Journal Articles
2020
Antaki, Majd; Deufel, Christopher L; Enger, Shirin A.
Fast mixed integer optimization (FMIO) for high dose rate brachytherapy Journal Article
In: Physics in Medicine and Biology, vol. 65, no. 21, pp. 215005, 2020, ISSN: 1361-6560.
Abstract | Links | BibTeX | Tags: Algorithms, Brachytherapy, Computer-Assisted, Humans, Linear Models, Male, Monte Carlo Method, Prostatic Neoplasms, Radiation Dosage, Radiotherapy Dosage, Radiotherapy Planning, Software, Time Factors
@article{antaki_fast_2020,
title = {Fast mixed integer optimization (FMIO) for high dose rate brachytherapy},
author = {Majd Antaki and Christopher L Deufel and Shirin A. Enger},
doi = {10.1088/1361-6560/aba317},
issn = {1361-6560},
year = {2020},
date = {2020-12-01},
journal = {Physics in Medicine and Biology},
volume = {65},
number = {21},
pages = {215005},
abstract = {The purpose of this work was to develop an efficient quadratic mixed integer programming algorithm for high dose rate (HDR) brachytherapy treatment planning problems and integrate the algorithm into an open-source Monte Carlo based treatment planning software, RapidBrachyMCTPS. The mixed-integer algorithm yields a globally optimum solution to the dose volume histogram (DVH) based problem and, unlike other methods, is not susceptible to local minimum trapping. A hybrid linear-quadratic penalty model coupled to a mixed integer programming model was used to optimize treatment plans for 10 prostate cancer patients. Dose distributions for each dwell position were calculated with RapidBrachyMCTPS with type A uncertainties less than 0.2% in voxels within the planning target volume (PTV). The optimization process was divided into two parts. First, the data was preprocessed, in which the problem size was reduced by eliminating voxels that had negligible impact on the solution (e.g. far from the dwell position). Second, the best combination of dwell times to obtain a plan with the highest score was found. The dwell positions and dose volume constraints were used as input to a commercial mixed integer optimizer (Gurobi Optimization, Inc.). A penalty-based criterion was adopted for the scoring. The voxel-reduction technique successfully reduced the problem size by an average of 91%, without loss of quality. The preprocessing of the optimization process required on average 4 s and solving for the global maximum required on average 33 s. The total optimization time averaged 37 s, which is a substantial improvement over the ∼15 min optimization time reported in published literature. The plan quality was evaluated by evaluating dose volume metrics, including PTV D90, rectum and bladder D1cc and urethra D0.1cc. In conclusion, fast mixed integer optimization is an order of magnitude faster than current mixed-integer approaches for solving HDR brachytherapy treatment planning problems with DVH based metrics.},
keywords = {Algorithms, Brachytherapy, Computer-Assisted, Humans, Linear Models, Male, Monte Carlo Method, Prostatic Neoplasms, Radiation Dosage, Radiotherapy Dosage, Radiotherapy Planning, Software, Time Factors},
pubstate = {published},
tppubtype = {article}
}
The purpose of this work was to develop an efficient quadratic mixed integer programming algorithm for high dose rate (HDR) brachytherapy treatment planning problems and integrate the algorithm into an open-source Monte Carlo based treatment planning software, RapidBrachyMCTPS. The mixed-integer algorithm yields a globally optimum solution to the dose volume histogram (DVH) based problem and, unlike other methods, is not susceptible to local minimum trapping. A hybrid linear-quadratic penalty model coupled to a mixed integer programming model was used to optimize treatment plans for 10 prostate cancer patients. Dose distributions for each dwell position were calculated with RapidBrachyMCTPS with type A uncertainties less than 0.2% in voxels within the planning target volume (PTV). The optimization process was divided into two parts. First, the data was preprocessed, in which the problem size was reduced by eliminating voxels that had negligible impact on the solution (e.g. far from the dwell position). Second, the best combination of dwell times to obtain a plan with the highest score was found. The dwell positions and dose volume constraints were used as input to a commercial mixed integer optimizer (Gurobi Optimization, Inc.). A penalty-based criterion was adopted for the scoring. The voxel-reduction technique successfully reduced the problem size by an average of 91%, without loss of quality. The preprocessing of the optimization process required on average 4 s and solving for the global maximum required on average 33 s. The total optimization time averaged 37 s, which is a substantial improvement over the ∼15 min optimization time reported in published literature. The plan quality was evaluated by evaluating dose volume metrics, including PTV D90, rectum and bladder D1cc and urethra D0.1cc. In conclusion, fast mixed integer optimization is an order of magnitude faster than current mixed-integer approaches for solving HDR brachytherapy treatment planning problems with DVH based metrics. Famulari, Gabriel; Alfieri, Joanne; Duclos, Marie; Vuong, Té; Enger, Shirin A.
Can intermediate-energy sources lead to elevated bone doses for prostate and head & neck high-dose-rate brachytherapy? Journal Article
In: Brachytherapy, vol. 19, no. 2, pp. 255–263, 2020, ISSN: 1873-1449.
