- Email: [email protected]
21 Quantitative and qualitative evaluation of different techniques in breast irradiation
$20
Monday, September 26
requirement to use 3d dose compensation to ensure the dosimetry achieved ICRU 50/62 guidelines. The control which dose modulation allows over the dose distribution enables doses to be modulated within a fraction of treatment rather than changing fraction size and number. Different doses may then be delivered to different regions of the breast depending on the recurrence risk. However, this type of trial requires much better techniques for Iocalisation and verification of the tumour bed and more knowledge of possible changes throughout treatment. It is clear that modulated dose distributions may now be delivered to the breast, but much work is needed to ensure that these coincide with the target areas by exploring known, and new imaging, technologies and applications. 19 Few S e g m e n t s Forward I M R T f o r Breast I r r a d i a t i o n
J.B. Van de Kamer I, C.C. W~rl~m 2, G. Van Tienhoven I, A. Bel ~, C.P. Raaijmakers 2 I Department of Radiation Oncology, Academic Medical Center, Amsterdam, The Netherlands 2Department of Radiotherapy, University Medical Center, Utrecht, The Netherlands Purpose: to develop a Quick IMRT (Q-IMRT) method to achieve a homogeneous 3D dose distribution in breast irradiation. Methods: From patients with breast cancer CT datasets are acquired in treatment position for 3D treatment planning. The target volume (breast tissue) is outlined by a radiation oncologist. In the treatment planning system (Plato RTS v2.6) two tangential beams are set-up in the conventional way, using MLCs to conform to the target volume. From this beam geometry and the CT dataset, medial and lateral equivalent path length maps (EPM) are constructed. Regions with a large EP need more monitor units than regions with a small EP for optimal dose homogeneity. The intensity maps are downscaled to a resolution of 10x2 mm2 to account for leaf width and step size (step and shoot technique). The downscaled maps are segmented into a chosen number of intensity levels (typically 3-4). These segments are sequenced with a home-made tool to ascertain that the number of segments equals the number of intensity levels. Fixing this number is desired since the segments are manually weighted to obtain a homogeneous dose distribution. After optimisation, the multiple-beam plan is converted into a 2-beam IMRT plan for easy delivery on a linear accelerator (Elekta). Finally the dose distribution of the 2-beam IMRT plan is evaluated. The approved plan is transferred to the treatment unit for delivery. Results: Using 4 instead of 3 segments per beam improved the dose distributions considerably. Using more segments would make the procedure less practical; automatic weighting of the beams, which is currently under investigation, would be desirable. Starting with the tangential beams, the Q-IMRT procedure takes about 30 minutes to provide an optimized dose distribution. The Q-IMRT plans showed a slight reduction in both heart and lung dose compared to the conventional plans. An additional benefit of the use of Q-IMRT for breast irradiation is the reduction of the required number of MU (about 20%). This results in a lower overall dose for the non-target regions. The developed tools are now standard available at our planning departments. So far, 50 patients have been treated with this IMRT technique in the UMC, Utrecht. In the AMC, Amsterdam the technique is currently being examined for possible future use. Conclusion: The developed method of Quick IMRT improves the dose distribution and is feasible for daily clinical practice.
