First published online May 10, 2006
British Journal of Radiology (2006) 79, 828-836
© 2006 British Institute of Radiology
doi: 10.1259/bjr/26575438
Dosimetric and treatment planning considerations for radiotherapy of the chest wall
M M Aspradakis, PhD, MIPEM,
H M McCallum, PhD, MIPEM and
N Wilson, DCR(T)
Regional Medical Physics Department, Newcastle General Hospital, Westgate Road, Newcastle upon Tyne NE4 6BE, UK
Correspondence: Dr Hazel M McCallum
 |
Abstract
|
|---|
Radiotherapy treatment planning calculations of the chest wall are complex due to missing tissue, the thin chest wall and the presence of lung. The accuracy of calculated dose is dependent on the type of algorithm employed. This work evaluates a collapsed cone (CC) and a pencil beam (PB) convolution model for radiotherapy planning of the chest wall. Various irradiation geometries simulating the chest wall have been examined and calculations were compared with measurements with an ionization chamber in epoxy resin water substitute and in low-density lung substitute blocks. A retrospective treatment planning study comprising 6 patients was carried out to evaluate the differences in the dose distributions and monitor units predicted by the two algorithms. The calculated dose in unit density medium was within ±1% for the CC model and up to ±2% for PB. In low density medium and under full scatter conditions, CC overestimated the dose by 1% whereas PB overestimated the dose by 9%. In the tangential irradiation geometry with water and lung media, the PB overestimated dose to the isocentre by up to 10%, whereas the dose from CC was within 3%. From the treatment planning study calculated monitor units (MU) and doses were consistent with the experimental findings. The CC model is more accurate for radiotherapy treatment planning of the chest wall and especially when there is significant involvement of lung tissue.
|
Introduction
|
|---|
Breast cancer affects up to 1 in 10 women in the Western World and is the most common female cancer in the UK in terms of incidence and mortality [1]. Adjuvant radiotherapy to the chest wall has an established role in reducing the risk of locoregional recurrence following mastectomy. A meta-analysis of all randomized controlled trials published in 1987 showed that post mastectomy radiotherapy was associated with a 66% reduction in the risk of locoregional recurrence [2]. Post mastectomy radiotherapy may be indicated either due to inadequate surgical resection of tumour margins due to lymphovascular involvement or due to a tumour size greater than 5 cm.
Radiotherapy treatment planning calculations for the chest wall are complex due to missing tissue and the presence of lung tissue within the treatment field. Accurate calculation of the dose distribution is important as doses to neighbouring organs at risk, such as heart, lungs and contralateral breast, needs to be minimized. Treatment planning systems (TPS) would, ideally, provide both accurate relative dose distributions and monitor unit settings for such irradiation geometries. The Helax-TMS (version 6.1a) TPS (Nucletron B.V., The Netherlands) has two dose calculation algorithms available for external photon beam planning, the pencil beam (PB) and the collapsed cone (CC) algorithms. Dose calculations for patients on Helax-TMS use density information from CT based on the conversion of Hounsfield numbers (HN) to material density, thus accounting for the attenuation properties of different tissues, as well as using information on the material composition. The two algorithms differ primarily in how they model radiation transport and calculate dose in heterogeneous media, with the CC algorithm better approximating the dose directly to the medium [3–5]. Their characteristics and dosimetric accuracy in simple and complex phantom irradiation geometries have been well documented [6–9]. It has been shown that there is little or no clinically significant difference in the calculation of dose by the two algorithms when no large non-unit density heterogeneities are present and under full scatter conditions, such as in the irradiation of the pelvis [9]. This excludes cases of pelvic irradiation with high density metal prosthesis [10].
In cases of treatment planning in the thorax region (for oesophagus and lung), differences between CC and PB in the calculation of dose (predicted monitor units (MU)), planning target volume (PTV) coverage and minimum dose to the PTV has been reported [11]. Such differences can potentially be clinically significant and affect tumour control. At tangential beam irradiations for a range of beam energies (4 MV, 6 MV, 15 MV) on a homogeneous medium, experimental verification of the Helax-TMS algorithms using a 15 cmx15 cm square field had shown the CC algorithm to be accurate to within ±2% of the measured dose and to model closely the reduction in dose due to the missing tissue [9]. From PB the dose close to phantom boundaries was overestimated by 4–5% at 4 MV, but less so at higher energies.