Abstract | Links | BibTeX | Tags: Bone and Bones, Brachytherapy, Cobalt Radioisotopes, Computer Simulation, Computer-Assisted, Dose calculation, Gadolinium, Humans, Intermediate-energy source, Iridium Radioisotopes, Male, Monte Carlo, Prostatic Neoplasms, Radiation Dosage, Radioisotopes, Radiotherapy Dosage, Radiotherapy Planning, Selenium Radioisotopes, Tissue composition, Tongue Neoplasms, Ytterbium
@article{famulari_can_2020,
title = {Can intermediate-energy sources lead to elevated bone doses for prostate and head & neck high-dose-rate brachytherapy?},
author = {Gabriel Famulari and Joanne Alfieri and Marie Duclos and Té Vuong and Shirin A. Enger},
doi = {10.1016/j.brachy.2019.12.004},
issn = {1873-1449},
year = {2020},
date = {2020-04-01},
journal = {Brachytherapy},
volume = {19},
number = {2},
pages = {255--263},
abstract = {PURPOSE: Several radionuclides with high (60Co, 75Se) and intermediate (169Yb, 153Gd) energies have been investigated as alternatives to 192Ir for high-dose-rate brachytherapy. The purpose of this study was to evaluate the impact of tissue heterogeneities for these five high- to intermediate-energy sources in prostate and head & neck brachytherapy. METHODS AND MATERIALS: Treatment plans were generated for a cohort of prostate (n = 10) and oral tongue (n = 10) patients. Dose calculations were performed using RapidBrachyMCTPS, an in-house Geant4-based Monte Carlo treatment planning system. Treatment plans were simulated using 60Co, 192Ir, 75Se, 169Yb, and 153Gd as the active core of the microSelectron v2 source. Two dose calculation scenarios were presented: (1) dose to water in water (Dw,w), and (2) dose to medium in medium (Dm,m).
RESULTS: Dw,w overestimates planning target volume coverage compared with Dm,m, regardless of photon energy. The average planning target volume D90 reduction was ∼1% for high-energy sources, whereas larger differences were observed for intermediate-energy sources (1%-2% for prostate and 4%-7% for oral tongue). Dose differences were not clinically relevant (textless5%) for soft tissues in general. Going from Dw,w to Dm,m, bone doses were increased two- to three-fold for 169Yb and four- to five-fold for 153Gd, whereas the ratio was close to ∼1 for high-energy sources.
CONCLUSIONS: Dw,w underestimates the dose to bones and, to a lesser extent, overestimates the dose to soft tissues for radionuclides with average energies lower than 192Ir. Further studies regarding bone toxicities are needed before intermediate-energy sources can be adopted in cases where bones are in close vicinity to the tumor.},
keywords = {Bone and Bones, Brachytherapy, Cobalt Radioisotopes, Computer Simulation, Computer-Assisted, Dose calculation, Gadolinium, Humans, Intermediate-energy source, Iridium Radioisotopes, Male, Monte Carlo, Prostatic Neoplasms, Radiation Dosage, Radioisotopes, Radiotherapy Dosage, Radiotherapy Planning, Selenium Radioisotopes, Tissue composition, Tongue Neoplasms, Ytterbium},
pubstate = {published},
tppubtype = {article}
}
PURPOSE: Several radionuclides with high (60Co, 75Se) and intermediate (169Yb, 153Gd) energies have been investigated as alternatives to 192Ir for high-dose-rate brachytherapy. The purpose of this study was to evaluate the impact of tissue heterogeneities for these five high- to intermediate-energy sources in prostate and head & neck brachytherapy. METHODS AND MATERIALS: Treatment plans were generated for a cohort of prostate (n = 10) and oral tongue (n = 10) patients. Dose calculations were performed using RapidBrachyMCTPS, an in-house Geant4-based Monte Carlo treatment planning system. Treatment plans were simulated using 60Co, 192Ir, 75Se, 169Yb, and 153Gd as the active core of the microSelectron v2 source. Two dose calculation scenarios were presented: (1) dose to water in water (Dw,w), and (2) dose to medium in medium (Dm,m).
RESULTS: Dw,w overestimates planning target volume coverage compared with Dm,m, regardless of photon energy. The average planning target volume D90 reduction was ∼1% for high-energy sources, whereas larger differences were observed for intermediate-energy sources (1%-2% for prostate and 4%-7% for oral tongue). Dose differences were not clinically relevant (textless5%) for soft tissues in general. Going from Dw,w to Dm,m, bone doses were increased two- to three-fold for 169Yb and four- to five-fold for 153Gd, whereas the ratio was close to ∼1 for high-energy sources.
CONCLUSIONS: Dw,w underestimates the dose to bones and, to a lesser extent, overestimates the dose to soft tissues for radionuclides with average energies lower than 192Ir. Further studies regarding bone toxicities are needed before intermediate-energy sources can be adopted in cases where bones are in close vicinity to the tumor.2018
DeCunha, Joseph M.; Enger, Shirin A.
A new delivery system to resolve dosimetric issues in intravascular brachytherapy Journal Article
In: Brachytherapy, vol. 17, no. 3, pp. 634–643, 2018, ISSN: 1873-1449.
Abstract | Links | BibTeX | Tags: Brachytherapy, Catheterization, Catheters, Computer Simulation, Coronary Vessels, Humans, Intravascular, Monte Carlo Method, Physics, Radiation Dosage, Radiometry, Restenosis, Stents, Strontium Radioisotopes
@article{decunha_new_2018,
title = {A new delivery system to resolve dosimetric issues in intravascular brachytherapy},
author = {Joseph M. DeCunha and Shirin A. Enger},
doi = {10.1016/j.brachy.2018.01.012},
issn = {1873-1449},
year = {2018},
date = {2018-06-01},
journal = {Brachytherapy},
volume = {17},
number = {3},
pages = {634--643},
abstract = {PURPOSE: Renewed interest is being expressed in intravascular brachytherapy (IVBT). A number of unresolved issues exist in the discipline. Providing a homogeneous and adequate dose to the target remains difficult in IVBT. The guidewire that delivers the device to the target, arterial plaques, and stent struts are all known to reduce the dose delivered to target. The viability and efficacy of a proposed IVBT delivery system designed to resolve the issue of guidewire attenuation is evaluated and compared to that of a popular and commercially available IVBT device.
