- Email: [email protected]
Aperture-based IMRT inverse planning with incorporation of organ motion
Proceedings of the 46th Annual ASTRO Meeting
of IMRT for small lung lesions should be used cautiously as the advantage of the intensity modulated distribution could be compromised by respiratory motion.
2484
The Effects of Lateral Electronic Disequilibrium on Small IMRT Beamlets
A. O. Jones Dept of Radiation Oncology, Geisinger Medical Center, Danville, PA Purpose/Objective: IMRT has become increasingly prevalent in the past several years. Inherent in the delivery of IMRT is the use of small individual beams (beamlets) to create the desired dose distribution where beams of less than 1 cm2 are commonly employed. There is a trend toward using smaller beamlets in order to increase the resolution of the intensity map. Similarly sized beamlets may be employed in functional stereotactic radiosurgery (SRS). These small beams do not have lateral electronic equilibrium even in homogeneous conditions. When they impinge upon different densities found in the body large changes in central axis depth doses, relative to the homogeneous case, can occur. Clinically, IMRT can be used to effectively treat areas in the head and neck. In these regions there are a myriad of different densities and interfaces. Soft tissue, bone, and air may all occur along the pathlength of an individual beamlet. Similarly in the chest, soft tissue, bone, and lung may all be encountered by individual beamlets. It has been shown that large dose drops using small beams are seen in lung tissue due to an increase in the range of electrons and the resulting exacerbation of the lack of lateral electronic equilibrium.1 In bone, an increase in dose is expected as the electron range is decreased relative to water and electrons that would normally deposit dose outside of the field remain within the field. Materials/Methods: Monte Carlo simulations (EGSnrcMP, National Research Council, Canada) are used to calculate 0.1, 0.2, 0.3, 0.5, 1.0, and 3.0 cm diameter 6 MV photon beams in a phantom consisting of 3 cm water, 3 cm bone (density 1.85 g/cm3 ), and 6 cm water. Percent depth dose curves are generated to show the effects of the different densities on the central axis dose. The magnitude of the dose change is examined using the Dose Perturbation Factor (DPF), which is the ratio of the dose at a point in the heterogeneous phantom to the dose to that same point in the homogeneous phantom. Monte Carlo simulations are also conducted to examine the combined effects of bone and lung tissue. Beamlets of 0.1, 0.3, 0.5, 1.0, and 3.0 cm are simulated on a phantom consisting of 3 cm water, 1 cm bone, 10 cm lung, 2 cm bone, and 14 cm water. Results: Percent depth dose curves show an increase in central axis dose in the bone relative to water. The increase in dose is inversely related to the field size ranging from 28% for a 0.1 cm field to less than 5% for a 3 cm field. The maximum DPF occurs at the initial soft tissue to bone interface. The average DPF across the bone is 1.21 for the 0.1 cm field. The 3 cm field size has an average DPF of 0.995, which is less than the statistical variation of the simulations (1%). The DPF decreases with depth and with increased field size. Beyond the bone the dose is decreased for the smallest field sizes and the average DPF distal to the bone drops to 0.87 for the 0.1 cm field size. The 1.0 cm field size shows no dose change beyond the bone with an average DPF of 0.995, and the 3.0 cm field shows a dose enhancement beyond the bone with an average DPF of 1.05. Field size is inversely related to the magnitude of the DPF for both bone and lung tissue with large changes evident for field sizes up to 3.0 cm. Comparison to the collapsed cone convolution algorithm shows that the magnitude of the dose change in bone and in lung is not correctly calculated, although the convolution algorithm does predict a dose change in the correct direction. Conclusions: This work shows that bone has a large effect on the central axis dose of small photon beams commonly used in IMRT and SRS. The dose to the bone is substantially increased while the dose beyond the bone is decreased for field sizes up to 3.0 cm. Combining this effect with additional significant density changes (either air or lung) substantially changes the dose to tissues along the beamlet path. These effects may have clinical implications in the delivery of ultra-small fields in IMRT or SRS. 1 Jones AO, Das IJ, and Jones FL, A Monte Carlo Study of IMRT Beamlets in Inhomogeneous Media, Med. Phys. 30: 296 – 300 2003.