The difficulties encountered with treatment planning of the chest wall in particular have not been reported, nor has the dosimetric accuracy of treatment planning calculations for such extreme irradiation geometries with the presence of lung and/or the plan normalization point within thin tissue or within lung been investigated. It was the purpose of this work to investigate the dosimetric accuracy of the Helax-TMS convolution/superposition algorithms for treatment planning of the chest wall at the Northern Centre for Cancer Treatment (NCCT), in Newcastle upon Tyne, UK. The dosimetric differences between the Helax-TMS CC and PB algorithms were examined for clinically relevant scenarios of chest wall irradiations. A planning study comprising six patients was also undertaken to support changes in clinical practice.
|
Materials and methods
|
|---|
Dosimetric verification of Helax-TMS algorithms for chest wall irradiation
Experimental set-up
All measurements were carried out with a 6 MV photon beam on a Siemens Primus accelerator (Siemens, Erlangen, Germany). This is the most frequently used beam energy at NCCT for treatment of the breast and chest wall. Three experimental geometries relevant to chest wall irradiation were examined. The WT1-tissue and lung substitute blocks (Epoxy resin water substitute blocks with elemental composition: H(8.09); C(67.22); N(2.4); O(19.84); Ca(2.32); Cl(0.13) and relative electron density of 1.02 (mass density 1.04 g cm–3) [12]. Lung equivalent blocks of elemental composition: H(8.38); C(60.50); N(1.68); O(17.28); Cl(0.15); Si(0.84); Mg(11.17) and relative electron density of 0.25 (mass density of 0.26 g cm–3) [13, 14]) used in the measurements were manufactured by ScanPlus (St Bartholomews, London). The measurement system used was an NE2571 0.6 cm3 cylindrical ionization chamber and its associated electrometer with a valid calibration factor (traceable to the primary standard at the National Physical Laboratory) to convert the ionization reading to dose in water.
The irradiation geometries used are shown in Figure 1
. The ionization chamber was positioned in all cases at the isocentre. In Figure 1a, the lateral extent of the phantom normal to the beam central axis varied from 1 cm to 3 cm and a measurement was also made at full scatter conditions with a lateral thickness of 10 cm. In Figure 1b the phantom comprised lung and tissue equivalent blocks. Measurements were carried out with the chamber in the tissue equivalent block at 1 cm, 1.5 cm and 2 cm from the edge of the lung heterogeneity and the phantom boundary varied from 1 cm to 3 cm and at full scatter conditions at 10 cm in the lateral direction. Figure 1c was analogous to Figure 1b, but the calculation point was in lung equivalent material.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 1. Experimental geometries employed for measurements and/or calculations in this work. All measurements were carried out with a 6 MV photon beam, gantry angle of 270° and collimator angle of 90°.
|
|
The position of the normalization/prescription point with respect to the phantom boundary or the lung was chosen to be similar to that often selected in patient treatment plans. Simulations on Helax-TMS were carried out for the slice separations and dose matrix resolutions of 0.25 cm.
Low density dosemeter correction factor
For the measurement of dose in a low density medium using the NE2571 ionization chamber and its absorbed dose to water calibration factor (
), the chamber reading is converted to dose in an infinitesimally small volume of water within the low density medium (which is what is also modelled by the treatment planning system), without accounting for the high replacement perturbation effect caused by the chamber and its high density graphite wall in the low density medium. The NE2571 chamber perturbs fluence differently in water and in low density lung and this difference should be accounted for when converting the ionization reading to dose.
