British Journal of Radiology 75 (2002),812-818 © 2002 The British Institute of Radiology
A novel approach of dose mapping using a humanoid breast phantom in radiotherapy
D P Banjade, PhD, AMINSTP
1
B S Ng, BSc, MSc
2
M Zakir, Dip RT
2
A A Tajuddin, PhD
1 and
A Shukri, PhD
1
1 School of Physics, Universiti Sains Malaysia, 11800 Penang and 2 Hospital Pantai Mutiara, 82 Jalan Tengah, Bayan Baru, 11900 Penang, Malaysia
Correspondence: D P Banjade, BP Koirala Memorial Cancer Hospital, Bharatpur Chitawan, Nepal
 |
Abstract
|
|---|
A study of dose mapping techniques to investigate the dose distribution throughout a planned target volume (PTV) in a humanoid breast phantom exposed to a 6 MV photon beam similar to that of treatment conditions is described. For tangential breast irradiation using a 6 MV accelerator beam, the dose is mapped at various locations within the PTV using thermoluminescent dosemeters (TLDs) and radiographic films. An average size perspex breast phantom with the ability to hold the dosemeters was made. TLDs were exposed after packing them in various locations in a particular slice, as planned by the treatment planning system (TPS). To map the dose relative to the isocenter, films were exposed after tightly packing them in between phantom slices, parallel to the central axis of the beam. The dose received at every location was compared with the given dose as generated by the TPS. The mapped dose in each location in the isocentric slice from superficial to deep region was found to be in close agreement with the TPS generated dose to within ±2%. Doses at greater depths and distant medial and lateral ends, however, were found to be lower by as much as 9.4% at some points. The mapped dose towards the superior region and closest inferior region from the isocenter was found to agree with those for TPS. Conversely, results for the farthest inferior region were found to be significantly different with a variance as much as 17.4% at some points, which is believed to be owing to the variation in size and shape of the contour. Results obtained from films confirmed this, showing similar trends in dose mapping. Considering the importance of accurate doses in radiotherapy, evaluating dose distribution using this technique and tool was found to be useful. This provides the opportunity to choose a technique and plan to provide optimum dose delivery for radiotherapy to the breast.
|
Introduction
|
|---|
Apart from accuracy of the dose at the point concerned, a uniform dose distribution within the target volume is also crucial for successful radiotherapy. This is especially true in complex field geometry such as tangential beam irradiation in carcinoma of the breast. Recently, Perrin et al [1] described a phantom design to audit treatment of the prostate by radiotherapy, suggesting the possibility of developing a phantom to simulate other treatment sites. In the past, there have been concerns over radiotherapy techniques, including tangential breast irradiation, which aims to give a uniform dose to the target volume [2–6]. The irradiation technique to treat the intact breast for early stage breast cancer usually consists of two tangential parallel-opposed photon beams. Additional fields, such as electron irradiation, are also introduced to boost the dose to skin or post-operative scar depending on clinical need. It is generally accepted that variance in the dose delivered to the patient should not be greater than 5% at the reference point [7]. More recently, a tolerance of 3.5% has been suggested [8]. Subsequently, the International Commission on Radiation Units and Measurements (ICRU) report No. 50 [9] has recommended dose homogeneity of between -5% and +7% of the prescribed dose throughout the planned target volume (PTV). Homogeneity better than this could be achieved if the intact breast has full scatter conditions for the applied beams and minimal discrepancies in contour outline from superior to inferior border. However, dose inhomogeneities greater than this recommendation normally occur owing to the lack of full scatter condition, complexity in beam arrangement and the inconsistent shape of the PTV, which are almost inevitable in the breast. Furthermore, the geometry of water phantom measurements that are used for beam modelling in computer treatment planning systems (TPSs) does not reflect the geometry used in breast treatment, even though TPSs employ various corrections, including a scatter correction algorithm to improve dose distribution throughout the PTV. In breast irradiation, the PTV varies significantly from superior through to inferior, as well as from medial to lateral, depending on the original size and post-operative deformities of the patient's breasts. Furthermore, treatment planning for breast cancer is complex because of the close vicinity of vital organs, i.e. lungs and heart underneath the chest wall. The best treatment can only be achieved by directing the photon beams tangentially or irradiating the chest wall using electron beams. Although the transverse obliquity of the contour can be compensated for to some extent by using wedges, complete compensation of the variation is impossible in most cases owing to non-uniformity of coronal contours of the chest wall.
Modern TPSs provide opportunities to optimize dose distribution by manipulating the beams and technique to overcome the difficulties caused by target volume location. There is a unique opportunity to enhance treatment accuracy based on three-dimensional images. Using these images, dose distribution can be optimized by choosing the options available, such as wedges, and optimum beam arrangements. Despite all attempts to achieve better dose distribution, ideal field arrangements and planning are still difficult. Therefore, considerable care is required to choose the beams, treatment techniques and accessories to deliver optimum dose to the PTV, minimizing possible errors while treating the breast. In addition, to ensure better treatment, in vivo dosimetry and dose mapping inside the volume are necessary to optimize the dose distribution throughout the volume.