2485
Aperture-Based IMRT Inverse Planning with Incorporation of Organ Motion
Z. Shou, L. Xing Radiation Oncology, Stanford University School of Medicine, Stanford, CA Purpose/Objective: Accurate targeting is of paramount importance in intensity modulated radiation therapy (IMRT). In the presence of organ motion, the problem becomes complicated because of the interplay between the organ motion and MLC movement. The purpose of this work is to develop a general aperture-based inverse planning algorithm with consideration of organ motion to compensate or minimize the influence of the interplay between the two types of movements. Materials/Methods: We implemented a hybrid aperture-based inverse planning algorithm that is capable of producing optimal shapes and weights of a pre-selected number of segments for each incident field. In this algorithm the variables defining the shapes of the involved segments and the weights of the segments were dealt separately by using a fast simulated annealing (ASA) and a descent gradient optimization. Two different scenarios were investigated: (i) the shapes of the organs (target and/or sensitive structures) remain unchanged during the organ motion process; and (ii) the shapes of the organs are deformable. A pre-requisition of the proposed aperture-based inverse planning algorithm is that the temporal process of each point within a structure has been acquired either before the treatment or during the treatment through the use of an adaptive algorithm and an online imaging system. In either case, the overall goal of the inverse planning is to minimize the temporally integrated dose to the sensitive structures while delivering a uniform integrated dose distribution to the target volume. The algorithm was first examined by using a phantom case with sinusoidally moving target and/or sensitive structures. The proposed technique was then used to plan four breast cases and dose distributions were computed for three different situations: (a) conventional aperture-
S631
S632
I. J. Radiation Oncology
● Biology ● Physics
Volume 60, Number 1, Supplement, 2004
based IMRT treatment without considering the breathing motion (other than the use of appropriate margins); (b) new aperture-based IMRT treatment with consideration of breathing motion; and (c) a hypothetical treatment where the segmented beams derived from (i) was used to irradiate the beast target volumes with breathing motion. Two opposed tangential fields (the beam energy was either 6 MV or 15 MV), each consists of 12 to 25 segments, were used for these plans. The three types of plans were compared quantitatively. Results: In both phantom and clinical case studies, we found that the organ motion severely deteriorated the dose distributions obtained using conventional aperture-based IMRT inverse planning with static organs even when appropriate margins were used. Generally, the level of deterioration increases with the level of intensity modulation or the number of segments of the incident beams. With the new algorithm, it was possible to significantly reduce the adverse dosimetric effect of the organ motion. For the same number of segments, the quality of the plan with organ motion was generally inferior to that without organ motion. To achieve similar target dose coverage and sensitive structure sparing as the static case, 15%– 40% more segments were generally required when the organ motion was switched on. Further increase of the segment number would lead to better but clinically insignificant improvement. The demand for more segments tends to be less as the original number of segments for the static plan is large. The comparison results strongly depended on the amplitude of organ motion. It was also found that the deformation of the involved organs played less significant role provided that the level of deformation was within its practical range and provided that the deformation was accounted for on a case specific basis by using the inverse planning technique proposed here. Conclusions: An inverse planning system with incorporation of organ motion was developed for aperture-based IMRT and the influence of organ motion on the final dose was investigated systematically. It was shown that the adverse dosimetric effect of organ motion can be compensated or minimized by the modified inverse planning algorithm with moderate increase of segmented fields. Together with organ tracking/imaging techniques, the approach may provide a practical solution to the clinically intractable problem of organ motion.