Symposia 2O The influence of h e t e o g e n e o u s dose distributions on the risk of radiation-induced cancer in the c o n t r a l a t e r a l breast
S. JohanseG E. Malinen Institute for Cancer Research, hospital, Oslo, Norway
The Norwegian
Radium
In risk estimates of radiation-induced cancers, the effect of heterogeneous dose distributions in organs at risk is rarely addressed. In such organs, the distribution of potentially malignant cells, with respect to carcinogenesis, is not necessarily homogeneous. For heterogeneous distributions of dose and potentially malignant cells, the risk is not directly proportional to the average dose in the tissue. The purpose of this study was to investigate the possible influence of the heterogeneous dose distribution in the contralateral breast following radiotherapy of the breast and regional lymph nodes on the risk of secondary cancer. 8 patients undergoing radiotherapy of the breast and regional lymph nodes were selected. Using the TMS (v. 6.1 a) treatment planning system, the contralateral breast volume was outlined in the CT slices and divided into 4 quadrants. The dose in the entire breast and in each quadrant was calculated by the TMS collapsed cone algorithm. The relative distribution of potentially malignant cells in the breast was estimated from published epidemiological data of a population-averaged distribution of "spontaneous" tumors in the breast. It is assumed that the cellular origin of "spontaneous" and radiation-induced cancers is identical. From these data, the relative number of relevant cells in the upper inner, upper outer, lower inner and lower outer quadrant of the breast is roughly 0.2, 0.6, 0.1 and 0.1, respectively. Thus, in the current framework, the upper outer quadrant is considered to be the most important part of the breast with respect to radiation-induced cancers. For risk estimates of secondary cancer in the contralateral breast, a linear response model, incorporating the dose and relative number of potentially malignant cells in each of the 4 quadrants, was utilized. The estimates were compared with a standard risk model, employing the average dose to the entire contralateral breast. The average dose to the entire contralateral breast was 4.6 % of the prescribed dose. The average dose to the upper inner, upper outer, lower inner and lower outer quadrant of the contralateral breast was 11.4, 3.3, 2.2 and 1.2 %, respectively. The excess risk estimated by taking the heterogeneous dose distribution into account was only about 3 % lower than the risk predicted by the "standard risk model. Therefore, in the current case, the more advanced risk estimates was not significantly different from standard estimates. However, other dose distributions may change these conclusions. 21 Q u a n t i t a t i v e and q u a l i t a t i v e e v a l u a t i o n t e c h n i q u e s in breast irradiation
of d i f f e r e n t
M. Rosa I, K. Jacob 1, A. Soares 1, F. Serra I, M. Ramalho I, S. O/iveira 1, R. Rodrigues 2, P. Ferreira I, N. Teixeira 3 1Medical Consult/HCD, Radiotherapy, Lisbon, Portugal 2Hosp. Cu# Descobertas, Radiotherapy, Lisbon, Portugal 3Medical Consult/HCD / ESTeS Lisboa, Radiotherapy, Lisbon, Portugal The side effects of breast irradiation in lungs and heart were always a serious problem in breast irradiation. Recent techniques were developed to improve dose homogeneity in the planning target volume (PTV) and to reduce the levels of toxicity in the vicinity tissues and in the organs at risk (OAR). The purpose of this study is to evaluate four different irradiation techniques using dose/volume histograms and isodose distributions in the Xio treatment planning system, for breast irradiation aith an Elekta accelerator.
Monday, September 26
Symposia The most conventional techniques analyzed consist of two isocentric tangential opposed fields, with and without wedge, applying the matching penumbras method. However, areas of under and over dose are often generated. Recent techniques try to solve some of these problems, like the two-segment techniques, which congregate the two conventional tangential fields with two small fields (same weight). The limitation of the wedge concerning the position of the collimator of the Elekta accelerator that exists at CUF Descobertas Hospital (Lisbon), does not allow an efficient protection of OAR if the conventional two field technique with wedge is used. However it's possible to overcome this problem with the four equivalent fields technique: two with wedges and two without wedges (the first ones provide a uniform irradiation of the PTV while, the last ones supplies adequate protections). The evaluation of this study was based in a sample of 82 clinical cases. The results obtained show that the new techniques improve the dose distribution in the areas of high gradient in breast, providing at the same time a better protection of OAR. 22 Clinical e x p e r i e n c e w i t h a setup correction protocol for breast cancer patients R. Kollaardj C. Hurkmans, R. Janssen Catharina Hospital, Radiotherapy Department, Eindhoven, The Netherlands Background and purpose: Offline setup correction can be used to reduce the systematic error in patient setup. The purpose of this presentation is to describe the clinical application of a setup correction protocol for breast cancer patients and to compare different protocols. Materials a n d m e t h o d s : Electronic portal images were made of the medio-lateral tangential treatment field and were matched with reference images from the CT. A shrinking action level (SAL) protocol (eL = 10 ram, Nmax = 3) was applied for a group of 97 patients. The inter-observer variability in matching was investigated for 4 observers and 70 portal images. The setup errors with and without the use of a correction protocol were obtained from the clinical data. Different protocols were simulated using our clinical data. Results: The setup errors are reported in 2D. The average inter-observer variation was 1.5 mm (1 SD) for the matches of the portal images. Due to the protocol, the mean population error reduced from 1.3 to 0.6 mm, the systematic error reduced from 3.8 to 2.3 mm (1 SD) and the random error increased from 3.3 to 3.8 mm (1 SD). 