Neal et al [10, 11] found greater dose inhomogeneities than the ICRU recommendation with tangential breast irradiation, particularly in women with large breasts. Knoos et al [12] also found that the absorbed doses measured in the breast were 2–6% lower than those calculated using the TPS. A significant volume of high dose (hot spot) has also been reported [3, 11] in large breasted women owing to rapidly changing contour, as well as larger separation. Total irradiated volume of the breast is significantly smaller than most other clinical volumes. Verification of dose distribution in breast planning is particularly important as dose distribution for tangential breast irradiation might not follow the expected pattern as shown by the computer. Therefore verification of dose distribution through a dose mapping technique could provide more options to comply with the requirements of the radiotherapy technique. It is also important to identify the factors that can be used to manipulate the beams using the various options available. Dose mapping in a humanoid breast phantom seems to be one of the quality assurance methods for checking a computer generated dose distribution prior to breast irradiation. Therefore a novel and practical dose mapping technique is proposed that would help the physicist analyze computer generated treatment plans and confidently accept parameters or re-evaluate plans to optimize treatment technique.
|
Materials and methods
|
|---|
A breast phantom of average size and shape was made using 1 cm thick perspex slices, which were shaped appropriately. The slices were fixed in a block and the outer edge of each slice was made smooth. 1 mm diameter horizontal holes were drilled in various locations in the slices in relation to the isocentre of the central slice in order to hold the thermoluminescent dosemeters (TLDs) within the phantom. As well as the central slice, one superior and two inferior slices were selected that were assumed liable to receive a higher or lower dose because of differences in their contour and the scattering cross-sectional area of the beam compared with the isocentric slice. The superior slice was 5 cm away from the isocentric slice whereas the inferior slices were 2 cm and 6 cm away from the isocentric slice, where the off centre dose was measured. Treatment fields were simulated using a simulator (Simview NT; Siemens Inc., Concord, CA) and planned using a plato TPS (Nucleotron, Veenendaal, The Netherlands) using an isocentric technique. A 6 MV beam from a Siemens Primus version 7 linear accelerator (Siemens Inc.) with dual energy photon and electron beams was used in this study. Lateral and medial tangential fields were exposed with a 15° wedge to provide a 1 Gy dose to the isocentre and TPS-defined volume. Since the reliability of LiF TLD and film dosemeters were already established for in vivo measurements [13, 14], 1 mm x 6 mm LiF:Mg:Ti TLD rods (TLD-100, Bicron NE, OH) and Kodak ready packed films (Eastman Kodak, New York, NY) were chosen as radiation detectors. TLDs were exposed by packing them in pre-determined locations in a particular slice, as planned by the TPS. Exposed TLDs were then read in a Harshaw model 3500 (Harshaw, OH) TLD reader. Films were also exposed after tightly packing them in between phantom slices, parallel to the central axis of the beam in exactly the same position and geometrical arrangement as the TLDs. For the isocentric slice, relative dose distribution using the films was also compared with TLD results. Film density was measured with a RMI (RMI, Middleton, WI) densitometer.
TLDs were calibrated against the absolute dose of 1 Gy in a "plastic water" (PW) phantom (Nuclear Associates, Chicago, IL) using the same beam, and a conversion factor was applied to correct the dose received by the perspex phantom. Since the results of different radiotherapy protocols show very small differences [15], any standard protocol can be used to calibrate beam output. However, in this particular study the linear accelerator was calibrated using the AAPM Task Group 21 protocol [16] in accordance with the departmental calibration practice. In addition, the AAPM TG-21 protocol provides an algorithm to convert the dose from one medium to another as follows: 
where Dwater(dmax) is the dose (D) in water at depth (d) of dose maximum, Dmed is the dose in measuring medium at the same point, ESC is an excess scatter correction factor and 
is the mass energy absorption coefficient ratio of water to medium.
