Todd Bossenberger

Effect of MLC leaf width on the planning and delivery of SMLC IMRT using the CORVUS inverse treatment planning system

This study investigates the influence of multileaf collimator (MLC) leaf width on intensity modulated radiation therapy (IMRT) plans delivered via the segmented multileaf collimator (SMLC) technique. IMRT plans were calculated using the Corvus treatment planning system for three brain, three prostate, and three pancreas cases using leaf widths of 0.5 and 1 cm. Resulting differences in plan quality and complexity are presented here. Plans calculated using a 1 cm leaf width were chosen over the 0.5 cm leaf width plans in seven out of nine cases based on clinical judgment. Conversely, optimization results revealed a superior objective function result for the 0.5 cm leaf width plans in seven out of the nine comparisons. The 1 cm leaf width objective function result was superior only for very large target volumes, indicating that expanding the solution space for plan optimization by using narrower leaves may result in a decreased probability of finding the global minimum. In the remaining cases, we can conclude that we are often not utilizing the objective function as proficiently as possible to meet our clinical goals. There was often no apparent clinically significant difference between the two plans, and in such cases the issue becomes one of plan complexity. A comparison of plan complexity revealed that the average 1 cm leaf width plan required roughly 60% fewer segments and over 40% fewer monitor units than required by 0.5 cm leaf width plans. This allows a significant decrease in whole body dose and total treatment time. For very complex IMRT plans, the treatment delivery time may affect the biologically effective dose. A clinically significant improvement in plan quality from using narrower leaves was evident only in cases with very small target volumes or those with concavities that are small with respect to the MLC leaf width. For the remaining cases investigated in this study, there was no clinical advantage to reducing the MLC leaf width from 1 to 0.5 cm. In such cases, there is no justification for the increased treatment time and whole body dose associated with the narrower MLC leaf width.

Single-fraction spine SBRT end-to-end testing on TomoTherapy, Vero, TrueBeam, and CyberKnife treatment platforms using a novel anthropomorphic phantom

Spine SBRT involves the delivery of very high doses of radiation to targets adjacent to the spinal cord and is most commonly delivered in a single fraction. Highly conformal planning and accurate delivery of such plans is imperative for successful treatment without catastrophic adverse effects. End–to‐end testing is an important practice for evaluating the entire treatment process from simulation through treatment delivery. We performed end‐to‐end testing for a set of representative spine targets planned and delivered using four different treatment planning systems (TPSs) and delivery systems to evaluate the various capabilities of each. An anthropomorphic E2E SBRT phantom was simulated and treated on each system to evaluate agreement between measured and calculated doses. The phantom accepts ion chambers in the thoracic region and radiochromic film in the lumbar region. Four representative targets were developed within each region (thoracic and lumbar) to represent different presentations of spinal metastases and planned according to RTOG 0631 constraints. Plans were created using the TomoTherapy TPS for delivery using the Hi·Art system, the iPlan TPS for delivery using the Vero system, the Eclipse TPS for delivery using the TrueBeam system in both flattened and flattening filter free (FFF), and the MultiPlan TPS for delivery using the CyberKnife system. Delivered doses were measured using a 0.007 cm3 ion chamber in the thoracic region and EBT3 GAFCHROMIC film in the lumbar region. Films were scanned and analyzed using an Epson Expression 10000XL flatbed scanner in conjunction with FilmQAPro2013. All treatment platforms met all dose constraints required by RTOG 0631. Ion chamber measurements in the thoracic targets delivered an overall average difference of 1.5%. Specifically, measurements agreed with the TPS to within 2.2%, 3.2%, 1.4%, 3.1%, and 3.0% for all three measureable cases on TomoTherapy, Vero, TrueBeam (FFF), TrueBeam (flattened), and CyberKnife, respectively. Film measurements for the lumbar targets resulted in average global gamma index passing rates of 100% at 3%/3 mm, 96.9% at 2%/2 mm, and 61.8% at 1%/1 mm, with a 10% minimum threshold for all plans on all platforms. Local gamma analysis was also performed with similar results. While gamma passing rates were consistently accurate across all platforms through 2%/2 mm, treatment beam‐on delivery times varied greatly between each platform with TrueBeam FFF being shortest, averaging 4.4 min, TrueBeam using flattened beam at 9.5 min, TomoTherapy at 30.5 min, Vero at 19 min, and CyberKnife at 46.0 min. In spite of the complexity of the representative targets and their proximity to the spinal cord, all treatment platforms were able to create plans meeting all RTOG 0631 dose constraints and produced exceptional agreement between calculated and measured doses. However, there were differences in the plan characteristics and significant differences in the beam‐on delivery time between platforms. Thus, clinical judgment is required for each particular case to determine most appropriate treatment planning/delivery platform. PACS number: 87.53.Ly