The application of a perturbation correction to the measurement in lung in this work was based on the work of Krieger and Sauer [15]. The replacement perturbation factor prepl(A) at field size A was defined as the ratio of dosemeter reading with zero chamber wall thickness to the reading at normal wall thickness. Krieger and Sauer [15] measured doses in Styrofoam (with mass density of 0.035 g cm–3) for variable chamber wall thicknesses made of plastic (PMMA) and at different field sizes, using a PTW 31003 ionization chamber which has a PMMA wall and wall thickness of 0.0655 cm (or 0.078 g cm–2) and mass density of 1.19 g cm–3. From an exponential curve fit to this data, dose values were calculated at zero wall thickness. For the purpose of our work, we assumed that the atomic compositions of graphite and PMMA are similar, that there is a linear relationship for the perturbation correction with medium density and that differences in the geometric characteristics of the chambers have a small influence on the perturbation in low density medium. The dose values in Styrofoam measured using a chamber with a PMMA wall were converted to dose values in Styrofoam measured with a chamber with a graphite wall. The replacement perturbation factor for the NE2571 in lung was further derived from linear interpolation between replacement perturbation factors for this chamber in Styrofoam and in water (prepl(A) in water is unity). Using this replacement perturbation correction, ionization readings (
) measured for a field size A with the NE2571 in a homogeneous lung medium, were converted to dose to water in an infinitesimally small volume of water within the low density lung using:
|
In the case of a measurement in a heterogeneous medium comprising lung and water where the chamber is placed within lung, a replacement perturbation correction was applied to the dose from the part of the field irradiating the lung medium (
). Thus, if 
is the measured dose in lung in the heterogeneous medium and 
is the replacement perturbation correction for the equivalent field size Aeq irradiating the low density medium, the dose to water in an infinitesimally small volume of water within the lung in the heterogeneous medium is calculated from [15]:
Thus for the derivation of the perturbation correction in the heterogeneous geometry, the dose in a homogeneous lung medium under full scatter conditions was also measured.
Treatment planning study
Standard breast planning technique at NCCT
All patients at NCCT undergoing external beam radiotherapy treatment planning for the breast have a three-dimensional (3D) planning study based on images from helical CT scanning using a 5 mm slice reconstruction. Patients are immobilized in a reproducible position, lying supine on a PosiboardTM-2 breast backrest (manufactured by Sinmed BV) and with couch position registration. Both arms are positioned above the patient's head, facilitating the tangential field arrangements of the treatment plan. Ball bearings are placed at stable points of the patient to define a CT reference point to aid positional verification during treatment delivery. A planning CT scan incorporating these markers is acquired with a 0.5 cm slice separation and this dataset is transferred to Helax-TMS. The patient external contour (skin) and organs at risk are outlined, namely involved lungs and heart, if appropriate.
The isocentric breast planning technique at NCCT treats the breast with two tangential fields and, in some cases, a single supraclavicular field. Non-divergent beam edges are produced at the match plane with the supraclavicular field and at the posterior edge of the tangential fields [16–18]. The plan normalization/prescription point is always set at the isocentre. The generation of a dose distribution within Helax-TMS is a two stage process. A dose plan is generated using fast interactive optimization of open and wedged beams. Interactive dose calculations (in the Beam Modelling module on Helax-TMS) are possible only with the PB algorithm. A final dose calculation is generated once the plan is placed for non-interactive computation (in the Evaluation module). Non-interactive dose computations are also possible with the CC algorithm, but this algorithm cannot be employed during manual optimization. Although a full 3D image set is used in the computation of dose distributions, current clinical practice at NCCT for breast planning dictates that optimization and evaluation of treatment plans are carried out only on the slice containing the normalization point.
Definition of the planning target volume (PTV)
It is not standard practice at NCCT to define a planning target volume (PTV) for planning of the breast or chest wall. For the purpose of this study, however, a PTV was defined to enable the comparison of dose–volume histograms (DVHs) from calculations employing different calculation algorithms. The PTV was defined by the geometrical beam limits of the superior and inferior field borders and the pleura, and drawn 5 mm inside the external outline. With Helax-TMS, a structure cannot be defined once a beam geometry has been added to the study dataset. In order to define a PTV retrospectively and for the purpose of this work, each patient study set (images and structures) was exported in DICOM RT format to the Exomio (Version 2.0) virtual simulation software. PTVs were added on Exomio and the patient structure set was re-imported into Helax-TMS.
Dose calculations on Helax-TMS
All six patient studies were planned with a 6 MV beam using the standard breast planning technique. The PTV for each patient was drawn based on the criteria described previously. Dose calculations were carried out with both algorithms, both with and without inhomogeneity correction. The option of no-inhomogeneity correction simply assigns all tissue in the patient geometry to unit mass density. This was carried out in order to exclude the influence of the lung tissue on the dose calculations. The dose calculation grid was 0.5 cm.