Measuring dose using TLDs, the dosemeter response signal produced by an individual TLD is converted to dose by applying the calibration factor (CFq), which is derived as follows:
where "Calibration dose" is the given dose at the depth of calibration and TL is the dosemeter response of the TLD in a PW phantom from a particular beam with the beam quality "q". The CF and measured TL readings from the dose received by that particular TLD in the phantom are then used to determine the real dose (D) in the breast phantom, which can be expressed as:
where 
is the total dosemeter correction factors for linearity, field size, orientation and fading effect of the dosemeters. The parameter 
is applied because the measuring medium is different from the calibrating phantom based on Equation 1
.
|
Results
|
|---|
The aim of this study was to assess the actual dose distribution in all directions from the isocenter to compare with the dose distribution planned by the TPS, to demonstrate that dose inhomogeneity in tangential breast irradiation is greater than expected, and to prevent any hidden discrepancies in dose. Before measurements were carried out, the isodose distribution is generated by the computer TPS, which is illustrated in Figures 1 and 2. The plan was optimized to cover the whole PTV by 100% isodose line focussing the central slice (Figure 1a). After careful optimization of the plan, the dose distribution for other slices in the superior to inferior direction was also studied to judge if it was covered by the equivalent isodose line (Figure 1b). However, the equivalent distribution of the central slice could only be achieved in the nearest inferior slice (Figure 2a). A satisfactory distribution in the farthest inferior and superior slices could only be achieved around the shallow area and the coverage of dose gradually fell towards the inferior part of these slices (Figure 2b-c). The distribution is significantly low in the farthest inferior slice, up to 80% near the interior border (Figure 2c) of the beams. Nevertheless, the measurement points in this area were selected within 90% isodose line and compared with the measured dose. A hot spot was also seen at superficially bulging areas of the contour. The mapped dose in all chosen slices, including the isocentric slice, was analyzed to identify discrepancies in measured and planned dose distribution throughout the PTV.
View larger version (46K):
[in this window]
[in a new window]
|
Figure 1. (a) The treatment planning system (TPS) generated two-dimensional dose distribution in the isocentric slice of the breast phantom. (b) The TPS generated dose distribution in the isocentric slice of the breast phantom (a three-dimensional observer's eye view image).
|
|
View larger version (67K):
[in this window]
[in a new window]
|
Figure 2. (a) The treatment planning system (TPS) generated dose distribution in the slice 2 cm inferior from the isocentric slice of the breast phantom. (b) The TPS generated dose distribution in the slice 6 cm inferior from the isocentric slice of the breast phantom. (c) The TPS generated dose distribution in the slice 5 cm superior from the isocentric slice of the breast phantom.
|
|
Results obtained from the TLD and film dosemeters in the isocentric slice were grouped according to their locations/point of interest in the irradiation volume. As shown in Figure 3, measurement points in the isocentric slice were classified into the regions of shallow, deep1, deep2 and deep3. The doses measured in this slice from shallow to deep2 regions using TLDs were found to be in close agreement, to within ±2%, compared with the TPS generated dose. The doses at greater depths and at distant medial and lateral ends, however, were found to be lower by as much as 9.4% at some points. The comparative results of dose mapping obtained from the films along with the TLDs are summarized in Figure 4. Results obtained from films are the normalized results to the isocenter density at which the dose is supposed to be proportional to the TPS dose, i.e. 1 Gy. As illustrated, the relative density corresponding to the TLD measurements shows a similar trend, or is closer to the TPS, than that of TLD results. However, since density values are relative to the isocenter, the dose cannot be predicted accurately even though results are closer. Nevertheless, these are sufficiently reliable for dose mapping purposes. Figure 5 provides the results of mapped doses in off axis slices inside the PTV using TLDs. As indicated in Figure 5, the corresponding results obtained in superior and inferior slices exhibited a similar trend as the central slice, but the measured dose in some points differed significantly from those for TPS generated values. The variation between mapped and TPS doses in the slice 2 cm inferior from the isocenter was found to be compatible, the maximum discrepancies being not more than 6.8% (Figure 5a). Similar measurements for the slice 6 cm inferior from the isocenter showed significant differences from those planned by the TPS, especially in the peripheral regions (Figure 5b). As shown in the figure, the mapped dose is found to be significantly lower than that of the TPS at distant lateral and medial regions by as much as 17.4% and 11.3%, respectively. This must be owing to the variation of contour shape and larger separation than that of the central slice, which is of course the focus of treatment planning. In addition, the build up/down effects might be responsible to some extent for a lower dose towards the beam entry and exit areas. However, the results for the distant superior slice 5 cm superior to the isocentre are within the maximum discrepancies of not more than 6.7% with those of the TPS (Figure 5c), as the slice contour is not very different from the central slice. The overall results of measured and planned doses are in agreement, showing the same pattern of dose distribution in most cases. However, the results obtained from the off-center slices show that the computer plan might overestimate the dose to some extent, particularly towards the inner and distant medial and lateral border of the volume, if the contour of the PTV varies significantly. This indicates that the measured and TPS doses are comparable, provided that the contour is uniform throughout the PTV and measurements are not influenced by the build up/down effect of the beams. Conversely, the distribution would vary significantly, especially in entry and exit regions, if the criteria are not met.
