skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: SU-F-T-342: Dosimetric Constraint Prediction Guided Automatic Mulit-Objective Optimization for Intensity Modulated Radiotherapy

Abstract

Purpose: For intensity modulated radiotherapy, the plan optimization is time consuming with difficulties of selecting objectives and constraints, and their relative weights. A fast and automatic multi-objective optimization algorithm with abilities to predict optimal constraints and manager their trade-offs can help to solve this problem. Our purpose is to develop such a framework and algorithm for a general inverse planning. Methods: There are three main components contained in this proposed multi-objective optimization framework: prediction of initial dosimetric constraints, further adjustment of constraints and plan optimization. We firstly use our previously developed in-house geometry-dosimetry correlation model to predict the optimal patient-specific dosimetric endpoints, and treat them as initial dosimetric constraints. Secondly, we build an endpoint(organ) priority list and a constraint adjustment rule to repeatedly tune these constraints from their initial values, until every single endpoint has no room for further improvement. Lastly, we implement a voxel-independent based FMO algorithm for optimization. During the optimization, a model for tuning these voxel weighting factors respecting to constraints is created. For framework and algorithm evaluation, we randomly selected 20 IMRT prostate cases from the clinic and compared them with our automatic generated plans, in both the efficiency and plan quality. Results: For each evaluatedmore » plan, the proposed multi-objective framework could run fluently and automatically. The voxel weighting factor iteration time varied from 10 to 30 under an updated constraint, and the constraint tuning time varied from 20 to 30 for every case until no more stricter constraint is allowed. The average total costing time for the whole optimization procedure is ∼30mins. By comparing the DVHs, better OAR dose sparing could be observed in automatic generated plan, for 13 out of the 20 cases, while others are with competitive results. Conclusion: We have successfully developed a fast and automatic multi-objective optimization for intensity modulated radiotherapy. This work is supported by the National Natural Science Foundation of China (No: 81571771)« less

Authors:
;  [1];  [2]
  1. Southern Medical University, Guangzhou, Guangdong (China)
  2. Beihang University, Beijing, Beijing (China)
Publication Date:
OSTI Identifier:
22648944
Resource Type:
Journal Article
Resource Relation:
Journal Name: Medical Physics; Journal Volume: 43; Journal Issue: 6; Other Information: (c) 2016 American Association of Physicists in Medicine; Country of input: International Atomic Energy Agency (IAEA)
Country of Publication:
United States
Language:
English
Subject:
60 APPLIED LIFE SCIENCES; 61 RADIATION PROTECTION AND DOSIMETRY; ALGORITHMS; DOSIMETRY; FORECASTING; LIMITING VALUES; OPTIMIZATION; PLANNING; RADIOTHERAPY

Citation Formats

Song, T, Zhou, L, and Li, Y. SU-F-T-342: Dosimetric Constraint Prediction Guided Automatic Mulit-Objective Optimization for Intensity Modulated Radiotherapy. United States: N. p., 2016. Web. doi:10.1118/1.4956527.
Song, T, Zhou, L, & Li, Y. SU-F-T-342: Dosimetric Constraint Prediction Guided Automatic Mulit-Objective Optimization for Intensity Modulated Radiotherapy. United States. doi:10.1118/1.4956527.
Song, T, Zhou, L, and Li, Y. 2016. "SU-F-T-342: Dosimetric Constraint Prediction Guided Automatic Mulit-Objective Optimization for Intensity Modulated Radiotherapy". United States. doi:10.1118/1.4956527.
