Abstract:Purpose/Objective: Leaf-sequencing in the Corvus (Nomos, PA) planning system takes place after the optimization of the treatment plan to be delivered in step-and-shoot IMRT. The Corvus leaf-sequencing algorithm generates a large number of apertures, which can result in significant treatment delivery time. We present a new approach for the intensity map decomposition problem that is solved to obtain leaf sequencing.Materials/Methods: Mathematically, the version of the problem that we consider may be modeled as the optimal decomposition of a matrix M of nonnegative integer intensities into a weighted sum: M = σ ci Si where the intensities ci and the binary matrices Si are to be determined by the algorithm so as to make the number of matrices (segments) and sum of the intensities σ ci in the decomposition as small as possible. For certain linacs (e.g., Elekta), the transition time between segments can dominate the actual radiation delivery time. In order to define the difference matrices used in our approach, we assume that we are given an m × n intensity map M = (Mij) and we add bordering columns on the left and right via ghost entries Mi,-1 = Mi,n+1 = 0 for i=1,…,m. After this augmentation, the difference operator Δ and initial difference matrix D are defined as follows: D=ΔM=(Dij) is the m × (n+1) matrix such that Dij=Mi,j-Mi,j-1. The algorithm generates a finite sequence of intensity maps and corresponding difference matrices, at each stage extracting from the current intensity map a shape matrix and its corresponding intensity multiplier, followed by a corresponding update of the intensity map and difference matrices. The extraction process is guided by examination of the current difference matrix. The chosen extraction is the one that zeros out the maximum number of entries in the difference matrix. This approach was applied to clinical test cases: head and neck and prostate. Plans were generated with 5, 10 and 100 intensity levels. The number of apertures (segments) and monitor units (beam-on time) resulting from our approach was compared with the output from Corvus and a previously published method by Boland et al (BHL).Results: A 5-field and 6-field plan was generated for the head and neck and prostate cases, respectively. The prescription doses were 45 Gy and 75.6 Gy for the head and neck and prostate cases, respectively. The reduction in the number of segments with our approach was 72%, 67% and 32% with 5, 10 and 100 intensity levels, respectively compared with Corvus for the head and neck case. The corresponding reduction compared to the BHL method was 74%, 62% and 13%, respectively. For the prostate case, the reduction in segments was 65%, 62% and 21% with 5, 10 and 100 intensity levels, respectively between our method and Corvus. The corresponding reduction in comparison with the BHL method was 60%, 51% and 6%. The number of monitor units is reduced by > 20% compared with the Corvus plans. We have processed the aperture shapes from our approach and the Corvus leaf sequence through a Monte Carlo calculation. The intensity profiles outside the beam’s-eye-view (BEV) projection of the tumor display leakage radiation with the Corvus leaf sequence which, can be attributed to the use of the partial transmission block used by Corvus in conjunction with Elekta linear accelerators. The leakage radiation outside the BEV from our approach is significantly reduced. However, tongue-and-groove effects are more prominent in our approach. The impact of tongue-and-groove has been studied by calculating a full 3-D dose distribution of the Corvus leaf sequence and the output from our method. Differences in the dose distribution resulting from our sequencing approach and the Corvus plan are negligible.Conclusions: Our results demonstrate that we have developed a fast heuristic for leaf sequencing which reduces the number of aperture shapes and beam-on time significantly compared with the output from Corvus and existing methods. We estimate that the overall delivery time will be 15 to 20 minutes with 5 to 10 intensity levels, a significant reduction compared to current delivery times. Purpose/Objective: Leaf-sequencing in the Corvus (Nomos, PA) planning system takes place after the optimization of the treatment plan to be delivered in step-and-shoot IMRT. The Corvus leaf-sequencing algorithm generates a large number of apertures, which can result in significant treatment delivery time. We present a new approach for the intensity map decomposition problem that is solved to obtain leaf sequencing. Materials/Methods: Mathematically, the version of the problem that we consider may be modeled as the optimal decomposition of a matrix M of nonnegative integer intensities into a weighted sum: M = σ ci Si where the intensities ci and the binary matrices Si are to be determined by the algorithm so as to make the number of matrices (segments) and sum of the intensities σ ci in the decomposition as small as possible. For certain linacs (e.g., Elekta), the transition time between segments can dominate the actual radiation delivery time. In order to define the difference matrices used in our approach, we assume that we are given an m × n intensity map M = (Mij) and we add bordering columns on the left and right via ghost entries Mi,-1 = Mi,n+1 = 0 for i=1,…,m. After this augmentation, the difference operator Δ and initial difference matrix D are defined as follows: D=ΔM=(Dij) is the m × (n+1) matrix such that Dij=Mi,j-Mi,j-1. The algorithm generates a finite sequence of intensity maps and corresponding difference matrices, at each stage extracting from the current intensity map a shape matrix and its corresponding intensity multiplier, followed by a corresponding update of the intensity map and difference matrices. The extraction process is guided by examination of the current difference matrix. The chosen extraction is the one that zeros out the maximum number of entries in the difference matrix. This approach was applied to clinical test cases: head and neck and prostate. Plans were generated with 5, 10 and 100 intensity levels. The number of apertures (segments) and monitor units (beam-on time) resulting from our approach was compared with the output from Corvus and a previously published method by Boland et al (BHL). Results: A 5-field and 6-field plan was generated for the head and neck and prostate cases, respectively. The prescription doses were 45 Gy and 75.6 Gy for the head and neck and prostate cases, respectively. The reduction in the number of segments with our approach was 72%, 67% and 32% with 5, 10 and 100 intensity levels, respectively compared with Corvus for the head and neck case. The corresponding reduction compared to the BHL method was 74%, 62% and 13%, respectively. For the prostate case, the reduction in segments was 65%, 62% and 21% with 5, 10 and 100 intensity levels, respectively between our method and Corvus. The corresponding reduction in comparison with the BHL method was 60%, 51% and 6%. The number of monitor units is reduced by > 20% compared with the Corvus plans. We have processed the aperture shapes from our approach and the Corvus leaf sequence through a Monte Carlo calculation. The intensity profiles outside the beam’s-eye-view (BEV) projection of the tumor display leakage radiation with the Corvus leaf sequence which, can be attributed to the use of the partial transmission block used by Corvus in conjunction with Elekta linear accelerators. The leakage radiation outside the BEV from our approach is significantly reduced. However, tongue-and-groove effects are more prominent in our approach. The impact of tongue-and-groove has been studied by calculating a full 3-D dose distribution of the Corvus leaf sequence and the output from our method. Differences in the dose distribution resulting from our sequencing approach and the Corvus plan are negligible. Conclusions: Our results demonstrate that we have developed a fast heuristic for leaf sequencing which reduces the number of aperture shapes and beam-on time significantly compared with the output from Corvus and existing methods. We estimate that the overall delivery time will be 15 to 20 minutes with 5 to 10 intensity levels, a significant reduction compared to current delivery times.