METHODS AND MATERIALS: Monte Carlo simulations are conducted to determine distributions of absorbed dose around an existing and proposed IVBT delivery system.
RESULTS: For the Novoste Beta-Cath 3.5F (TeamBest®), dose in water varies by 10% as a function of angle in the plane perpendicular to the delivery catheter due to off-centering of seeds in the catheter. Dose is reduced by 52% behind a stainless steel guidewire and 64% behind a guidewire, arterial plaque, and stent strut for the Novoste Beta-Cath 3.5F. Dose is not perturbed by the presence of a guidewire for the proposed device and is reduced by 46% by an arterial plaque and stent strut.
CONCLUSIONS: Dose attenuation by guidewire is likely the single greatest source of dose attenuation in IVBT in terms of absolute dose reduction and is greater than previously reported for the Novoste Beta-Cath 3.5F. The Novoste Beta-Cath 3.5F delivers an inhomogeneous dose to target. A delivery system is proposed, which resolves the issue of guidewire attenuation in IVBT and should reduce treatment times.},
keywords = {Brachytherapy, Catheterization, Catheters, Computer Simulation, Coronary Vessels, Humans, Intravascular, Monte Carlo Method, Physics, Radiation Dosage, Radiometry, Restenosis, Stents, Strontium Radioisotopes},
pubstate = {published},
tppubtype = {article}
}
PURPOSE: Renewed interest is being expressed in intravascular brachytherapy (IVBT). A number of unresolved issues exist in the discipline. Providing a homogeneous and adequate dose to the target remains difficult in IVBT. The guidewire that delivers the device to the target, arterial plaques, and stent struts are all known to reduce the dose delivered to target. The viability and efficacy of a proposed IVBT delivery system designed to resolve the issue of guidewire attenuation is evaluated and compared to that of a popular and commercially available IVBT device.
METHODS AND MATERIALS: Monte Carlo simulations are conducted to determine distributions of absorbed dose around an existing and proposed IVBT delivery system.
RESULTS: For the Novoste Beta-Cath 3.5F (TeamBest®), dose in water varies by 10% as a function of angle in the plane perpendicular to the delivery catheter due to off-centering of seeds in the catheter. Dose is reduced by 52% behind a stainless steel guidewire and 64% behind a guidewire, arterial plaque, and stent strut for the Novoste Beta-Cath 3.5F. Dose is not perturbed by the presence of a guidewire for the proposed device and is reduced by 46% by an arterial plaque and stent strut.
CONCLUSIONS: Dose attenuation by guidewire is likely the single greatest source of dose attenuation in IVBT in terms of absolute dose reduction and is greater than previously reported for the Novoste Beta-Cath 3.5F. The Novoste Beta-Cath 3.5F delivers an inhomogeneous dose to target. A delivery system is proposed, which resolves the issue of guidewire attenuation in IVBT and should reduce treatment times.2012
Enger, Shirin A.; Ahnesjö, Anders; Verhaegen, Frank; Beaulieu, Luc
Dose to tissue medium or water cavities as surrogate for the dose to cell nuclei at brachytherapy photon energies Journal Article
In: Physics in Medicine and Biology, vol. 57, no. 14, pp. 4489–4500, 2012, ISSN: 1361-6560.
Abstract | Links | BibTeX | Tags: Brachytherapy, Cell Line, Cell Nucleus, Humans, Monte Carlo Method, Photons, Radiation Dosage, Radiotherapy Dosage, Water
@article{enger_dose_2012,
title = {Dose to tissue medium or water cavities as surrogate for the dose to cell nuclei at brachytherapy photon energies},
author = {Shirin A. Enger and Anders Ahnesjö and Frank Verhaegen and Luc Beaulieu},
doi = {10.1088/0031-9155/57/14/4489},
issn = {1361-6560},
year = {2012},
date = {2012-07-01},
journal = {Physics in Medicine and Biology},
volume = {57},
number = {14},
pages = {4489--4500},
abstract = {It has been suggested that modern dose calculation algorithms should be able to report absorbed dose both as dose to the local medium, D(m,m,) and as dose to a water cavity embedded in the medium, D(w,m), using conversion factors from cavity theory. Assuming that the cell nucleus with its DNA content is the most important target for biological response, the aim of this study is to investigate, by means of Monte Carlo (MC) simulations, the relationship of the dose to a cell nucleus in a medium, D(n,m,) to D(m,m) and D(w,m), for different combinations of cell nucleus compositions and tissue media for different photon energies used in brachytherapy. As D(n,m) is very impractical to calculate directly for routine treatment planning, while D(m,m) and D(w,m) are much easier to obtain, the questions arise which one of these quantities is the best surrogate for D(n,m) and which cavity theory assumptions should one use for its estimate. The Geant4.9.4 MC code was used to calculate D(m,m,) D(w,m) and D(n,m) for photon energies from 20 (representing the lower energy end of brachytherapy for ¹⁰³Pd or ¹²⁵I) to 300 keV (close to the mean energy of (¹⁹²Ir) and for the tissue media adipose, breast, prostate and muscle. To simulate the cell and its nucleus, concentric spherical cavities were placed inside a cubic phantom (10 × 10 × 10 mm³). The diameter of the simulated nuclei was set to 14 µm. For each tissue medium, three different setups were simulated; (a) D(n,m) was calculated with nuclei embedded in tissues (MC-D(n,m)). Four different published elemental compositions of cell nuclei were used. (b) D(w,m) was calculated with MC (MC-D(w,m)) and compared with large cavity theory