 40 centres that defined an acceptable gradient 62.5% (25) did not keep within this for the larger breast size outline. The beam energy used to plan the large outline and the resultant dose gradient is shown in Figure 4. All dose gradients quoted are from central slice isodose distributions. Seven centres routinely produce distributions off axis, but these were not considered as part of this study.
View larger version (22K):
[in this window]
[in a new window]
|
Figure 4. The beam energies used to plan the larger breast size outline and the resultant central axis dose gradient demonstrating that a number of departments exceed a 10% gradient for this size of patient.
|
|
The range of wedge angles used is shown in Figure 5. The chest wall data show the use of a large range of wedge angles across centres. Only four centres set the field width using full asymmetry resulting in a non-divergent posterior beam edge, two limiting its use to the chest wall outline only. Three further centres used partial asymmetry to position the field isocentre appropriately, most commonly for the chest wall to bring the isocentre from air into tissue.
View larger version (16K):
[in this window]
[in a new window]
|
Figure 5. Box plot of wedge compensation. Horizontal dashed line shows median, box represents interquartile range, bars represent range of values excluding outliers (these are shown as * and  ). Medial field (MED), lateral field (LAT). The chest wall results demonstrate a large range of angles in use across centres. The data also suggest that the majority of departments surveyed use similar wedge angles for both medial and lateral fields.
|
|
|
Discussion
|
|---|
The breast has a complex shape, with changes in contour and tissue separation throughout the length of the treatment volume. Interpatient variation in breast size and shape, beam obliquity and dose limiting organs at risk (lung and heart), increase the problems with treatment planning. Care with planning optimization by selecting the most appropriate beam parameters, along with consistent patient positioning can lead to an improvement in breast dosimetry. However, most centres in the UK determine the breast dose distribution by planning on a two-dimensional (2D) contour taken along the central plane of the breast. Variations in the breast contour either side of this central plane are not taken into account and there is increasing evidence that dose inhomogeneity within the breast is greater than at other anatomical sites, particularly in women with large breasts [6].
In the majority of departments, the aim for breast planning is to produce a homogeneous dose distribution on the central axis outline. Where high dose areas occur, these are evenly distributed. The maximum lung depth and lung volume treated have an impact on the overall isodose distribution the appearance of which differs depending on whether or not lung density correction is used. Lung tissue has a lower electron density than muscle and fat, resulting in less attenuation of the primary beam. Unless corrections are made for lung density, it is likely that the true dose delivered to the posterior aspect of the breast will be higher than that shown by the representative dose on the treatment plan. This effect will be compounded in cases where the resultant distribution, without correction for lung density, shows high dose regions in the medial and lateral areas. This type of distribution was seen in four centres submitting treatment plans for the standard outlines (Table 3
). Discrepancies in dosimetry when changing from unit density to lung inhomogeneity corrected plans have been assessed [7, 8] concluding that the influence of lung tissue on the dose distribution can result in a calculated dose increase of up to 7%. Webb et al [9] have shown dose inhomogeneities on the central axis ranging from 4% to 9% on non-lung corrected plans whereas the actual dose inhomogeneity ranged between 7% and 18% when lung density correction was incorporated.
Where no lung density correction factor is included in the treatment plan calculation, a distribution with a high dose region at the breast apex may intentionally be produced. This will compensate for the reduced absorption of radiation through the lung in the treatment field and resultant higher dose in adjacent tissues. Where lung correction is not applied, the choice of wedge filter should take into account the expected increase in dose to the medial and lateral parts of the breast due to reduced attenuation by the lung. Where centres used no wedge or very low wedge angles to plan the medium breast size and larger breast size outlines, this type of distribution results. Nine centres planned with the maximum dose region towards the apex of the breast; seven of these did not use lung correction factors in calculating the resultant plan and accounted for this in their isodose distribution. Some assessment should be made of the amount of lung in the treatment fields and its impact on the resultant dose distribution even when CT is not available. Accurate estimation of lung volume is only possible with CT outlining of the chest wall but other estimates are preferable to no correction at all.
Wedge compensators are used to compensate for the breast contour in one plane. Where a wedge was used to plan the chest wall outline the size of the wedge angle appears to have only a small impact on the distribution. Consistent with the previous data for the medium breast size and larger breast size, those centres using no wedge for the chest wall produced a distribution which had higher doses towards the anterior aspect of the chest wall.