View larger version (90K):
[in this window]
[in a new window]
|
Figure 4. Results of mapped dose in an isocentric slice (Figure 1a) of the phantom, using thermoluminescent dosemeters (red data) and films (blue data) against the treatment planning system dose of 100 cGy throughout the planned target volume.
|
|
|
Discussion and conclusion
|
|---|
Although the TPS dose is considered reasonably accurate, there may be some intrinsic errors in complex beam geometries, such as tangential breast irradiation, because of the many algorithms and extrapolation/interpolation involved, as well as corrections needed for dosimetry protocols. Furthermore, it seems more difficult to compensate for the surface irregularities of the breast contour from superior to inferior. Compensation could only be possible by using multidirectional wedges or using a customized tissue compensator, which is a rather complex procedure and may increase risk by introducing other complications in treatment. Therefore, planning practice should involve the whole area of the target volume from superior to inferior as well as in vivo dose mapping to provide optimum dose to the PTV. In this article we presented a practical approach of dose mapping to estimate the dose at various points of interest in a breast phantom from superior to inferior slices, so as to detect any possible uncertainties in dose distribution within the PTV. As a result, a low dose area can be compensated by using additional options, including treatment such as a boost with an electron beam. Estimated uncertainties in the resulting dose at every measurement point in this study were observed to be less than ±3%. From these results we recommend that any significant dose variation due to contour irregularities should be taken into account to avoid under dosing or over dosing, which could lead to tumour recurrence or normal tissue complications in radiotherapy.
Received for publication December 3, 2001.
Revision received May 7, 2002.
Accepted for publication May 8, 2002.
|
References
|
|---|
- Perrin BA, Jordan TJ, Hounsell AR. The design and evaluation of a phantom for audit of the treatment chain for prostate radiotherapy. Radiother Oncol 2001;60:37–43.[Medline]
- Baird CT, Starkschall G, Liu HH, Buchholz TA, Hogstrom KR. Verification of tangential breast treatment dose calculations in a commercial 3D treatment planning system. J Appl Clin Med Phys 2001;2:72–84.
- Buchholz TA, Gurgoze E, Bice WS, Prestidge BR. Dosimetric analysis of intact breast irradiation in off-axis planes. Int J Radiat Oncol Biol Phys 1997;39:261–7.[Medline]
- Cozzi L, Fogliata-Cozi A. Quality assurance in radiation oncology. A study of feasibility and impact on action levels of an in vivo dosimetry program during breast cancer irradiation. Radiother Oncol 1998;47:29–36.[Medline]
- Delaney G, Beckham W, Veness M, Ahern V, Back M, et al. Three dimensional dose distribution of tangential breast irradiation: results of multicentre phantom dosimetry study. Radiother Oncol 2000;57:61–8.[Medline]
- Kirsner SM, Prado KL, Tailor RC, Bencomo JA. Verification of the accuracy of 3D calculations of breast dose during tangential irradiation: measurements in a breast phantom. J Appl Clin Med Phys 2001;2:149–56.[Medline]
- International Commission on Radiation Units and Measurements. ICRU Report 24. Determination of absorbed dose in a patient irradiated by beams of X-or gamma-rays in radiotherapy procedures. Washington DC: ICRU, 1976.
- Mijnheer BJ, Battermann JJ, Wambersie A. What degree of accuracy is required and can be achieved in photon and neutron therapy? Radiother Oncol 1987;8:237–52.[Medline]
- International Commission on Radiation Units and Measurements. ICRU Report 50. Prescribing, recording, and reporting photon beam therapy. Bethesda, MD: ICRU, 1993.
- 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/Free Full Text]
- Neal AJ, Torr M, Helyer S, Yarnold JR. Correlation of breast dose heterogeneity with breast size using 3D CT planning and dose volume histogram. Radiother Oncol 1995;34:210–8.[Medline]
- Knoos T, Ahlgren L, Nilsson M. Comparison of measured and calculated absorbed doses from tangential irradiation of the breast. Radiother Oncol 1986;7:81–6.[Medline]
- Kron T. Thermoluminescence dosimetry and its applications in medicine—Part 2: history and applications. Australas Phys Eng Sci Med 1995;18:1–25.[Medline]
- van Bree NA, Idzes MH, Huizenga H, Mijnheer BJ. Film dosimetry for radiotherapy treatment planning verification of a 6 MV tangential breast irradiation. Radiother Oncol 1994;31:251–5.[Medline]
- Banjade DP, Tajuddin AA, Shukri A. Determination of absorbed dose for photon and electron beams from a linear accelerator by comparing various protocols on different phantoms. Appl Radiat Iso 2001;55:235–43.
- AAPM (American Association of Physicists in Medicine), TG-21. A protocol for the determination of absorbed dose from high-energy photon and electron beams. Med Phys 1983;10:741–71.[Medline]
Top
Abstract
Introduction
, film; 
, thermoluminescent dosemeter (TLD); 
, treatment planning system (TPS).