calculated D(w,m) (LCT-D(w,m)), and small cavity theory calculated D(w,m) (SCT-D(w,m)). (c) D(m,m) was calculated with MC (MC-D(m,m)). MC-D(w,m) is a good substitute for MC-D(n,m) for all photon energies and for all simulated nucleus compositions and tissue types. SCT-D(w,m) can be used for most energies in brachytherapy, while LCT-D(w,m) should only be considered for source spectra well below 50 keV, since contributions to the absorbed dose inside the nucleus to a large degree stem from electrons released in the surrounding medium. MC-D(m,m) is not an appropriate substitute for MC-D(n,m) for the lowest photon energies for adipose and breast tissues. The ratio of MC-D(m,m) to MC-D(n,m) for adipose and breast tissue deviates from unity by 34% and 15% respectively for the lowest photon energy (20 keV), whereas the ratio is close to unity for higher energies. For prostate and muscle tissue MC-D(m,m) is a good substitute for MC-D(n,m). However, for all photon energies and tissue types the nucleus composition with the highest hydrogen content behaves differently than other compositions. Elemental compositions of the tissue and nuclei affect considerably the absorbed dose to the cell nuclei for brachytherapy sources, in particular those at the low-energy end of the spectrum. Thus, there is a need for more accurate data for the elemental compositions of tumours and healthy cells. For the nucleus compositions and tissue types investigated, MC-D(w,m) is a good substitute to MC-D(n,m) for all simulated photon energies. Whether other studied surrogates are good approximations to MC-D(n,m) depends on the target size, target composition, composition of the surrounding tissue and photon energy.},
keywords = {Brachytherapy, Cell Line, Cell Nucleus, Humans, Monte Carlo Method, Photons, Radiation Dosage, Radiotherapy Dosage, Water},
pubstate = {published},
tppubtype = {article}
}
It has been suggested that modern dose calculation algorithms should be able to report absorbed dose both as dose to the local medium, D(m,m,) and as dose to a water cavity embedded in the medium, D(w,m), using conversion factors from cavity theory. Assuming that the cell nucleus with its DNA content is the most important target for biological response, the aim of this study is to investigate, by means of Monte Carlo (MC) simulations, the relationship of the dose to a cell nucleus in a medium, D(n,m,) to D(m,m) and D(w,m), for different combinations of cell nucleus compositions and tissue media for different photon energies used in brachytherapy. As D(n,m) is very impractical to calculate directly for routine treatment planning, while D(m,m) and D(w,m) are much easier to obtain, the questions arise which one of these quantities is the best surrogate for D(n,m) and which cavity theory assumptions should one use for its estimate. The Geant4.9.4 MC code was used to calculate D(m,m,) D(w,m) and D(n,m) for photon energies from 20 (representing the lower energy end of brachytherapy for ¹⁰³Pd or ¹²⁵I) to 300 keV (close to the mean energy of (¹⁹²Ir) and for the tissue media adipose, breast, prostate and muscle. To simulate the cell and its nucleus, concentric spherical cavities were placed inside a cubic phantom (10 × 10 × 10 mm³). The diameter of the simulated nuclei was set to 14 µm. For each tissue medium, three different setups were simulated; (a) D(n,m) was calculated with nuclei embedded in tissues (MC-D(n,m)). Four different published elemental compositions of cell nuclei were used. (b) D(w,m) was calculated with MC (MC-D(w,m)) and compared with large cavity theory calculated D(w,m) (LCT-D(w,m)), and small cavity theory calculated D(w,m) (SCT-D(w,m)). (c) D(m,m) was calculated with MC (MC-D(m,m)). MC-D(w,m) is a good substitute for MC-D(n,m) for all photon energies and for all simulated nucleus compositions and tissue types. SCT-D(w,m) can be used for most energies in brachytherapy, while LCT-D(w,m) should only be considered for source spectra well below 50 keV, since contributions to the absorbed dose inside the nucleus to a large degree stem from electrons released in the surrounding medium. MC-D(m,m) is not an appropriate substitute for MC-D(n,m) for the lowest photon energies for adipose and breast tissues. The ratio of MC-D(m,m) to MC-D(n,m) for adipose and breast tissue deviates from unity by 34% and 15% respectively for the lowest photon energy (20 keV), whereas the ratio is close to unity for higher energies. For prostate and muscle tissue MC-D(m,m) is a good substitute for MC-D(n,m). However, for all photon energies and tissue types the nucleus composition with the highest hydrogen content behaves differently than other compositions. Elemental compositions of the tissue and nuclei affect considerably the absorbed dose to the cell nuclei for brachytherapy sources, in particular those at the low-energy end of the spectrum. Thus, there is a need for more accurate data for the elemental compositions of tumours and healthy cells. For the nucleus compositions and tissue types investigated, MC-D(w,m) is a good substitute to MC-D(n,m) for all simulated photon energies. Whether other studied surrogates are good approximations to MC-D(n,m) depends on the target size, target composition, composition of the surrounding tissue and photon energy.
Journal Articles
2020
Antaki, Majd; Deufel, Christopher L; Enger, Shirin A.
Fast mixed integer optimization (FMIO) for high dose rate brachytherapy Journal Article
In: Physics in Medicine and Biology, vol. 65, no. 21, pp. 215005, 2020, ISSN: 1361-6560.