The beam energy has implications for production of the resultant dose distribution. When planning patients with a very large separation it is unlikely that a dose gradient of 10% or less can be achieved using 6 MV. However, higher beam energies may compromise the dose to superficial tissues. Although surface dose does not appear to differ significantly between 6 MV and 10 MV, for open fields, (the depth of the 90% isodose) when using photon energies of 6 MV is ![{els]](/math/els.gif)
le;4 mm, whereas for 10 MV photon beams a dose of 90% of the prescription dose occurs at a maximum depth of 8 mm [10]. This difference in the depth of the 90% isodose may be compensated in part by using the lower and higher beam energies in combination.
In most cases, an adequate dose distribution is achieved using appropriate beam energy and wedge filters [11], the exception being those patients with large breast separations. The number of plans for the larger breast size in this study having a greater than 10% dose gradient was 54% (26). A considerable number, 33% (16), had a resultant dose gradient of above 12% (not achieving the degree of homogeneity as described by ICRU 50). A further study [12] has demonstrated suboptimal dosimetry in a sample of patients, when looking at 3D data, and suggests that the degree of dose heterogeneity is significantly correlated with breast size. It was reported that an inhomogeneous dose distribution might result in breast tissue receiving doses of either less than or more than 5% different to the prescribed dose and that this increase or decrease in dose can be linked to breast volume. Areas of low dose may contribute to local treatment failure while high dose areas result in increased dose leading to increased late normal tissue radiation morbidity and possibly a poor cosmetic result.
Patients with larger breast sizes do have proportionately greater changes in the external contour in both the cranial and caudal direction away from the central axis and in most cases the maximum and minimum doses will not be represented on the central axis slice [11, 13]. In these cases Cheng et al [14] recommend 3D planning to appreciate the actual dose distribution throughout the treatment volume and using 3D compensation to improve the dosimetry for this subset of patients.
Dose to the contralateral breast is increased as a result of scattered radiation and is influenced by a number of factors including wedge compensation and divergent tangential beams. In this study, wedge compensation was achieved using either a fixed metal wedge (internal or external), a universal wedge located in the head of the treatment machine or a dynamic wedge where the movement of the field defining jaws determines the extent of wedging.
Use of external fixed wedge systems contributes an increased dose to the contralateral breast [15, 16]. When using fixed wedges, placement of all or most of the wedge on the lateral field can produce a reduction of dose to the contralateral breast [17]. Ikner et al [18] reported on wedging the lateral field only and concluded that a reasonable dose distribution could be obtained for those patients with separations of le;22 cm. Our data suggest that the majority of centres in the UK use similar wedge angles for both medial and lateral fields with no attempt to reduce the contralateral breast dose by using uneven wedging.
Geometric divergence of the tangential treatment fields can either be calculated to align the posterior edges of the tangential fields and achieve beam coincidence or beam asymmetry used to achieve half beam blocking. Correcting for the divergence of the tangential beams has the additional advantage of avoiding unnecessary irradiation of underlying lung and heart.
|
Conclusion
|
|---|
The baseline information from this questionnaire has enabled a number of observations to be made with respect to breast planning in the UK. For the majority of patients it is possible to obtain an acceptable distribution on the 2D central axis breast plan. However, away from this plane the radiation dose may be much less homogeneous. This is especially true for women with larger breast sizes. There is also a need for appreciation of the impact of lung tissue irradiated within tangential breast treatment fields and the compensatory methods required to achieve improved dosimetry.
In conclusion, this initial survey has allowed an insight into the criteria used for the production of dose distributions for tangential breast irradiation and highlighted possible inconsistencies between centres in use of available beam energies, wedge filter and planning software and their perception of the optimal dose distribution in this setting. The quality assurance programme for the trial will monitor dose distributions for one in three patients randomized, recording data with respect to irradiation using modern techniques.
|
Footnotes
|
|---|
START Trial Management Group included Edwin Aird, Jane Barrett, Peter Barrett-Lee, Judith Bliss, Jackie Brown, Clare Cruickshank, Amanda Deighton, John Dewar, Jane Dobbs, Sue Griffiths, Penny Hopwood, Peter Hoskin, Eileen Parkin (RAGE Observer), Caroline Jackson, Pat Lawton, Brian Magee, David Morgan, Roger Owen, Joyce Pritchard (RAGE Observer), Val Speechely, David Spooner, Karen Venables, Elizabeth Winfield and John Yarnold. 