Assessment and Minimization of Contralateral Breast Dose for Conventional and Intensity Modulated Breast Radiotherapy

Breast radiotherapy is associated with an increased risk of contralateral breast cancer (CBC) in women under age 45 at the time of treatment. This risk increases with increasing absorbed dose to the contralateral breast. The use of intensity modulated radiotherapy (IMRT) is expected to substantially reduce the dose to the contralateral breast by eliminating scattered radiation from physical beam modifiers. The absorbed dose to the contralateral breast was measured for 5 common radiotherapy techniques, including paired 15 degrees wedges, lateral 30 degrees wedge only, custom-designed physical compensators, aperture based (field-within-field) IMRT with segments chosen by the planner, and inverse planned IMRT with segments chosen by a leaf sequencing algorithm after dose volume histogram (DVH)-based fluence map optimization. Further reduction in contralateral breast dose through the use of lead shielding was also investigated. While shielding was observed to have the most profound impact on surface dose, the radiotherapy technique proved to be most important in determining internal dose. Paired wedges or compensators result in the highest contralateral breast doses (nearly 10% of the prescription dose on the medial surface), while use of IMRT or removal of the medial wedge results in significantly lower doses. Aperture-based IMRT results in the lowest internal doses, primarily due to the decrease in the number of monitor units required and the associated reduction in leakage dose. The use of aperture-based IMRT reduced the average dose to the contralateral breast by greater than 50% in comparison to wedges or compensators. Combined use of IMRT and 1/8-inch-thick lead shielding reduced the dose to the interior and surface of the contralateral breast by roughly 60% and 85%, respectively. This reduction may warrant the use of IMRT for younger patients who have a statistically significant risk of contralateral breast cancer associated with breast radiotherapy.

Single fraction radiosurgery/stereotactic body radiation therapy (SBRT) for spine metastasis: A dosimetric comparison of multiple delivery platforms

Abstract There are numerous commercial radiotherapy systems capable of delivering single fraction spine radiosurgery/SBRT. We aim to compare the capabilities of several of these systems to deliver this treatment when following standardized criteria from a national protocol. Four distinct target lesions representing various case presentations of spine metastases were contoured in both the thoracic and lumbar spine of an anthropomorphic SBRT phantom. Single fraction radiosurgery/SBRT plans were designed for each target with each of our treatment platforms. Plans were prescribed to 16 Gy in one fraction to cover 90% of the target volume using normal tissue and target constraints from RTOG 0631. We analyzed these plans with priority on the dose to 10% of the partial spinal cord and dose to 0.03 cc of the spinal cord. Each system was able to maintain 90% target coverage while meeting all the constraints of RTOG 0631. On average, CyberKnife was able to achieve the lowest spinal cord doses overall and also generated the sharpest dose falloff as indicated by the Paddick gradient index. Treatment times varied widely depending on the modality utilized. On average, treatment can be delivered faster with Flattening Filter Free RapidArc and Tomotherapy, compared to Vero and Cyberknife. While all systems analyzed were able to meet the dose constraints of RTOG 0631, unique characteristics of individual treatment modalities may guide modality selection. Specifically, certain modalities performed better than the others for specific target shapes and locations, and delivery time varied significantly among the different modalities. These findings could provide guidance in determining which of the available modalities would be preferable for the treatment of spine metastases based on individualized treatment goals.