Abstract | Links | BibTeX | Tags: Algorithms, Brachytherapy, Computer-Assisted, Humans, Linear Models, Male, Monte Carlo Method, Prostatic Neoplasms, Radiation Dosage, Radiotherapy Dosage, Radiotherapy Planning, Software, Time Factors
@article{antaki_fast_2020,
title = {Fast mixed integer optimization (FMIO) for high dose rate brachytherapy},
author = {Majd Antaki and Christopher L Deufel and Shirin A. Enger},
doi = {10.1088/1361-6560/aba317},
issn = {1361-6560},
year = {2020},
date = {2020-12-01},
journal = {Physics in Medicine and Biology},
volume = {65},
number = {21},
pages = {215005},
abstract = {The purpose of this work was to develop an efficient quadratic mixed integer programming algorithm for high dose rate (HDR) brachytherapy treatment planning problems and integrate the algorithm into an open-source Monte Carlo based treatment planning software, RapidBrachyMCTPS. The mixed-integer algorithm yields a globally optimum solution to the dose volume histogram (DVH) based problem and, unlike other methods, is not susceptible to local minimum trapping. A hybrid linear-quadratic penalty model coupled to a mixed integer programming model was used to optimize treatment plans for 10 prostate cancer patients. Dose distributions for each dwell position were calculated with RapidBrachyMCTPS with type A uncertainties less than 0.2% in voxels within the planning target volume (PTV). The optimization process was divided into two parts. First, the data was preprocessed, in which the problem size was reduced by eliminating voxels that had negligible impact on the solution (e.g. far from the dwell position). Second, the best combination of dwell times to obtain a plan with the highest score was found. The dwell positions and dose volume constraints were used as input to a commercial mixed integer optimizer (Gurobi Optimization, Inc.). A penalty-based criterion was adopted for the scoring. The voxel-reduction technique successfully reduced the problem size by an average of 91%, without loss of quality. The preprocessing of the optimization process required on average 4 s and solving for the global maximum required on average 33 s. The total optimization time averaged 37 s, which is a substantial improvement over the ∼15 min optimization time reported in published literature. The plan quality was evaluated by evaluating dose volume metrics, including PTV D90, rectum and bladder D1cc and urethra D0.1cc. In conclusion, fast mixed integer optimization is an order of magnitude faster than current mixed-integer approaches for solving HDR brachytherapy treatment planning problems with DVH based metrics.},
keywords = {Algorithms, Brachytherapy, Computer-Assisted, Humans, Linear Models, Male, Monte Carlo Method, Prostatic Neoplasms, Radiation Dosage, Radiotherapy Dosage, Radiotherapy Planning, Software, Time Factors},
pubstate = {published},
tppubtype = {article}
}
The purpose of this work was to develop an efficient quadratic mixed integer programming algorithm for high dose rate (HDR) brachytherapy treatment planning problems and integrate the algorithm into an open-source Monte Carlo based treatment planning software, RapidBrachyMCTPS. The mixed-integer algorithm yields a globally optimum solution to the dose volume histogram (DVH) based problem and, unlike other methods, is not susceptible to local minimum trapping. A hybrid linear-quadratic penalty model coupled to a mixed integer programming model was used to optimize treatment plans for 10 prostate cancer patients. Dose distributions for each dwell position were calculated with RapidBrachyMCTPS with type A uncertainties less than 0.2% in voxels within the planning target volume (PTV). The optimization process was divided into two parts. First, the data was preprocessed, in which the problem size was reduced by eliminating voxels that had negligible impact on the solution (e.g. far from the dwell position). Second, the best combination of dwell times to obtain a plan with the highest score was found. The dwell positions and dose volume constraints were used as input to a commercial mixed integer optimizer (Gurobi Optimization, Inc.). A penalty-based criterion was adopted for the scoring. The voxel-reduction technique successfully reduced the problem size by an average of 91%, without loss of quality. The preprocessing of the optimization process required on average 4 s and solving for the global maximum required on average 33 s. The total optimization time averaged 37 s, which is a substantial improvement over the ∼15 min optimization time reported in published literature. The plan quality was evaluated by evaluating dose volume metrics, including PTV D90, rectum and bladder D1cc and urethra D0.1cc. In conclusion, fast mixed integer optimization is an order of magnitude faster than current mixed-integer approaches for solving HDR brachytherapy treatment planning problems with DVH based metrics.
Famulari, Gabriel; Alfieri, Joanne; Duclos, Marie; Vuong, Té; Enger, Shirin A.
Can intermediate-energy sources lead to elevated bone doses for prostate and head & neck high-dose-rate brachytherapy? Journal Article
In: Brachytherapy, vol. 19, no. 2, pp. 255–263, 2020, ISSN: 1873-1449.
Abstract | Links | BibTeX | Tags: Bone and Bones, Brachytherapy, Cobalt Radioisotopes, Computer Simulation, Computer-Assisted, Dose calculation, Gadolinium, Humans, Intermediate-energy source, Iridium Radioisotopes, Male, Monte Carlo, Prostatic Neoplasms, Radiation Dosage, Radioisotopes, Radiotherapy Dosage, Radiotherapy Planning, Selenium Radioisotopes, Tissue composition, Tongue Neoplasms, Ytterbium
@article{famulari_can_2020,
title = {Can intermediate-energy sources lead to elevated bone doses for prostate and head & neck high-dose-rate brachytherapy?},
author = {Gabriel Famulari and Joanne Alfieri and Marie Duclos and Té Vuong and Shirin A. Enger},
doi = {10.1016/j.brachy.2019.12.004},
issn = {1873-1449},
year = {2020},
date = {2020-04-01},
journal = {Brachytherapy},
volume = {19},
number = {2},
pages = {255--263},
abstract = {PURPOSE: Several radionuclides with high (60Co, 75Se) and intermediate (169Yb, 153Gd) energies have been investigated as alternatives to 192Ir for high-dose-rate brachytherapy. The purpose of this study was to evaluate the impact of tissue heterogeneities for these five high- to intermediate-energy sources in prostate and head & neck brachytherapy. METHODS AND MATERIALS: Treatment plans were generated for a cohort of prostate (n = 10) and oral tongue (n = 10) patients. Dose calculations were performed using RapidBrachyMCTPS, an in-house Geant4-based Monte Carlo treatment planning system. Treatment plans were simulated using 60Co, 192Ir, 75Se, 169Yb, and 153Gd as the active core of the microSelectron v2 source. Two dose calculation scenarios were presented: (1) dose to water in water (Dw,w), and (2) dose to medium in medium (Dm,m).