2486
IMRT Simultaneous Integrated Boost for High-Risk Prostate Cancer
X. Li,1,2 J. Z. Wang,2 P. Jursinic,1 C. A. Lawton1 1
Radiation Oncology, Medical College of Wisconsin, Milwaukee, WI, 2Radiation Oncology, University of Maryland, Baltimore, MD Purpose/Objective: A sequential two-phase process, initial and boost irradiation, is commonly used for high risk prostate cancer. In this work, we propose to use IMRT simultaneous integrated boost (SIB) to replace this two-phase process. The SIB which is a single-phase process can simultaneously deliver high dose to prostate and lower dose to pelvic nodes. Materials/Methods: The SIB is designed to irradiate pelvic nodes with the same fractionations as conventional two-phase treatment (e.g., 25 ⫻ 1.8 Gy) and to deliver higher doses to prostate in the same 25 fractions (i.e., hypofractionation). The concept of equivalent uniform dose (EUD) was used to determine suitable SIB fractionations that deliver the biologically equivalent doses to prostate. The EUD for prostate is estimated based on LQ model. The most recent LQ parameters derived from clinical data for prostate cancer were used. The EUD for normal tissue was computed based on the Lyman model. To compare biological effectiveness voxel by voxel, we propose a new concept, termed the voxel equivalent dose (VED). The calculation of VED is similar to that for EUD except it is done within a voxel. The 3D VED distributions and the VED-based DVHs were used to compare biological and dosimetric merits for newly-designed SIB regimens. To demonstrate dosimetric feasibility and advantages of the proposed SIB, we have performed a planning study on selected patient CT using a commercial IMRT and a 3D planning systems. Four treatment scenarios were considered: (1) 3D initial plus 3D boost, (2) 3D initial plus IMRT boost, (3) IMRT initial plus IMRT boost, and (4) IMRT SIB. EUDs and VED-based DVHs for prostate, pelvic nodes, small bowel, rectum, bladder and other tissue for all 4 scenarios are compared to show dosimetric and biological benefits of the IMRT SIB. Results: A series of equivalent hypofractionation regimens suitable for the IMRT SIB are obtained. The table(table 1) below lists the required SIB prescribed doses that are biologically equivalent to a series of desired EUDs. The data are calculated using LQ parameters of ␣ ⫽ 0.15 Gy-1, ␣/ ⫽ 3.1 Gy. The data in the table may be used to design appropriate SIB schemes that are biologically equivalent to the conventional treatments. For example, the conventional regimen of 42 ⫻ 1.8 Gy (EUD ⫽ 75.4 Gy) would be equivalent to a SIB regimen of 25 ⫻ 2.54 Gy. From comparisons of 3D VED distributions and DVHs between the SIB and the conventional two-phase treatments, we found that the SIB offers more conformal dose distributions to prostate and pelvic nodes and provides better sparing to the critical structures. The figure below compares VED-based DVHs for rectum for the 4 scenarios with an EUD of 75.4 Gy. Conclusions: A new IMRT simultaneous integrated boost strategy that irradiates prostate with hypofractionations while irradiating pelvic nodes with the conventional fractionations is proposed for high-risk prostate cancer. Compared with the conventional two-phase treatment, the proposed SIB technique offers potential advantages including better sparing of critical structures, more efficient delivery, shorter treatment duration, and higher biological effectiveness.
of IMRT for small lung lesions should be used cautiously as the advantage of the intensity modulated distribution could be compromised by respiratory motion.
2484
The Effects of Lateral Electronic Disequilibrium on Small IMRT Beamlets
A. O. Jones Dept of Radiation Oncology, Geisinger Medical Center, Danville, PA Purpose/Objective: IMRT has become increasingly prevalent in the past several years. Inherent in the delivery of IMRT is the use of small individual beams (beamlets) to create the desired dose distribution where beams of less than 1 cm2 are commonly employed. There is a trend toward using smaller beamlets in order to increase the resolution of the intensity map. Similarly sized beamlets may be employed in functional stereotactic radiosurgery (SRS). These small beams do not have lateral electronic equilibrium even in homogeneous conditions. When they impinge upon different densities found in the body large changes in central axis depth doses, relative to the homogeneous case, can occur. Clinically, IMRT can be used to effectively treat areas in the head and neck. In these regions there are a myriad of different densities and interfaces. Soft tissue, bone, and air may all occur along the pathlength of an individual beamlet. Similarly in the chest, soft tissue, bone, and lung may all be encountered by individual beamlets. It has been shown that large dose drops using small beams are seen in lung tissue due to an increase in the range of electrons and the resulting exacerbation of the lack of lateral electronic equilibrium.1 In bone, an increase in dose is expected as the electron range is decreased relative to water and electrons that would normally deposit dose outside of the field remain within the field. Materials/Methods: Monte Carlo simulations (EGSnrcMP, National Research Council, Canada) are used to calculate 0.1, 0.2, 0.3, 0.5, 1.0, and 3.0 cm diameter 6 MV photon beams in a phantom consisting of 3 cm water, 3 cm bone (density 1.85 g/cm3 ), and 6 cm water. Percent depth dose curves are generated to show the effects of the different densities on the central axis dose. The magnitude of the dose change is examined using the Dose Perturbation Factor (DPF), which is the ratio of the dose at a point in the heterogeneous phantom to the dose to that same point in the homogeneous phantom. Monte Carlo simulations are also conducted to examine the combined effects of bone and lung tissue. Beamlets of 0.1, 0.3, 0.5, 1.0, and 3.0 cm are simulated on a phantom consisting of 3 cm water, 1 cm bone, 10 cm lung, 2 cm bone, and 14 cm water. Results: Percent depth dose curves show an increase in central axis dose in the bone relative to water. The increase in dose is inversely related to the field size ranging from 28% for a 0.1 cm field to less