Dose escalation in prostate cancer using intensity modulated neutron radiotherapy

BACKGROUND AND PURPOSE Initial promising results of 3D conformal neutron radiotherapy (3D-CNRT) were subsequently limited by high normal tissue toxicities. It is now possible to deliver intensity modulated neutron radiotherapy (IMNRT). The present work compares photon IMRT, 3D-CNRT and IMNRT for three prostate patients to quantify the benefits of IMNRT. MATERIALS AND METHODS We compare updated 3D-CNRT plans, IMNRT plans, and conventional IMRT plans by translating neutron DVHs into effective photon DVHs using the dose dependent radiobiological effectiveness (RBE) for each structure. RBE curves are parameterized for a range of normal tissue and prostate tumor values. Generalized equivalent uniform dose (gEUD) and gEUD in 2Gy fractions (gEUD(2)) is calculated for each structure, plan, and parameterization. Rectal sparing and dose to prostate-GTV are compared for 3D-CNRT, IMNRT, and IMRT as a function of normal tissue and prostate RBE. RESULTS The closer the RBE values of prostate tumor and normal tissue, the greater the advantage of IMNRT over 3D-CNRT. The rectal sparing achieved using IMNRT ranged from ∼5% to 13% depending upon the choice of RBE for rectum and the α/β value of prostate tumor. IMNRT may provide a theoretical dose advantage over photon IMRT if the α/β value of prostate is 1.5 and the RBEs of prostate and rectum differ by more than 5%. For higher values of prostate α/β any advantages of IMNRT over IMRT could require that the RBEs of prostate and rectum differ by as much as 20%. CONCLUSIONS IMNRT provides a clear normal tissue sparing advantage over 3D-CNRT. The advantage increases when the RBEs of the target structure and the normal tissue are similar. This RBE translation method could help identify clinical sites where the dose sparing advantages of IMNRT would allow for the exploitation of the radiobiological advantages of high-LET neutron radiotherapy.