RESULTS: Dw,w overestimates planning target volume coverage compared with Dm,m, regardless of photon energy. The average planning target volume D90 reduction was ∼1% for high-energy sources, whereas larger differences were observed for intermediate-energy sources (1%-2% for prostate and 4%-7% for oral tongue). Dose differences were not clinically relevant (textless5%) for soft tissues in general. Going from Dw,w to Dm,m, bone doses were increased two- to three-fold for 169Yb and four- to five-fold for 153Gd, whereas the ratio was close to ∼1 for high-energy sources.
CONCLUSIONS: Dw,w underestimates the dose to bones and, to a lesser extent, overestimates the dose to soft tissues for radionuclides with average energies lower than 192Ir. Further studies regarding bone toxicities are needed before intermediate-energy sources can be adopted in cases where bones are in close vicinity to the tumor.},
keywords = {Bone and Bones, Brachytherapy, Cobalt Radioisotopes, Computer Simulation, Computer-Assisted, Dose calculation, Gadolinium, Humans, Intermediate-energy source, Iridium Radioisotopes, Male, Monte Carlo, Prostatic Neoplasms, Radiation Dosage, Radioisotopes, Radiotherapy Dosage, Radiotherapy Planning, Selenium Radioisotopes, Tissue composition, Tongue Neoplasms, Ytterbium},
pubstate = {published},
tppubtype = {article}
}
PURPOSE: Several radionuclides with high (60Co, 75Se) and intermediate (169Yb, 153Gd) energies have been investigated as alternatives to 192Ir for high-dose-rate brachytherapy. The purpose of this study was to evaluate the impact of tissue heterogeneities for these five high- to intermediate-energy sources in prostate and head & neck brachytherapy. METHODS AND MATERIALS: Treatment plans were generated for a cohort of prostate (n = 10) and oral tongue (n = 10) patients. Dose calculations were performed using RapidBrachyMCTPS, an in-house Geant4-based Monte Carlo treatment planning system. Treatment plans were simulated using 60Co, 192Ir, 75Se, 169Yb, and 153Gd as the active core of the microSelectron v2 source. Two dose calculation scenarios were presented: (1) dose to water in water (Dw,w), and (2) dose to medium in medium (Dm,m).
RESULTS: Dw,w overestimates planning target volume coverage compared with Dm,m, regardless of photon energy. The average planning target volume D90 reduction was ∼1% for high-energy sources, whereas larger differences were observed for intermediate-energy sources (1%-2% for prostate and 4%-7% for oral tongue). Dose differences were not clinically relevant (textless5%) for soft tissues in general. Going from Dw,w to Dm,m, bone doses were increased two- to three-fold for 169Yb and four- to five-fold for 153Gd, whereas the ratio was close to ∼1 for high-energy sources.
CONCLUSIONS: Dw,w underestimates the dose to bones and, to a lesser extent, overestimates the dose to soft tissues for radionuclides with average energies lower than 192Ir. Further studies regarding bone toxicities are needed before intermediate-energy sources can be adopted in cases where bones are in close vicinity to the tumor.
RESULTS: Dw,w overestimates planning target volume coverage compared with Dm,m, regardless of photon energy. The average planning target volume D90 reduction was ∼1% for high-energy sources, whereas larger differences were observed for intermediate-energy sources (1%-2% for prostate and 4%-7% for oral tongue). Dose differences were not clinically relevant (textless5%) for soft tissues in general. Going from Dw,w to Dm,m, bone doses were increased two- to three-fold for 169Yb and four- to five-fold for 153Gd, whereas the ratio was close to ∼1 for high-energy sources.
CONCLUSIONS: Dw,w underestimates the dose to bones and, to a lesser extent, overestimates the dose to soft tissues for radionuclides with average energies lower than 192Ir. Further studies regarding bone toxicities are needed before intermediate-energy sources can be adopted in cases where bones are in close vicinity to the tumor.
2018
DeCunha, Joseph M.; Enger, Shirin A.
A new delivery system to resolve dosimetric issues in intravascular brachytherapy Journal Article
In: Brachytherapy, vol. 17, no. 3, pp. 634–643, 2018, ISSN: 1873-1449.
Abstract | Links | BibTeX | Tags: Brachytherapy, Catheterization, Catheters, Computer Simulation, Coronary Vessels, Humans, Intravascular, Monte Carlo Method, Physics, Radiation Dosage, Radiometry, Restenosis, Stents, Strontium Radioisotopes
@article{decunha_new_2018,
title = {A new delivery system to resolve dosimetric issues in intravascular brachytherapy},
author = {Joseph M. DeCunha and Shirin A. Enger},
doi = {10.1016/j.brachy.2018.01.012},
issn = {1873-1449},
year = {2018},
date = {2018-06-01},
journal = {Brachytherapy},
volume = {17},
number = {3},
pages = {634--643},
abstract = {PURPOSE: Renewed interest is being expressed in intravascular brachytherapy (IVBT). A number of unresolved issues exist in the discipline. Providing a homogeneous and adequate dose to the target remains difficult in IVBT. The guidewire that delivers the device to the target, arterial plaques, and stent struts are all known to reduce the dose delivered to target. The viability and efficacy of a proposed IVBT delivery system designed to resolve the issue of guidewire attenuation is evaluated and compared to that of a popular and commercially available IVBT device.
METHODS AND MATERIALS: Monte Carlo simulations are conducted to determine distributions of absorbed dose around an existing and proposed IVBT delivery system.