Evaluation of plans
For the evaluation of treatment plans, isodose distributions, monitor units (to deliver 1 Gy at the normalization point) and data from DVHs were compared. To aid comparisons and relate the changes observed for the individual chest wall geometry and to current clinical practice, geometric measurements were taken from all CT studies on the central planning slice containing the normalization point. These measurements, as shown in Figure 2 were: the distance from the normalization point to the skin surface dskin-norm point, the thickness of breast, i.e. the distance from the lung-tissue interface to the skin surface dbreast, and the maximum thickness of lung tissue involved in this central slice dmax lung. For three patients, the normalization point was in breast tissue (patients 1, 2, 3) and for the others, in the lung (patients 4, 5, 6). The differences in monitor units from open and wedged beams for the PB and CC algorithms were analysed and the minimum, maximum and mean percentage doses in the PTV and lung were recorded from the DVH data.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 2. Distances measured on the central planning slice containing the isocentre.dskin-norm point is the distance from the normalization point (isocentre) to the skin of the patient, dbreast is the thickness on the chest wall at the level of the isocentre and dmax lungthe maximum lung involvement in the field.
|
|
|
Results
|
|---|
Experimental verification
The CC algorithm modelled the dose at the isocentre in a homogeneous unit density medium under tangential irradiation (Figure 1a
) with varying lateral thickness of tissue more accurately than PB. The uncertainties in the measured values did not exceed 0.5%. Deviations from measurement for CC did not exceed ±1%, whereas for PB and for lateral thickness equal to or less than 2.5 cm, these were greater than +1% (maximum of +2.3%). These results were consistent with previous findings and generally the differences between the two algorithms for this beam energy are not large [9].
Figure 3 shows the variation of prepl with field size for the NE2571 in lung at 6 MV for collimator settings greater than 5 cmx5 cm (the use of this chamber for dosimetry in smaller field sizes is not appropriate due to its size in relation to the size of the beam). The uncertainty in the derivation of prepl, as quoted by Krieger and Sauer, is 2.3% at 5 cmx5 cm and is reduced to 1.5% at 15 cmx15 cm [15]. In the case of a homogeneous lung medium irradiated with a 10 cmx10 cm field under full scatter conditions, it was found that the CC model predicted the dose to within 0.9% of the measurement, whereas the PB overestimated dose by +8.8%. The combined uncertainty in the dose measurements did not exceed 2.1%.
In Figure 4a–c, it can be seen how the measured dose varies with distance from lung tissue and phantom boundary, when the chamber is in unit density medium (experiment shown in Figure 1b). As the distance from the phantom boundary decreases, the dose decreases and it decreases even more (by 1.6%) once its proximity to the low density inhomogeneity is equal to or less than 1.5 cm. In this case, the overall deviation from measurement for CC is ±0.6%, where errors by PB vary from +2.5% to +7%.
For tangential irradiation with the normalization point in lung, the combined uncertainties in the measurement are expected to be of the order of 3% due to the assumptions in the derivation of the perturbation correction [15]. The dose predicted by the PB model varied between +10.7% to +8.2%, depending on the effective lateral thickness from the calculation point (Figure 4d). The CC predicted values between +3.6% and +2.3% for increasing effective thickness. This is in agreement with other workers who, in heterogeneous media and within low density and close to an interface, have reported the CC model to be more accurate than PB [3, 15, 19]. Here, the CC model predicted a higher dose than the measured dose, but one has to consider the uncertainty in the derivation of the perturbation correction for the measurement in lung. In the case of the heterogeneous medium with the point of measurement in lung and under complex missing tissue irradiation geometry, the CC predicts dose values close to measurement and within 3% in the case of lateral electronic equilibrium (effective lateral thickness greater than the 15 mm).
Treatment planning study
Figure 5 shows isodoses for patient 5 of this study on the CT slice containing the normalization point within lung. The isodose distribution calculated with CC follows the lung contour, predicting an overall lower dose to lung tissue and a relatively higher dose in tissue. This is because the energy released in the lung is scattered laterally to the adjacent breast and lung tissue and PB does not model this.