@article{osti_22648944,
title = {SU-F-T-342: Dosimetric Constraint Prediction Guided Automatic Mulit-Objective Optimization for Intensity Modulated Radiotherapy},
author = {Song, T and Zhou, L and Li, Y},
abstractNote = {Purpose: For intensity modulated radiotherapy, the plan optimization is time consuming with difficulties of selecting objectives and constraints, and their relative weights. A fast and automatic multi-objective optimization algorithm with abilities to predict optimal constraints and manager their trade-offs can help to solve this problem. Our purpose is to develop such a framework and algorithm for a general inverse planning. Methods: There are three main components contained in this proposed multi-objective optimization framework: prediction of initial dosimetric constraints, further adjustment of constraints and plan optimization. We firstly use our previously developed in-house geometry-dosimetry correlation model to predict the optimal patient-specific dosimetric endpoints, and treat them as initial dosimetric constraints. Secondly, we build an endpoint(organ) priority list and a constraint adjustment rule to repeatedly tune these constraints from their initial values, until every single endpoint has no room for further improvement. Lastly, we implement a voxel-independent based FMO algorithm for optimization. During the optimization, a model for tuning these voxel weighting factors respecting to constraints is created. For framework and algorithm evaluation, we randomly selected 20 IMRT prostate cases from the clinic and compared them with our automatic generated plans, in both the efficiency and plan quality. Results: For each evaluated plan, the proposed multi-objective framework could run fluently and automatically. The voxel weighting factor iteration time varied from 10 to 30 under an updated constraint, and the constraint tuning time varied from 20 to 30 for every case until no more stricter constraint is allowed. The average total costing time for the whole optimization procedure is ∼30mins. By comparing the DVHs, better OAR dose sparing could be observed in automatic generated plan, for 13 out of the 20 cases, while others are with competitive results. Conclusion: We have successfully developed a fast and automatic multi-objective optimization for intensity modulated radiotherapy. This work is supported by the National Natural Science Foundation of China (No: 81571771)},
doi = {10.1118/1.4956527},
journal = {Medical Physics},
number = 6,
volume = 43,
place = {United States},
year = 2016,
month = 6
}
  • The objective of this work is to develop a dosimetric phantom quality assurance (QA) of linear accelerators capable of cone-beam CT (CBCT) image guided and intensity-modulated radiotherapy (IG-IMRT). This phantom is to be used in an integral test to quantify in real-time both the performance of the image guidance and the dose delivery systems in terms of dose localization. The prototype IG-IMRT QA phantom consisted of a cylindrical imaging phantom (CatPhan) combined with an array of 11 radiation diodes mounted on a 10 cm diameter disk, oriented perpendicular to the phantom axis. Basic diode response characterization was performed for 6more » and 18 MV photons. The diode response was compared to planning system calculations in the open and penumbrae regions of simple and complex beam arrangements. The clinical use of the QA phantom was illustrated in an integral test of an IG-IMRT treatment designed for a clinical spinal radiosurgery case. The sensitivity of the phantom to multileaf collimator (MLC) calibration and setup errors in the clinical setting was assessed by introducing errors in the IMRT plan or by displacing the phantom. The diodes offered good response linearity and long-term reproducibility for both 6 and 18 MV. Axial dosimetry of coplanar beams (in a plane containing the beam axes) was made possible with the nearly isoplanatic response of the diodes over 360 deg. of gantry (usually within {+-}1%). For single beam geometry, errors in phantom placement as small as 0.5 mm could be accurately detected (in gradient {>=}1%/mm). In clinical setting, MLC systematic errors of 1 mm on a single MLC bank introduced in the IMRT plan were easily detectable with the QA phantom. The QA phantom demonstrated also sufficient sensitivity for the detection of setup errors as small as 1 mm for the IMRT delivery. These results demonstrated that the prototype can accurately and efficiently verify the entire IG-IMRT process. This tool, in conjunction with image guidance capabilities has the potential to streamline this QA process and improve the level of performance of image guided and intensity modulated radiotherapy.« less