The START trial is funded jointly by The Medical Research Council, the Cancer Research Campaign and the Department of Health, developed under the auspices of the UKCCCR Breast Cancer Subcommittee.
Received for publication May 15, 2002.
Revision received November 15, 2002.
Accepted for publication January 23, 2003.
|
References
|
|---|
- START Trial Management Group. Standardisation of breast radiotherapy (START) trial. Clin Oncol 1999;11:145–7.
- Ptaszynski A, Van der Bogaert W, Horiot J, Poortmans P, Fourquet A, Struikmans H, et al. Radiation dose homogeneity in an EORTC multi-centre trial on breast irradiation. Acta Oncologica 1999;13:19–27.
- Winfield E, Deighton A, Venables K, Hoskin PJ, Aird EGA. Survey of UK breast radiotherapy techniques: Background prior to the introduction of the quality assurance programme for the START (standardisation of radiotherapy) trial in breast cancer. Clin Oncol 2002;14:267–71.
- International Commission on Radiation Units, Measurements. Prescribing, recording and reporting photon beam therapy, ICRU Report 50. Bethesda, MD: ICRU, 1993.
- Redpath AT, Williams JR, Thwaites DI. Treatment planning for external beam therapy. In: JR Williams, DI Thwaites, editors. Radiotherapy physics in practice. Oxford: Oxford University Press, 1993:135–85.
- Neal AJ, Mayles WPM, Yarnold JR. Invited review: Tangential breast irradiation—rationale and methods for improving dosimetry. Br J Radiol 1994;67:1149–54.[Abstract]
- Redpath AT, Thwaites DI, Rodger A, Aitken MW, Hardiman PDJ. A multidisciplinary approach to improving the quality of tangential chest wall and breast irradiation for carcinoma of the breast. Radiother Oncol 1992;23:118–26.[Medline]
- Mijnheer BJ, Heukelom S, Lanson JH, van Battum LJ, van Bree NAM, van Tienhoven G. Should inhomogeneity corrections be applied during treatment planning of tangential breast irradiation? Radiother Oncol 1991;22:239–44.[Medline]
- Webb S, Leach MO, Bentley RE, Maureemootoo K, Yarnold JR, Toms MA, et al. Clinical dosimetry for radiotherapy to the breast based on imaging with the prototype Royal Marsden Hospital CT simulator. Phys Med Biol 1987;7:835–45.[CrossRef]
- McCullough EC. A measurement and analysis of buildup region dose for open field photon beams (Cobalt-60 through 24 MV). Med Dosim 1994;19:5–14.[Medline]
- Moody AM, Mayles WPM, Bliss JM, A'Hearn RP, Owen JR, Regan J, et al. The influence of breast size on late radiation effects and association with radiotherapy dose inhomogeneity. Radiother Oncol 1994;33:106–12.[CrossRef][Medline]
- Neal AJ, Torr M, Helyer S, Yarnold JR. Correlation of breast dose heterogeneity with breast size using 3D CT planning and dose-volume histograms. Radiother Oncol 1995;34:210–8.[Medline]
- Vincent D, Beckham W, Delaney G. An assessment of the number of CT slices necessary to plan breast radiotherapy. Radiother Oncol 1999;52:179–83.[Medline]
- Cheng C, Das IJ, Tang W, Chang S, Tsai J, Ceberg C, et al. Dosimetric comparison of treatment planning systems in irradiation of breast with tangential fields. Int J Radiat Oncol Biol Phys 1997;38:835–42.[Medline]
- Warlick WB, O'Rear JH, Earley L, Moeller JH, Gaffney DK, Leavitt DD. Dose to the contralateral breast: a comparison of two techniques using the enhanced dynamic wedge versus a standard wedge. Med Dosim 1997;22:185–91.[Medline]
- Weides CD, Mok EC, Chang WC, Findley DO Shostak CA. Evaluating the dose to the contralateral breast when using a dynamic wedge versus a regular wedge. Med Dosim 1995;20:287–93.[Medline]
- Epstein RJ, Kelly SA, Cook M, Bateman A, Paddick I, Kam KC, et al. Active minimisation of radiation scatter during breast radiotherapy: management implications for young patients with good prognosis neoplasms. Radiother Oncol 1996;40:69–74.[Medline]
- Ikner CL, Russo R, Podgorsak MB, Proulx GM Lee RJ. Comparison of the homogeneity of breast dose distributions with and without the medial wedge. Med Dosim 1998;23:89–94.[CrossRef][Medline]
Top
Abstract
Introduction