View larger version (42K):
[in this window]
[in a new window]
|
Figure 5. Typical dose distributions on the central planning slice generated by(a) the pencil beam (PB) and (b) the collapsed cone (CC) dose algorithms. These data are for patient 5 of this study.
|
|
Table 1 summarizes the patients included in this study and their measurements. The effective lateral distance from the normalization point to the skin is defined as the distance of tissue scaled with mass density. Table 2 summarizes the results from the two algorithms in terms of absolute difference in MU per beam, the average percentage difference in MU for the open fields only and the percentage difference in minimum, maximum and mean doses to the PTV and lung extracted from DVHs. The analogous results for the calculations without an inhomogeneity correction are shown in Table 3. Figure 6 illustrates how the percentage differences in mean doses to PTV and lung calculated by PB and CC vary with breast tissue thickness, maximum lung thickness and distance from the normalization point to the skin.
View this table:
[in this window]
[in a new window]
|
Table 2. The difference in monitor units calculated between a pencil beam(PB) and collapsed cone (CC), their average percentage differences for the open fields and the percentage difference in dose to the planning target volume (PTV) and lung, with the inhomogeneity correction switched on
|
|
View this table:
[in this window]
[in a new window]
|
Table 3. The difference in monitor units calculated between pencil beam(PB) and collapsed cone (CC) and their average percentage differences for the open fields with the inhomogeneity correction switched off
|
|
As seen from Table 2, the CC model predicted higher MU than the PB, the latter overestimating dose, and this was consistent with our experimental findings. It can be seen (Figure 6) that the greater the thickness of breast tissue the smaller the differences between the two algorithms. For patient 3 a thickness of 4.4 cm of breast tissue, resulted in, on average, 3 MU difference per beam between algorithms, a 4.1% difference to the dose at the isocentre from open fields and a 0.4% difference in the mean PTV dose. For patient 5 the thickness of breast tissue was 1.0 cm and this resulted in 5 MU difference between algorithms, thus 16.5% difference to the dose contributions at the isocentre from the open beams and 11% difference at the mean dose to the PTV. For patients with the normalization point placed in breast tissue (patients 1, 2 and 3) the differences between the two algorithms were the lowest.
Figure 6 shows that with increasing lung involvement in the plan, the mean dose to the PTV and lung increases. Patients 4 and 5 had maximum lung involvement of 2.4 cm and 2.8 cm, and for these the average percentage difference in MU was the highest, 16.5% and 16.6%, respectively. For patient 3, with similar lung involvement (2.5 cm), an analogous result was expected. But for this patient the normalization point was within 4.4 cm of breast tissue and the percentage difference in MU and the difference to the mean dose in the PTV between algorithms was the lowest observed. This would indicate that when there is sufficient breast tissue laterally, the influence of lung on the dose at the normalization point is reduced.
With the inhomogeneity correction switched off, the differences in MU between PB and CC were smaller (Table 3). For patient 5, for example, the difference in MU was 2 (3%) when the lung was not accounted for, as opposed to 9 (16.6%) when accounted for in the calculation. This confirms that it is important to account for the presence of lung in the dose calculation and when a lung heterogeneity is present, differences between algorithms increase with increasing involvement of lung.
In Table 2 it is also seen that the minimum dose difference to the PTV between the algorithms was small in some cases and in particular for patient 3. As pointed out previously, the optimization of these plans was carried out based on the distribution calculated on the central slice of the study and this is why the results from DVHs are not optimal in terms of the minimum dose to the PTV. In addition to this, these PTVs were drawn close to the patient skin and close to the penumbral region of both fields that would contribute to small dose values.
|
Discussion and conclusions
|
|---|
The dose at the normalization point in the case of isocentric breast irradiation using a 6 MV beam, with no lung involvement and in a region of lateral electronic equilibrium, is modelled to within +3% by the PB algorithm and within ±1% by CC.
For breast cancer patients who have undergone mastectomy and are also to be treated with external beam radiotherapy, the presence of lung tissue and often the small thickness of the chest wall present in the treatment volume (less than 2.5 cm), is a concern due to the known limitations of dose calculation algorithms in modelling dose in lung, and at lung-tissue interfaces, due to the missing tissue geometry and due to patient movement during treatment delivery. In terms of treatment planning, the CC algorithm calculates dose more accurately under most of these conditions (to within ±1%), with the exception of positions very close to media interfaces. The PB algorithm can underestimate the required MU by up to 7% in some cases and in all cases generates a misleading dose distribution, overestimating the dose in lung and underestimating the dose to the PTV. The differences in MU between PB and CC depend on the thickness of chest wall and the position of the normalization point with respect to the lung. Our findings indicate that for a chest wall thickness equal or less than 2.5 cm and with 2 cm or more lung tissue involvement, the CC algorithm should be used instead of PB.