Modeling the Agility MLC in the Monaco treatment planning system

We investigate the relationship between the various parameters in the Monaco MLC model and dose calculation accuracy for an Elekta Agility MLC. The vendor‐provided MLC modeling procedure — completed first with external vendor participation and then exclusively in‐house — was used in combination with our own procedures to investigate several sets of MLC modeling parameters to determine their effect on dose distributions and point‐dose measurements. Simple plans provided in the vendor procedure were used to elucidate specific mechanical characteristics of the MLC, while ten complex treatment plans — five IMRT and five VMAT — created using TG‐119‐based structure sets were used to test clinical dosimetric effects of particular parameter choices. EDR2 film was used for the vendor fields to give high spatial resolution, while a combination of MapCHECK and ion chambers were used for the in‐house TG‐119‐based procedures. The vendor‐determined parameter set provided a reasonable starting point for the MLC model and largely delivered acceptable gamma pass rates for clinical plans — including a passing external evaluation using the IROC H&N phantom. However, the vendor model did not provide point‐dose accuracy consistent with that seen in other treatment systems at our center. Through further internal testing it was found that there existed many sets of MLC parameters, often at opposite ends of their allowable ranges, that provided similar dosimetric characteristics and good agreement with planar and point‐dose measurements. In particular, the leaf offset and tip leakage parameters compensated for one another if adjusted in opposite directions, which provided a level curve of acceptable parameter sets across all plans. Interestingly, gamma pass rates of the plans were less dependent upon parameter choices than point‐dose measurements, suggesting that MLC modeling using only gamma evaluation may be generally an insufficient approach. It was also found that exploring all parameters of the very robust MLC model to find the best match to the vendor‐provided QA fields can reduce the pass rates of the TG‐119‐based clinical distributions as compared to simpler models. A wide variety of parameter sets produced MLC models capable of meeting RPC passing criteria for their H&N IMRT phantom. The most accurate models were achievable using a combination of vendor‐provided and in‐house procedures. The potential existed for an over‐modeling of the Agility MLC in an effort to obtain the fine structure of certain quality assurance fields, which led to a reduction in agreement between calculation and measurement of more typical clinical dose distributions. PACS number(s): 87.56.nk, 87.53.Kn, 87.55.km, 87.55.QrWe investigate the relationship between the various parameters in the Monaco MLC model and dose calculation accuracy for an Elekta Agility MLC. The vendor-provided MLC modeling procedure - completed first with external vendor participation and then exclusively in-house - was used in combination with our own procedures to investigate several sets of MLC modeling parameters to determine their effect on dose distributions and point-dose measurements. Simple plans provided in the vendor procedure were used to elucidate specific mechanical characteristics of the MLC, while ten complex treatment plans - five IMRT and five VMAT - created using TG-119-based structure sets were used to test clinical dosimetric effects of particular parameter choices. EDR2 film was used for the vendor fields to give high spatial resolution, while a combination of MapCHECK and ion chambers were used for the in-house TG-119-based procedures. The vendor-determined parameter set provided a reasonable starting point for the MLC model and largely delivered acceptable gamma pass rates for clinical plans - including a passing external evaluation using the IROC H&N phantom. However, the vendor model did not provide point-dose accuracy consistent with that seen in other treatment systems at our center. Through further internal testing it was found that there existed many sets of MLC parameters, often at opposite ends of their allowable ranges, that provided similar dosimetric characteristics and good agreement with planar and point-dose measurements. In particular, the leaf offset and tip leakage parameters compensated for one another if adjusted in opposite directions, which provided a level curve of acceptable parameter sets across all plans. Interestingly, gamma pass rates of the plans were less dependent upon parameter choices than point-dose measurements, suggesting that MLC modeling using only gamma evaluation may be generally an insufficient approach. It was also found that exploring all parameters of the very robust MLC model to find the best match to the vendor-provided QA fields can reduce the pass rates of the TG-119-based clinical distributions as compared to simpler models. A wide variety of parameter sets produced MLC models capable of meeting RPC passing criteria for their H&N IMRT phantom. The most accurate models were achievable using a combination of vendor-provided and in-house procedures. The potential existed for an over-modeling of the Agility MLC in an effort to obtain the fine structure of certain quality assurance fields, which led to a reduction in agreement between calculation and measurement of more typical clinical dose distributions. PACS number(s): 87.56.nk, 87.53.Kn, 87.55.km, 87.55.Qr.

Intensity modulated neutron radiotherapy optimization by photon proxy

PURPOSE Introducing intensity modulation into neutron radiotherapy (IMNRT) planning has the potential to mitigate some normal tissue complications seen in past neutron trials. While the hardware to deliver IMNRT plans has been in use for several years, until recently the IMNRT planning process has been cumbersome and of lower fidelity than conventional photon plans. Our in-house planning system used to calculate neutron therapy plans allows beam weight optimization of forward planned segments, but does not provide inverse optimization capabilities. Commercial treatment planning systems provide inverse optimization capabilities, but currently cannot model our neutron beam. METHODS We have developed a methodology and software suite to make use of the robust optimization in our commercial planning system while still using our in-house planning system to calculate final neutron dose distributions. Optimized multileaf collimator (MLC) leaf positions for segments designed in the commercial system using a 4 MV photon proxy beam are translated into static neutron ports that can be represented within our in-house treatment planning system. The true neutron dose distribution is calculated in the in-house system and then exported back through the MATLAB software into the commercial treatment planning system for evaluation. RESULTS The planning process produces optimized IMNRT plans that reduce dose to normal tissue structures as compared to 3D conformal plans using static MLC apertures. The process involves standard planning techniques using a commercially available treatment planning system, and is not significantly more complex than conventional IMRT planning. Using a photon proxy in a commercial optimization algorithm produces IMNRT plans that are more conformal than those previously designed at our center and take much less time to create. CONCLUSIONS The planning process presented here allows for the optimization of IMNRT plans by a commercial treatment planning optimization algorithm, potentially allowing IMNRT to achieve similar conformality in treatment as photon IMRT. The only remaining requirements for the delivery of very highly modulated neutron treatments are incremental improvements upon already implemented hardware systems that should be readily achievable.