than 5% for a 3 cm field. The maximum DPF occurs at the initial soft tissue to bone interface. The average DPF across the bone is 1.21 for the 0.1 cm field. The 3 cm field size has an average DPF of 0.995, which is less than the statistical variation of the simulations (1%). The DPF decreases with depth and with increased field size. Beyond the bone the dose is decreased for the smallest field sizes and the average DPF distal to the bone drops to 0.87 for the 0.1 cm field size. The 1.0 cm field size shows no dose change beyond the bone with an average DPF of 0.995, and the 3.0 cm field shows a dose enhancement beyond the bone with an average DPF of 1.05. Field size is inversely related to the magnitude of the DPF for both bone and lung tissue with large changes evident for field sizes up to 3.0 cm. Comparison to the collapsed cone convolution algorithm shows that the magnitude of the dose change in bone and in lung is not correctly calculated, although the convolution algorithm does predict a dose change in the correct direction. Conclusions: This work shows that bone has a large effect on the central axis dose of small photon beams commonly used in IMRT and SRS. The dose to the bone is substantially increased while the dose beyond the bone is decreased for field sizes up to 3.0 cm. Combining this effect with additional significant density changes (either air or lung) substantially changes the dose to tissues along the beamlet path. These effects may have clinical implications in the delivery of ultra-small fields in IMRT or SRS. 1 Jones AO, Das IJ, and Jones FL, A Monte Carlo Study of IMRT Beamlets in Inhomogeneous Media, Med. Phys. 30: 296 – 300 2003.
2485
Aperture-Based IMRT Inverse Planning with Incorporation of Organ Motion
Z. Shou, L. Xing Radiation Oncology, Stanford University School of Medicine, Stanford, CA Purpose/Objective: Accurate targeting is of paramount importance in intensity modulated radiation therapy (IMRT). In the presence of organ motion, the problem becomes complicated because of the interplay between the organ motion and MLC movement. The purpose of this work is to develop a general aperture-based inverse planning algorithm with consideration of organ motion to compensate or minimize the influence of the interplay between the two types of movements. Materials/Methods: We implemented a hybrid aperture-based inverse planning algorithm that is capable of producing optimal shapes and weights of a pre-selected number of segments for each incident field. In this algorithm the variables defining the shapes of the involved segments and the weights of the segments were dealt separately by using a fast simulated annealing (ASA) and a descent gradient optimization. Two different scenarios were investigated: (i) the shapes of the organs (target and/or sensitive structures) remain unchanged during the organ motion process; and (ii) the shapes of the organs are deformable. A pre-requisition of the proposed aperture-based inverse planning algorithm is that the temporal process of each point within a structure has been acquired either before the treatment or during the treatment through the use of an adaptive algorithm and an online imaging system. In either case, the overall goal of the inverse planning is to minimize the temporally integrated dose to the sensitive structures while delivering a uniform integrated dose distribution to the target volume. The algorithm was first examined by using a phantom case with sinusoidally moving target and/or sensitive structures. The proposed technique was then used to plan four breast cases and dose distributions were computed for three different situations: (a) conventional aperture-
S631
S632
I. J. Radiation Oncology
● Biology ● Physics
Volume 60, Number 1, Supplement, 2004
based IMRT treatment without considering the breathing motion (other than the use of appropriate margins); (b) new aperture-based IMRT treatment with consideration of breathing motion; and (c) a hypothetical treatment where the segmented beams derived from (i) was used to irradiate the beast target volumes with breathing motion. Two opposed tangential fields (the beam energy was either 6 MV or 15 MV), each consists of 12 to 25 segments, were used for these plans. The three types of plans were compared quantitatively. Results: In both phantom and clinical case studies, we found that the organ motion severely deteriorated the dose distributions obtained using conventional aperture-based IMRT inverse planning with static organs even when appropriate margins were used. Generally, the level of deterioration increases with the level of intensity modulation or the number of segments of the incident beams. With the new algorithm, it was possible to significantly reduce the adverse dosimetric effect of the organ motion. For the same number of segments, the quality of the plan with organ motion was generally inferior to that without organ motion. To achieve similar target dose coverage and sensitive structure sparing as the static case, 15%– 40% more segments were generally required when the organ motion was switched on. Further increase of the segment number would lead to better but clinically insignificant improvement. The demand for more segments tends to be less as the original number of segments for the static plan is large. The comparison results strongly depended on the amplitude of organ motion. It was also found that the deformation of the involved organs played less significant role provided that the level of deformation was within its practical range and provided that the deformation was accounted for on a case specific basis by using the inverse planning technique proposed here. Conclusions: An inverse planning system with incorporation of organ motion was developed for aperture-based IMRT and the influence of organ motion on the final dose was investigated systematically. It was shown that the adverse dosimetric effect of organ motion can be compensated or minimized by the modified inverse planning algorithm with moderate increase of segmented fields. Together with organ tracking/imaging techniques, the approach may provide a practical solution to the clinically intractable problem of organ motion.