10% of the patients had an initial systematic setup error of 5.9 mm or more (5.9-13.2 mm). After application of the protocol, 0% of the patients had such a large systematic setup error. 4 out of 7 patients with a setup error over 7 mm had a large breast volume (at least 1264 cc). Simulations show that the setup correction can be improved with a no-action level (NAL) protocol (using 3 fractions) or with another choice of SAL parameters (e.g. cz = 7 mm, N~na× = 2). The average number of portal images would be lower for the NAL protocol (3) and comparable for the SAL protocol (6.9). The systematic setup error would be reduced to 1.7 mm (1 SD) for the SAL protocol and 1.3 mm (1 SD) for the NAL protocol. The frequency of setup correction is 57% for this SAL protocol and 100% for the NAL protocol. Currently, only 25% of the patients receive a setup correction. Conclusions: Large setup errors can be completely avoided with the current SAL protocol. Suggestions for further improvement of the patient setup were found using simulations based on the clinical data. With a setup correction protocol, more accurate irradiation of the target volume is possible and more advanced treatment techniques can be applied.
100% 90%
S21
: - ....... ithoutc0n'ecSonprotocol ith SAL protocol(N=3,alpha=10)
80%
70% 60% 50% 40%
"~ 30% z 20% 10% 0% 0
2
4
6
8
10
12
2D systematicsetup error (ram)
14
16
Monday, September 26
requirement to use 3d dose compensation to ensure the dosimetry achieved ICRU 50/62 guidelines. The control which dose modulation allows over the dose distribution enables doses to be modulated within a fraction of treatment rather than changing fraction size and number. Different doses may then be delivered to different regions of the breast depending on the recurrence risk. However, this type of trial requires much better techniques for Iocalisation and verification of the tumour bed and more knowledge of possible changes throughout treatment. It is clear that modulated dose distributions may now be delivered to the breast, but much work is needed to ensure that these coincide with the target areas by exploring known, and new imaging, technologies and applications. 19 Few S e g m e n t s Forward I M R T f o r Breast I r r a d i a t i o n
J.B. Van de Kamer I, C.C. W~rl~m 2, G. Van Tienhoven I, A. Bel ~, C.P. Raaijmakers 2 I Department of Radiation Oncology, Academic Medical Center, Amsterdam, The Netherlands 2Department of Radiotherapy, University Medical Center, Utrecht, The Netherlands Purpose: to develop a Quick IMRT (Q-IMRT) method to achieve a homogeneous 3D dose distribution in breast irradiation. Methods: From patients with breast cancer CT datasets are acquired in treatment position for 3D treatment planning. The target volume (breast tissue) is outlined by a radiation oncologist. In the treatment planning system (Plato RTS v2.6) two tangential beams are set-up in the conventional way, using MLCs to conform to the target volume. From this beam geometry and the CT dataset, medial and lateral equivalent path length maps (EPM) are constructed. Regions with a large EP need more monitor units than regions with a small EP for optimal dose homogeneity. The intensity maps are downscaled to a resolution of 10x2 mm2 to account for leaf width and step size (step and shoot technique). The downscaled maps are segmented into a chosen number of intensity levels (typically 3-4). These segments are sequenced with a home-made tool to ascertain that the number of segments equals the number of intensity levels. Fixing this number is desired since the segments are manually weighted to obtain a homogeneous dose distribution. After optimisation, the multiple-beam plan is converted into a 2-beam IMRT plan for easy delivery on a linear accelerator (Elekta). Finally the dose distribution of the 2-beam IMRT plan is evaluated. The approved plan is transferred to the treatment unit for delivery. Results: Using 4 instead of 3 segments per beam improved the dose distributions considerably. Using more segments would make the procedure less practical; automatic weighting of the beams, which is currently under investigation, would be desirable. Starting with the tangential beams, the Q-IMRT procedure takes about 30 minutes to provide an optimized dose distribution. The Q-IMRT plans showed a slight reduction in both heart and lung dose compared to the conventional plans. An additional benefit of the use of Q-IMRT for breast irradiation is the reduction of the required number of MU (about 20%). This results in a lower overall dose for the non-target regions. The developed tools are now standard available at our planning departments. So far, 50 patients have been treated with this IMRT technique in the UMC, Utrecht. In the AMC, Amsterdam the technique is currently being examined for possible future use. Conclusion: The developed method of Quick IMRT improves the dose distribution and is feasible for daily clinical practice.