For cases when the plan normalization point has to be placed in the lung and close to an interface with breast tissue, one needs to be aware of the limitations of the TPS algorithms in such regions. It is preferable to avoid placement of the plan normalization point close to interfaces, but as seen in some of the clinical cases examined here, the thin chest wall often leaves no other option. In our study, in some cases the breast tissue received higher than the prescribed dose (at least 5% higher) and the clinician should be advised to prescribe to the clinically significant isodose level covering the chest wall tissue.
This work investigated the dosimetric accuracy of isocentric chest wall irradiation using 6 MV photon beams, which is the energy used for such irradiations at our centre. A previous comparison between the PB and CC algorithms with the plan normalization point (isocentre) in homogeneous water under a tangential irradiation geometry and for a range of beam energies (4 MV, 6 MV, 15 MV) has shown that the difference from measurement is greater at the lowest energy and in particular for the PB algorithm. At higher beam energies, for a tangential irradiation with the normalization point in low density medium, it is expected that the performance of both calculation algorithms against measurement could worsen, because at higher energies the range of travel of secondary particles is greater and modelling the deposition of their energy becomes more complex, especially in the lateral direction and close to media interfaces.
Routine clinical implementation of the CC dose calculation algorithm on Helax-TMS is hindered because it is not possible to implement this interactively. However, in the special case of chest wall irradiation it is advised that an optimum treatment plan is produced using the PB algorithm and the same plan is also calculated with the CC algorithm for the clinician to make an informed decision on the prescription to the chest wall tissue. At NCCT the Helax-TMS system will soon be replaced with the Oncentra (OTP) TPS (Nucletron B.V.). Both PB and CC dose calculation engines are available on OTP, and CC could be used interactively. Therefore, CC would be the algorithm of choice for treatment planning of the chest wall, and other sites with significant involvement of non-unit density heterogeneity.
The isocentric breast planning technique at NCCT will develop further with the optimization and evaluation of plans based on distributions generated on all CT slices. New approaches have been investigated for plan optimization for the breast, such as electronic compensation using field-in-field (forward IMRT) [20, 21]. For these to be implemented clinically it is required that a PTV is defined on all slices of the CT study. This necessitates additional input by the clinician in treatment planning of the breast and/or the revision of guidelines for breast planning to be followed by experienced dosimetrists. New, complex treatment planning techniques for breast planning also necessitate a revision of the methods used for the independent check of monitor units generated by TPSs. Patient movement during treatment has not been addressed here. Future efforts in improving the treatment of chest wall patients would also need to account for chest wall movement [22, 23].
Current address for M M Aspradakis: Klinik für Radio-Onkologie, Universitäts Spital Zürich, Rämistrasse 100, Zürich 8091, Switzerland.
|
Acknowledgments
|
|---|
The authors are thankful to Dr Otto Sauer (Department of Radiotherapy, Julius-Maximilians-University of Würzburg, Germany) for discussions on the derivation of chamber perturbation correction for measurements in lung. Mr Geoff Lambert (Head of Radiotherapy Physics, Regional Medical Physics Department, Newcastle upon Tyne) and Ms Gill Lawrence (Consultant Clinical Scientist in Radiotherapy Physics, Regional Medical Physics Department) are gratefully acknowledged for their support and helpful comments to the manuscript. Dr Ujjal Mallick (Consultant Clinical Oncologist, Northern Centre for Cancer Treatment) is acknowledged for his clinical input to this work.
Received for publication October 21, 2005.
Revision received February 14, 2006.