Commissioning of a dedicated commercial Co‐60 total body irradiation unit

Abstract We describe the commissioning of the first dedicated commercial total body irradiation (TBI) unit in clinical operation. The Best Theratronics GammaBeam 500 is a Co‐60 teletherapy unit with extended field size and imaging capabilities. Radiation safety, mechanical and imaging systems, and radiation output are characterized. Beam data collection, calibration, and external dosimetric validation are described. All radiation safety and mechanical tests satisfied relevant requirements and measured dose distributions meet recommendations of American Association of Physicists in Medicine (AAPM) Report #17. At a typical treatment distance, the dose rate in free space per unit source activity using the thick flattening filter is 1.53 × 10−3 cGy*min−1*Ci−1. With a 14,000 Ci source, the resulting dose rate at the midplane of a typical patient is approximately 17 and 30 cGy/min using the thick and thin flattening filters, respectively, using the maximum source to couch distance. The maximum useful field size, defined by the 90% isodose line, at this location is 225 × 78 cm with field flatness within 5% over the central 178 × 73 cm. Measured output agreed with external validation within 0.5%. End‐to‐end testing was performed in a modified Rando phantom. In‐house MATLAB software was developed to calculate patient‐specific dose distributions using DOSXYZnrc, and fabricate custom 3D‐printed forms for creating patient‐specific lung blocks. End‐to‐end OSLD and diode measurements both with and without lung blocks agreed with Monte Carlo calculated doses to within 5% and in‐phantom film measurements validated dose distribution uniformity. Custom lung block transmission measurements agree well with design criteria and provide clinically favorable dose distributions within the lungs. Block placement is easily facilitated using the flat panel imaging system with an exposure time of 0.01 min. In conclusion, a novel dedicated TBI unit has been commissioned and clinically implemented. Its mechanical, dosimetric, and imaging capabilities are suitable to provide state‐of‐the‐art TBI for patients as large as 220 cm in height and 78 cm in width.

MO‐F‐108‐09: Single Fraction Spine SBRT End to End Testing On TomoTherapy and Vero Treatment Platforms Using a Novel Anthropomorphic Phantom

PURPOSE Single fraction spine SBRT involves very high doses delivered to targets adjacent to the spinal cord. Highly conformal planning and accurate delivery of such plans is imperative for successful treatment without catastrophic adverse affects. End to end testing is an important practice for evaluating the entire treatment process from simulation through treatment delivery. We performed end-to-end testing for a set of representative spine SBRT targets planned and delivered using different treatment planning systems (TPSs) and delivery systems. METHODS An anthropomorphic E2E SBRT phantom (IMT, Troy,NY) accepting ion chambers in the thoracic region and film in the lumbar region was simulated and treated to evaluate agreement between measured and calculated doses. Four representative targets were developed within each region (thoracic and lumbar) to represent different presentations of spine metastases and planned according to RTOG 0631 constraints. Plans were created using the TomoTherapy TPS for delivery using the Hi-ART system (Accuray, Sunnyvale,CA) and the iPlan TPS for delivery using the Vero system (BrainLAB, Feldkirchen,Germany). Delivered doses were measured using 0.007cc ion chambers (Standard Imaging, Middleton,WI) in the thoracic region and EBT3 Gafchromic film (Ashland, Wayne,NJ) in the lumbar region. RESULTS Both treatment platforms met all dose constraints required by RTOG 0631. Ion chamber measurements in the 4 thoracic targets delivered on TomoTherapy all agreed within 2.2%, with an average difference of 0.8%. Similar measurements on the Vero system agreed within 3.3%. Film measurements for the 4 lumbar targets resulted in gamma index passing rates over 99.7% at 3%/3mm, 99.2% at 2%/2mm, and 90% at 1%/1mm for all plans on both platforms. CONCLUSION In spite of the complexity of the representative targets and their proximity to the spinal cord, both treatment platforms were able to create plans meeting all RTOG dose constraints and produced exceptional agreement between calculated and measured doses.