  • To determine whether residual interfraction seminal vesicle (SV) displacement necessitates specific planning target volume (PTV) margins during fiducial-guided intensity modulated radiation therapy (IMRT) of the prostate. A planning computed tomography (CT) scan and 2 subsequent CT scans were prospectively obtained for 20 prostate cancer patients with intraprostatic fiducial markers. After CT registration, SV displacement relative to the prostate was quantified as a function of margin size for both the proximal (1 cm) SV (PSV) and the full SV (FSV). Two IMRT plans were simulated for each patient (prostate + PSV and prostate + FSV) both with a uniform 5-mm PTVmore » margin. Minimum clinical target volume (CTV) dose (D{sub min}) and the volume of SV receiving 95% of the prescription dose (V{sub 95%}) were assessed during treatment and compared with the initial plan. In all cases, SV displacement with respect to the prostate was greater for the FSV compared with the PSV. To ensure at least 95% geometrical coverage of the CTV for 90% of patients, margins of 5 and 8 mm were required for the PSV and FSV, respectively. Dosimetrically, residual SV displacement had minimal impact on PSV coverage compared with FSV coverage. For the PSV D{sub min} was {>=}95% of the prescribed dose in 90% of patients with an overall mean V{sub 95%} of 99.6 {+-} 0.8%; for the FSV D{sub min} was {>=}95% of the prescribed dose in only 45% of patients with a mean V{sub 95%} of 97.9 {+-} 2.4%. The SVs move differentially from the prostate and exhibit greater variation with increasing distance from the prostate. For plans targeting just the prostate and PSVs, 5-mm PTV expansions are adequate. However, despite daily localization of the prostate, larger PTV margins are required for cases where the intent is to completely cover the FSV.« less
  • To develop a quality assurance (QA) of XVI cone beam system (XVIcbs) for its optimal imaging-guided radiotherapy (IGRT) implementation, and to construe prostate tumor margin required for intensity-modulated radiation therapy (IMRT) if IGRT is unavailable. XVIcbs spatial accuracy was explored with a humanoid phantom; isodose conformity to lesion target with a rice phantom housing a soap as target; image resolution with a diagnostic phantom; and exposure validation with a Radcal ion chamber. To optimize XVIcbs, rotation flexmap on coincidency between gantry rotational axis and that of XVI cone beam scan was investigated. Theoretic correlation to image quality of XVIcbs rotationalmore » axis stability was elaborately studied. Comprehensive QA of IGRT using XVIcbs has initially been explored and then implemented on our general IMRT treatments, and on special IMRT radiotherapies such as head and neck (H and N), stereotactic radiation therapy (SRT), stereotactic radiosurgery (SRS), and stereotactic body radiotherapy (SBRT). Fifteen examples of prostate setup accounted for 350 IGRT cone beam system were analyzed. IGRT accuracy results were in agreement {+-} 1 mm. Flexmap 0.25 mm met the manufacturer's specification. Films confirmed isodose coincidence with target (soap) via XVIcbs, otherwise not. Superficial doses were measured from 7.2-2.5 cGy for anatomic diameters 15-33 cm, respectively. Image quality was susceptible to rotational stability or patient movement. IGRT using XVIcbs on general IMRT treatments such as prostate, SRT, SRS, and SBRT for setup accuracy were verified; and subsequently coordinate shifts corrections were recorded. The 350 prostate IGRT coordinate shifts modeled to Gaussian distributions show central peaks deviated off the isocenter by 0.6 {+-} 3.0 mm, 0.5 {+-} 4.5 mm in the X(RL)- and Z(SI)-coordinates, respectively; and 2.0 {+-} 3.0 mm in the Y(AP)-coordinate as a result of belly and bladder capacity variations. Sixty-eight percent of confidence was within {+-} 4.5 mm coordinates shifting. IGRT using XVIcbs is critical to IMRT for prostate and H and N, especially SRT, SRS, and SBRT. To optimize this modality of IGRT, a vigilant QA program is indispensable. Prostate IGRT reveals treatment accuracy as subject to coordinates' adjustments; otherwise a 4.5-mm margin is required to allow for full dose coverage of the clinical target volume, notwithstanding toxicity to normal tissues.« less