Abstract:Purpose/ObjectiveAccurate 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/MethodsWe 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-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.ResultsIn 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.ConclusionsAn 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 Purpose/ObjectiveAccurate 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. 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/MethodsWe 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-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. 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-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. ResultsIn 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. 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. ConclusionsAn 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 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

Beam Angle Selection (BAS) Via Nested Partitions-based Global Search Using a Fast Beam Set Scoring Process

Quantitative Comparison of Aperture-Based and Beamlet-Based Inverse Planning Techniques

Impact Of Interfractional Motion On Hypofractionated Pencil Beam Scanning Proton Therapy And Vmat Delivery For Prostate Cancer

MO-D-351-03: A Unified Approach to Beam Angle Selection and Dose Optimization with High-Throughput Computing for IMRT

A Surrogate-Based Metaheuristic Global Search Method for Beam Angle Selection in Radiation Treatment Planning

WE-E-BRA-02: Clinical Evaluation of Fixed-Field IMRT Plans and IMAT Plans in Locally Advanced Head and Neck Cancer

SU-E-T-713: Study On Application of Beam Angle Optimization for Multiple Intracranial Metastases

Atlas-based Dose Optimization Via Linear Programming for IMRT Using a High-throughput Computing Environment

SU-FF-T-30: A Nested Partitions Framework for Beam Angle and Dose Optimization in IMRT

A data-driven approach to optimal beam/arc angle selection for liver stereotactic body radiation therapy treatment planning

Leaf sequencing via difference matrices reduces aperture number and beam-on time in IMRT

Quantification of Dosimetry Improvement With or Without Patient Surface Guidance

Aperture-based IMRT Inverse Planning with Incorporation of Organ Motion

A Nested Partitions Framework for Beam Angle Optimization in Intensity-Modulated Radiation Therapy

A Novel Dose Rate Optimization Method to Maximize Ultra-High Dose Rate Coverage of Critical Organs-at-risk Without Compromising Dosimetry Metrics in Proton Pencil Beam Scanning FLASH Radiotherapy

Beam specific planning target volumes incorporating 4DCT for pencil beam scanning proton therapy of thoracic tumors

Solving Beam-Angle Selection and Dose Optimization Simultaneously Via High-Throughput Computing

MO-D-351-05: A “slicing” Approach for IMRT Optimization Using Linear Programming

Selection of Beam Orientations in Intensity-Modulated Radiation Therapy Using Single-Beam Indices and Integer Programming

Guaranteed epsilon-optimal treatment plans with minimum number of beams for stereotactic body radiation therapy

SU‐FF‐T‐242: High‐Throughput Computing in Condor Using a Nested Partitions Framework for IMRT Beam and Dose Optimization