RESULTS: For the Novoste Beta-Cath 3.5F (TeamBest®), dose in water varies by 10% as a function of angle in the plane perpendicular to the delivery catheter due to off-centering of seeds in the catheter. Dose is reduced by 52% behind a stainless steel guidewire and 64% behind a guidewire, arterial plaque, and stent strut for the Novoste Beta-Cath 3.5F. Dose is not perturbed by the presence of a guidewire for the proposed device and is reduced by 46% by an arterial plaque and stent strut.
CONCLUSIONS: Dose attenuation by guidewire is likely the single greatest source of dose attenuation in IVBT in terms of absolute dose reduction and is greater than previously reported for the Novoste Beta-Cath 3.5F. The Novoste Beta-Cath 3.5F delivers an inhomogeneous dose to target. A delivery system is proposed, which resolves the issue of guidewire attenuation in IVBT and should reduce treatment times.},
keywords = {Brachytherapy, Catheterization, Catheters, Computer Simulation, Coronary Vessels, Humans, Intravascular, Monte Carlo Method, Physics, Radiation Dosage, Radiometry, Restenosis, Stents, Strontium Radioisotopes},
pubstate = {published},
tppubtype = {article}
}
PURPOSE: Renewed interest is being expressed in intravascular brachytherapy (IVBT). A number of unresolved issues exist in the discipline. Providing a homogeneous and adequate dose to the target remains difficult in IVBT. The guidewire that delivers the device to the target, arterial plaques, and stent struts are all known to reduce the dose delivered to target. The viability and efficacy of a proposed IVBT delivery system designed to resolve the issue of guidewire attenuation is evaluated and compared to that of a popular and commercially available IVBT device.
METHODS AND MATERIALS: Monte Carlo simulations are conducted to determine distributions of absorbed dose around an existing and proposed IVBT delivery system.
RESULTS: For the Novoste Beta-Cath 3.5F (TeamBest®), dose in water varies by 10% as a function of angle in the plane perpendicular to the delivery catheter due to off-centering of seeds in the catheter. Dose is reduced by 52% behind a stainless steel guidewire and 64% behind a guidewire, arterial plaque, and stent strut for the Novoste Beta-Cath 3.5F. Dose is not perturbed by the presence of a guidewire for the proposed device and is reduced by 46% by an arterial plaque and stent strut.
CONCLUSIONS: Dose attenuation by guidewire is likely the single greatest source of dose attenuation in IVBT in terms of absolute dose reduction and is greater than previously reported for the Novoste Beta-Cath 3.5F. The Novoste Beta-Cath 3.5F delivers an inhomogeneous dose to target. A delivery system is proposed, which resolves the issue of guidewire attenuation in IVBT and should reduce treatment times.
METHODS AND MATERIALS: Monte Carlo simulations are conducted to determine distributions of absorbed dose around an existing and proposed IVBT delivery system.
RESULTS: For the Novoste Beta-Cath 3.5F (TeamBest®), dose in water varies by 10% as a function of angle in the plane perpendicular to the delivery catheter due to off-centering of seeds in the catheter. Dose is reduced by 52% behind a stainless steel guidewire and 64% behind a guidewire, arterial plaque, and stent strut for the Novoste Beta-Cath 3.5F. Dose is not perturbed by the presence of a guidewire for the proposed device and is reduced by 46% by an arterial plaque and stent strut.
CONCLUSIONS: Dose attenuation by guidewire is likely the single greatest source of dose attenuation in IVBT in terms of absolute dose reduction and is greater than previously reported for the Novoste Beta-Cath 3.5F. The Novoste Beta-Cath 3.5F delivers an inhomogeneous dose to target. A delivery system is proposed, which resolves the issue of guidewire attenuation in IVBT and should reduce treatment times.
2012
Enger, Shirin A.; Ahnesjö, Anders; Verhaegen, Frank; Beaulieu, Luc
Dose to tissue medium or water cavities as surrogate for the dose to cell nuclei at brachytherapy photon energies Journal Article
In: Physics in Medicine and Biology, vol. 57, no. 14, pp. 4489–4500, 2012, ISSN: 1361-6560.