  • Purpose: Cranial irradiation in pediatric patients is associated with serious long-term adverse effects. We sought to determine whether both three-dimensional conformal proton radiotherapy (3D-PRT) and intensity-modulated proton therapy (IMPT) compared with intensity-modulated radiotherapy (IMRT) decrease integral dose to brain areas known to harbor neuronal stem cells, major blood vessels, and other normal brain structures for pediatric patients with craniopharyngiomas. Methods and Materials: IMRT, forward planned, passive scattering proton, and IMPT plans were generated and optimized for 10 pediatric patients. The dose was 50.4 Gy (or cobalt Gy equivalent) delivered in 28 fractions with the requirement for planning target volume (PTV)more » coverage of 95% or better. Integral dose data were calculated from differential dose-volume histograms. Results: The PTV target coverage was adequate for all modalities. IMRT and IMPT yielded the most conformal plans in comparison to 3D-PRT. Compared with IMRT, 3D-PRT and IMPT plans had a relative reduction of integral dose to the hippocampus (3D-PRT, 20.4; IMPT, 51.3%{sup Asterisk-Operator }), dentate gyrus (27.3, 75.0%{sup Asterisk-Operator }), and subventricular zone (4.5, 57.8%{sup Asterisk-Operator }). Vascular organs at risk also had reduced integral dose with the use of proton therapy (anterior cerebral arteries, 33.3{sup Asterisk-Operator }, 100.0%{sup Asterisk-Operator }; middle cerebral arteries, 25.9%{sup Asterisk-Operator }, 100%{sup Asterisk-Operator }; anterior communicating arteries, 30.8{sup Asterisk-Operator }, 41.7%{sup Asterisk-Operator }; and carotid arteries, 51.5{sup Asterisk-Operator }, 77.6{sup Asterisk-Operator }). Relative reduction of integral dose to the infratentorial brain (190.7{sup Asterisk-Operator }, 109.7%{sup Asterisk-Operator }), supratentorial brain without PTV (9.6, 26.8%{sup Asterisk-Operator }), brainstem (45.6, 22.4%{sup Asterisk-Operator }), and whole brain without PTV (19.4{sup Asterisk-Operator }, 34.4%{sup Asterisk-Operator }) were recorded with the use of proton therapy. ({sup Asterisk-Operator }Differences were significant based on Friedman's test with Bonferroni-Dunn correction, {alpha} = 0.05) Conclusions: The current study found that proton therapy was able to avoid excess integral radiation dose to a variety of normal structures at all dose levels while maintaining equal target coverage. Future studies will examine the clinical benefits of these dosimetric advantages.« less
  • Volumetric-modulated arc radiotherapy (VMAT) is an iteration of intensity-modulated radiotherapy (IMRT), both of which deliver highly conformal dose distributions. Studies have shown the superiority of VMAT and IMRT in comparison with 3-dimensional conformal radiotherapy (3D-CRT) in planning target volume (PTV) coverage and organs-at-risk (OARs) sparing. This is the first study examining the benefits of VMAT in pancreatic cancer for doses more than 55.8 Gy. A planning study comparing 3D-CRT, IMRT, and VMAT was performed in 20 patients with pancreatic cancer. Treatments were planned for a 25-fraction delivery of 45 Gy to a large field followed by a reduced-volume 8-fraction externalmore » beam boost to 59.4 Gy in total. OARs and PTV doses, conformality index (CI) deviations from 1.0, monitor units (MUs) delivered, and isodose volumes were compared. IMRT and VMAT CI deviations from 1.0 for the large-field and the boost plans were equivalent (large field: 0.032 and 0.046, respectively; boost: 0.042 and 0.037, respectively; p > 0.05 for all comparisons). Both IMRT and VMAT CI deviations from 1.0 were statistically superior to 3D-CRT (large field: 0.217, boost: 0.177; p < 0.05 for all comparisons). VMAT showed reduction of the mean dose to the boost PTV (VMAT: 61.4 Gy, IMRT: 62.4 Gy, and 3D-CRT: 62.3 Gy; p < 0.05). The mean number of MUs per fraction was significantly lower for VMAT for both the large-field and the boost plans. VMAT delivery time was less than 3 minutes compared with 8 minutes for IMRT. Although no statistically significant dose reduction to the OARs was identified when comparing VMAT with IMRT, VMAT showed a reduction in the volumes of the 100% isodose line for the large-field plans. Dose escalation to 59.4 Gy in pancreatic cancer is dosimetrically feasible with shorter treatment times, fewer MUs delivered, and comparable CIs for VMAT when compared with IMRT.« less