Accepted for publication February 20, 2006.
|
References
|
|---|
- Swerdlow A, Dos Santos Silva I, Doll R, editors. Cancer incidence and mortality in England and Wales: trends and risk factors. Oxford University Press, 2001:71–196
- Johnson S. Incidence and mortality in breast cancer. In: International handbook of breast cancer. Euromed Publications, 2002
- Ahnesjö A. Collapsed cone convolution of radiant energy for photon dose calculation in heterogeneous media. Med Phys 1989;16:577–92.[CrossRef][Medline]
- Ahnesjö A, Saxner M, Trepp A. A pencil beam model for photon dose calculation. Med Phys 1992;19:263–73.[CrossRef][Medline]
- Ahnesjö A, Knöös T, Montelius A. Application of the convolution method for calculation of output factors for therapy photon beams. Med Phys 1992;19:295–301.[CrossRef][Medline]
- Knöös T, Ceberg C, Weber L, Nilsson P. The dosimetric verification of a pencil beam based treatment planning system. Phys Med Biol 1994;39:1609–28.[CrossRef][Medline]
- Engelsman M, Damen EMF, Kiken PW, van't Veld A, Ingen KM, Mijnheer BJ. Impact of simple tissue inhomogeneity correction algorithms on conformal radiotherapy of lung tumours. Radiother Oncol 2001;60:299–309.[CrossRef][Medline]
- Weber L, Nilsson P. Verification of dose calculations with a clinical treatment planning system based on a point kernel dose engine. J Applied Clin Med Phys 2002;3:73–87.
- Aspradakis MM, Morrison RH, Richmond ND, Steele A. Experimental verification of convolution/superposition photon dose calculations for radiotherapy treatment planning. Phys Med Biol 2003;48:2873–93.[CrossRef][Medline]
- Wieslander E, Knöös T. Dose pertubation in the presence of metallic implants: treatment planning system versus Monte Carlo simulations. Phys Med Biol 2003;48:3295–305.[CrossRef][Medline]
- Irvine C, Morgan A, Crellin A, Nisbet A, Beange I. The clinical implications of the collapsed cone planning algorithm. Clin Oncol 2004;16:148–54.[CrossRef]
- Constantinou C, Attix FH, Bhudatt RP. A solid water phantom material for radiotherapy x-ray and gamma-ray beam calibrations. Med Phys 1982;9:436–41.[CrossRef][Medline]
- White DR, Martin RJ, Darlison R. Epoxy resin based tissue substitutes. Br J Radiol 1977;50:814–21.[Abstract]
- White DR, Constantinou C, Martin RJ. Technical Note: Foamed epoxy resin-based lung substitutes. Br J Radiol 1986;59:787–90.[Medline]
- Krieger T, Sauer O. Monte Carlo- versus pencil beam-/collapsed -cone-dose calculation in a heterogeneous multi-layer medium. Phys Med Biol 2005;50:859–68.[CrossRef][Medline]
- Siddon RL. Solution to treatment planning problems using coordinate transformations. Med Phys 1981;8:766–74.[CrossRef][Medline]
- Casebow MP. Matching of adjacent radiation beams for isocentric radiotherapy. Br J Radiol 1984;57:735–40.[Medline]
- Wilkinson J. Consensus management - a departmental study of the external beam therapy of carcinoma breast. Radiography Today 1992:17–20
- Carrasco P, Jornet N, Duch MA, Weber L, Ginjaume M, Eudaldo T, et al. Comparison of dose calculation algorithms in phantoms with lung equivalent heterogeneities under conditions of lateral electronic disequilibrium. Med Phys 2004;31:2899–911.[CrossRef][Medline]
- Richmond ND, Turner R, Dawes PDK, Lambert GD, Lawrence GL. Evaluation of the dosimetric consequences of adding a single asymmetric or MLC shaped field to a tangential breast radiotherapy technique. Radiother Oncol 2003;67:165–70.[CrossRef][Medline]
- Wilkinson M, Aspradakis MM. A study of electronic compensation methods in breast radiotherapy planning. In: IPEM Annual Scientific Meeting. York, UK: IPEM, 2004
- Pedersen AN, Korreman S, Nystrom H, Specht L. Breathing adapted radiotherapy of breast cancer: reduction of cardiac and pulmonary doses using voluntary inspiration breath-hold. Radiother Oncol 2004;72:53–60.[CrossRef][Medline]
- Korreman SS, Pedersen AN, Nottrup TJ, Specht L, Nystrom H. Breathing adapted radiotherapy for breast cancer: comparison of free breathing gating with the breath-hold technique. Radiother Oncol 2005;76:311–8.[CrossRef][Medline]
Top
Abstract
Introduction
, PTW31003 with PMMA wall in styrofoam (Krieger & Sauer, with permission); 
, NE2571 with graphite wall in styrofoam; •, NE2571 with graphite in wall in lung.
, collapsed cone; •, measurement.