2486
IMRT Simultaneous Integrated Boost for High-Risk Prostate Cancer
X. Li,1,2 J. Z. Wang,2 P. Jursinic,1 C. A. Lawton1 1
Radiation Oncology, Medical College of Wisconsin, Milwaukee, WI, 2Radiation Oncology, University of Maryland, Baltimore, MD Purpose/Objective: A sequential two-phase process, initial and boost irradiation, is commonly used for high risk prostate cancer. In this work, we propose to use IMRT simultaneous integrated boost (SIB) to replace this two-phase process. The SIB which is a single-phase process can simultaneously deliver high dose to prostate and lower dose to pelvic nodes. Materials/Methods: The SIB is designed to irradiate pelvic nodes with the same fractionations as conventional two-phase treatment (e.g., 25 ⫻ 1.8 Gy) and to deliver higher doses to prostate in the same 25 fractions (i.e., hypofractionation). The concept of equivalent uniform dose (EUD) was used to determine suitable SIB fractionations that deliver the biologically equivalent doses to prostate. The EUD for prostate is estimated based on LQ model. The most recent LQ parameters derived from clinical data for prostate cancer were used. The EUD for normal tissue was computed based on the Lyman model. To compare biological effectiveness voxel by voxel, we propose a new concept, termed the voxel equivalent dose (VED). The calculation of VED is similar to that for EUD except it is done within a voxel. The 3D VED distributions and the VED-based DVHs were used to compare biological and dosimetric merits for newly-designed SIB regimens. To demonstrate dosimetric feasibility and advantages of the proposed SIB, we have performed a planning study on selected patient CT using a commercial IMRT and a 3D planning systems. Four treatment scenarios were considered: (1) 3D initial plus 3D boost, (2) 3D initial plus IMRT boost, (3) IMRT initial plus IMRT boost, and (4) IMRT SIB. EUDs and VED-based DVHs for prostate, pelvic nodes, small bowel, rectum, bladder and other tissue for all 4 scenarios are compared to show dosimetric and biological benefits of the IMRT SIB. Results: A series of equivalent hypofractionation regimens suitable for the IMRT SIB are obtained. The table(table 1) below lists the required SIB prescribed doses that are biologically equivalent to a series of desired EUDs. The data are calculated using LQ parameters of ␣ ⫽ 0.15 Gy-1, ␣/ ⫽ 3.1 Gy. The data in the table may be used to design appropriate SIB schemes that are biologically equivalent to the conventional treatments. For example, the conventional regimen of 42 ⫻ 1.8 Gy (EUD ⫽ 75.4 Gy) would be equivalent to a SIB regimen of 25 ⫻ 2.54 Gy. From comparisons of 3D VED distributions and DVHs between the SIB and the conventional two-phase treatments, we found that the SIB offers more conformal dose distributions to prostate and pelvic nodes and provides better sparing to the critical structures. The figure below compares VED-based DVHs for rectum for the 4 scenarios with an EUD of 75.4 Gy. Conclusions: A new IMRT simultaneous integrated boost strategy that irradiates prostate with hypofractionations while irradiating pelvic nodes with the conventional fractionations is proposed for high-risk prostate cancer. Compared with the conventional two-phase treatment, the proposed SIB technique offers potential advantages including better sparing of critical structures, more efficient delivery, shorter treatment duration, and higher biological effectiveness.