Abstract | Links | BibTeX | Tags: Brachytherapy, Cell Line, Cell Nucleus, Humans, Monte Carlo Method, Photons, Radiation Dosage, Radiotherapy Dosage, Water
@article{enger_dose_2012,
title = {Dose to tissue medium or water cavities as surrogate for the dose to cell nuclei at brachytherapy photon energies},
author = {Shirin A. Enger and Anders Ahnesjö and Frank Verhaegen and Luc Beaulieu},
doi = {10.1088/0031-9155/57/14/4489},
issn = {1361-6560},
year = {2012},
date = {2012-07-01},
journal = {Physics in Medicine and Biology},
volume = {57},
number = {14},
pages = {4489--4500},
abstract = {It has been suggested that modern dose calculation algorithms should be able to report absorbed dose both as dose to the local medium, D(m,m,) and as dose to a water cavity embedded in the medium, D(w,m), using conversion factors from cavity theory. Assuming that the cell nucleus with its DNA content is the most important target for biological response, the aim of this study is to investigate, by means of Monte Carlo (MC) simulations, the relationship of the dose to a cell nucleus in a medium, D(n,m,) to D(m,m) and D(w,m), for different combinations of cell nucleus compositions and tissue media for different photon energies used in brachytherapy. As D(n,m) is very impractical to calculate directly for routine treatment planning, while D(m,m) and D(w,m) are much easier to obtain, the questions arise which one of these quantities is the best surrogate for D(n,m) and which cavity theory assumptions should one use for its estimate. The Geant4.9.4 MC code was used to calculate D(m,m,) D(w,m) and D(n,m) for photon energies from 20 (representing the lower energy end of brachytherapy for ¹⁰³Pd or ¹²⁵I) to 300 keV (close to the mean energy of (¹⁹²Ir) and for the tissue media adipose, breast, prostate and muscle. To simulate the cell and its nucleus, concentric spherical cavities were placed inside a cubic phantom (10 × 10 × 10 mm³). The diameter of the simulated nuclei was set to 14 µm. For each tissue medium, three different setups were simulated; (a) D(n,m) was calculated with nuclei embedded in tissues (MC-D(n,m)). Four different published elemental compositions of cell nuclei were used. (b) D(w,m) was calculated with MC (MC-D(w,m)) and compared with large cavity theory calculated D(w,m) (LCT-D(w,m)), and small cavity theory calculated D(w,m) (SCT-D(w,m)). (c) D(m,m) was calculated with MC (MC-D(m,m)). MC-D(w,m) is a good substitute for MC-D(n,m) for all photon energies and for all simulated nucleus compositions and tissue types. SCT-D(w,m) can be used for most energies in brachytherapy, while LCT-D(w,m) should only be considered for source spectra well below 50 keV, since contributions to the absorbed dose inside the nucleus to a large degree stem from electrons released in the surrounding medium. MC-D(m,m) is not an appropriate substitute for MC-D(n,m) for the lowest photon energies for adipose and breast tissues. The ratio of MC-D(m,m) to MC-D(n,m) for adipose and breast tissue deviates from unity by 34% and 15% respectively for the lowest photon energy (20 keV), whereas the ratio is close to unity for higher energies. For prostate and muscle tissue MC-D(m,m) is a good substitute for MC-D(n,m). However, for all photon energies and tissue types the nucleus composition with the highest hydrogen content behaves differently than other compositions. Elemental compositions of the tissue and nuclei affect considerably the absorbed dose to the cell nuclei for brachytherapy sources, in particular those at the low-energy end of the spectrum. Thus, there is a need for more accurate data for the elemental compositions of tumours and healthy cells. For the nucleus compositions and tissue types investigated, MC-D(w,m) is a good substitute to MC-D(n,m) for all simulated photon energies. Whether other studied surrogates are good approximations to MC-D(n,m) depends on the target size, target composition, composition of the surrounding tissue and photon energy.},
keywords = {Brachytherapy, Cell Line, Cell Nucleus, Humans, Monte Carlo Method, Photons, Radiation Dosage, Radiotherapy Dosage, Water},
pubstate = {published},
tppubtype = {article}
}
It has been suggested that modern dose calculation algorithms should be able to report absorbed dose both as dose to the local medium, D(m,m,) and as dose to a water cavity embedded in the medium, D(w,m), using conversion factors from cavity theory. Assuming that the cell nucleus with its DNA content is the most important target for biological response, the aim of this study is to investigate, by means of Monte Carlo (MC) simulations, the relationship of the dose to a cell nucleus in a medium, D(n,m,) to D(m,m) and D(w,m), for different combinations of cell nucleus compositions and tissue media for different photon energies used in brachytherapy. As D(n,m) is very impractical to calculate directly for routine treatment planning, while D(m,m) and D(w,m) are much easier to obtain, the questions arise which one of these quantities is the best surrogate for D(n,m) and which cavity theory assumptions should one use for its estimate. The Geant4.9.4 MC code was used to calculate D(m,m,) D(w,m) and D(n,m) for photon energies from 20 (representing the lower energy end of brachytherapy for ¹⁰³Pd or ¹²⁵I) to 300 keV (close to the mean energy of (¹⁹²Ir) and for the tissue media adipose, breast, prostate and muscle. To simulate the cell and its nucleus, concentric spherical cavities were placed inside a cubic phantom (10 × 10 × 10 mm³). The diameter of the simulated nuclei was set to 14 µm. For each tissue medium, three different setups were simulated; (a) D(n,m) was calculated with nuclei embedded in tissues (MC-D(n,m)). Four different published elemental compositions of cell nuclei were used. (b) D(w,m) was calculated with MC (MC-D(w,m)) and compared with large cavity theory calculated D(w,m) (LCT-D(w,m)), and small cavity theory calculated D(w,m) (SCT-D(w,m)). (c) D(m,m) was calculated with MC (MC-D(m,m)). MC-D(w,m) is a good substitute for MC-D(n,m) for all photon energies and for all simulated nucleus compositions and tissue types. SCT-D(w,m) can be used for most energies in brachytherapy, while LCT-D(w,m) should only be considered for source spectra well below 50 keV, since contributions to the absorbed dose inside the nucleus to a large degree stem from electrons released in the surrounding medium. MC-D(m,m) is not an appropriate substitute for MC-D(n,m) for the lowest photon energies for adipose and breast tissues. The ratio of MC-D(m,m) to MC-D(n,m) for adipose and breast tissue deviates from unity by 34% and 15% respectively for the lowest photon energy (20 keV), whereas the ratio is close to unity for higher energies. For prostate and muscle tissue MC-D(m,m) is a good substitute for MC-D(n,m). However, for all photon energies and tissue types the nucleus composition with the highest hydrogen content behaves differently than other compositions. Elemental compositions of the tissue and nuclei affect considerably the absorbed dose to the cell nuclei for brachytherapy sources, in particular those at the low-energy end of the spectrum. Thus, there is a need for more accurate data for the elemental compositions of tumours and healthy cells. For the nucleus compositions and tissue types investigated, MC-D(w,m) is a good substitute to MC-D(n,m) for all simulated photon energies. Whether other studied surrogates are good approximations to MC-D(n,m) depends on the target size, target composition, composition of the surrounding tissue and photon energy.