Symposia 2O The influence of h e t e o g e n e o u s dose distributions on the risk of radiation-induced cancer in the c o n t r a l a t e r a l breast
S. JohanseG E. Malinen Institute for Cancer Research, hospital, Oslo, Norway
The Norwegian
Radium
In risk estimates of radiation-induced cancers, the effect of heterogeneous dose distributions in organs at risk is rarely addressed. In such organs, the distribution of potentially malignant cells, with respect to carcinogenesis, is not necessarily homogeneous. For heterogeneous distributions of dose and potentially malignant cells, the risk is not directly proportional to the average dose in the tissue. The purpose of this study was to investigate the possible influence of the heterogeneous dose distribution in the contralateral breast following radiotherapy of the breast and regional lymph nodes on the risk of secondary cancer. 8 patients undergoing radiotherapy of the breast and regional lymph nodes were selected. Using the TMS (v. 6.1 a) treatment planning system, the contralateral breast volume was outlined in the CT slices and divided into 4 quadrants. The dose in the entire breast and in each quadrant was calculated by the TMS collapsed cone algorithm. The relative distribution of potentially malignant cells in the breast was estimated from published epidemiological data of a population-averaged distribution of "spontaneous" tumors in the breast. It is assumed that the cellular origin of "spontaneous" and radiation-induced cancers is identical. From these data, the relative number of relevant cells in the upper inner, upper outer, lower inner and lower outer quadrant of the breast is roughly 0.2, 0.6, 0.1 and 0.1, respectively. Thus, in the current framework, the upper outer quadrant is considered to be the most important part of the breast with respect to radiation-induced cancers. For risk estimates of secondary cancer in the contralateral breast, a linear response model, incorporating the dose and relative number of potentially malignant cells in each of the 4 quadrants, was utilized. The estimates were compared with a standard risk model, employing the average dose to the entire contralateral breast. The average dose to the entire contralateral breast was 4.6 % of the prescribed dose. The average dose to the upper inner, upper outer, lower inner and lower outer quadrant of the contralateral breast was 11.4, 3.3, 2.2 and 1.2 %, respectively. The excess risk estimated by taking the heterogeneous dose distribution into account was only about 3 % lower than the risk predicted by the "standard risk model. Therefore, in the current case, the more advanced risk estimates was not significantly different from standard estimates. However, other dose distributions may change these conclusions. 21 Q u a n t i t a t i v e and q u a l i t a t i v e e v a l u a t i o n t e c h n i q u e s in breast irradiation
of d i f f e r e n t
M. Rosa I, K. Jacob 1, A. Soares 1, F. Serra I, M. Ramalho I, S. O/iveira 1, R. Rodrigues 2, P. Ferreira I, N. Teixeira 3 1Medical Consult/HCD, Radiotherapy, Lisbon, Portugal 2Hosp. Cu# Descobertas, Radiotherapy, Lisbon, Portugal 3Medical Consult/HCD / ESTeS Lisboa, Radiotherapy, Lisbon, Portugal The side effects of breast irradiation in lungs and heart were always a serious problem in breast irradiation. Recent techniques were developed to improve dose homogeneity in the planning target volume (PTV) and to reduce the levels of toxicity in the vicinity tissues and in the organs at risk (OAR). The purpose of this study is to evaluate four different irradiation techniques using dose/volume histograms and isodose distributions in the Xio treatment planning system, for breast irradiation aith an Elekta accelerator.