Commissioning of intensity modulated neutron radiotherapy (IMNRT)

PURPOSE Intensity modulated neutron radiotherapy (IMNRT) has been developed using inhouse treatment planning and delivery systems at the Karmanos Cancer Center∕Wayne State University Fast Neutron Therapy facility. The process of commissioning IMNRT for clinical use is presented here. Results of commissioning tests are provided including validation measurements using representative patient plans as well as those from the TG-119 test suite. METHODS IMNRT plans were created using the Varian Eclipse optimization algorithm and an inhouse planning system for calculation of neutron dose distributions. Tissue equivalent ionization chambers and an ionization chamber array were used for point dose and planar dose distribution comparisons with calculated values. Validation plans were delivered to water and virtual water phantoms using TG-119 measurement points and evaluation techniques. Photon and neutron doses were evaluated both inside and outside the target volume for a typical IMNRT plan to determine effects of intensity modulation on the photon dose component. Monitor unit linearity and effects of beam current and gantry angle on output were investigated, and an independent validation of neutron dosimetry was obtained. RESULTS While IMNRT plan quality is superior to conventional fast neutron therapy plans for clinical sites such as prostate and head and neck, it is inferior to photon IMRT for most TG-119 planning goals, particularly for complex cases. This results significantly from current limitations on the number of segments. Measured and calculated doses for 11 representative plans (six prostate∕five head and neck) agreed to within -0.8 ± 1.4% and 5.0 ± 6.0% within and outside the target, respectively. Nearly all (22∕24) ion chamber point measurements in the two phantom arrangements were within the respective confidence intervals for the quantity [(measured-planned)∕prescription dose] derived in TG-119. Mean differences for all measurements were 0.5% (max = 7.0%) and 1.4% (max = 4.1%) in water and virtual water, respectively. The mean gamma pass rate for all cases was 92.8% (min = 88.6%). These pass rates are lower than typically achieved with photon IMRT, warranting development of a planar dosimetry system designed specifically for IMNRT and∕or the improvement of neutron beam modeling in the penumbral region. The fractional photon dose component did not change significantly in a typical IMNRT plan versus a conventional fast neutron therapy plan, and IMNRT delivery is not expected to significantly alter the RBE. All other commissioning results were considered satisfactory for clinical implementation of IMNRT, including the external neutron dose validation, which agreed with the predicted neutron dose to within 1%. CONCLUSIONS IMNRT has been successfully commissioned for clinical use. While current plan quality is inferior to photon IMRT, it is superior to conventional fast neutron therapy. Ion chamber validation results for IMNRT commissioning are also comparable to those typically achieved with photon IMRT. Gamma pass rates for planar dose distributions are lower than typically observed for photon IMRT but may be improved with improved planar dosimetry equipment and beam modeling techniques. In the meantime, patient-specific quality assurance measurements should rely more heavily on point dose measurements with tissue equivalent ionization chambers. No significant technical impediments are anticipated in the clinical implementation of IMNRT as described here.