Monday, September 26
Symposia The most conventional techniques analyzed consist of two isocentric tangential opposed fields, with and without wedge, applying the matching penumbras method. However, areas of under and over dose are often generated. Recent techniques try to solve some of these problems, like the two-segment techniques, which congregate the two conventional tangential fields with two small fields (same weight). The limitation of the wedge concerning the position of the collimator of the Elekta accelerator that exists at CUF Descobertas Hospital (Lisbon), does not allow an efficient protection of OAR if the conventional two field technique with wedge is used. However it's possible to overcome this problem with the four equivalent fields technique: two with wedges and two without wedges (the first ones provide a uniform irradiation of the PTV while, the last ones supplies adequate protections). The evaluation of this study was based in a sample of 82 clinical cases. The results obtained show that the new techniques improve the dose distribution in the areas of high gradient in breast, providing at the same time a better protection of OAR. 22 Clinical e x p e r i e n c e w i t h a setup correction protocol for breast cancer patients R. Kollaardj C. Hurkmans, R. Janssen Catharina Hospital, Radiotherapy Department, Eindhoven, The Netherlands Background and purpose: Offline setup correction can be used to reduce the systematic error in patient setup. The purpose of this presentation is to describe the clinical application of a setup correction protocol for breast cancer patients and to compare different protocols. Materials a n d m e t h o d s : Electronic portal images were made of the medio-lateral tangential treatment field and were matched with reference images from the CT. A shrinking action level (SAL) protocol (eL = 10 ram, Nmax = 3) was applied for a group of 97 patients. The inter-observer variability in matching was investigated for 4 observers and 70 portal images. The setup errors with and without the use of a correction protocol were obtained from the clinical data. Different protocols were simulated using our clinical data. Results: The setup errors are reported in 2D. The average inter-observer variation was 1.5 mm (1 SD) for the matches of the portal images. Due to the protocol, the mean population error reduced from 1.3 to 0.6 mm, the systematic error reduced from 3.8 to 2.3 mm (1 SD) and the random error increased from 3.3 to 3.8 mm (1 SD). 10% of the patients had an initial systematic setup error of 5.9 mm or more (5.9-13.2 mm). After application of the protocol, 0% of the patients had such a large systematic setup error. 4 out of 7 patients with a setup error over 7 mm had a large breast volume (at least 1264 cc). Simulations show that the setup correction can be improved with a no-action level (NAL) protocol (using 3 fractions) or with another choice of SAL parameters (e.g. cz = 7 mm, N~na× = 2). The average number of portal images would be lower for the NAL protocol (3) and comparable for the SAL protocol (6.9). The systematic setup error would be reduced to 1.7 mm (1 SD) for the SAL protocol and 1.3 mm (1 SD) for the NAL protocol. The frequency of setup correction is 57% for this SAL protocol and 100% for the NAL protocol. Currently, only 25% of the patients receive a setup correction. Conclusions: Large setup errors can be completely avoided with the current SAL protocol. Suggestions for further improvement of the patient setup were found using simulations based on the clinical data. With a setup correction protocol, more accurate irradiation of the target volume is possible and more advanced treatment techniques can be applied.
100% 90%
S21
: - ....... ithoutc0n'ecSonprotocol ith SAL protocol(N=3,alpha=10)
80%
70% 60% 50% 40%
"~ 30% z 20% 10% 0% 0
2
4
6
8
10
12
2D systematicsetup error (ram)
14
16