Image-guided radiation therapy using cone-beam computed tomography (CBCT) is becoming routine practice in modern radiation therapy. The purpose of this work was to develop an imaging QA program for CT and CBCT units in our department, based on the American College of Radiology (ACR) CT accreditation phantom. The phantom has four testing modules, permitting one to test CT number accuracy, slice width, low contrast resolution, image uniformity, in-plane distance accuracy, and high-contrast resolution reproducibly with suggested window/levels for image analysis. Additional tests for contrast-to-noise ratio (CNR) and noise were added using the polyethylene and acrylic plugs. Baseline values were obtained from CT simulator images acquired on a Phillips Brilliance Big Bore CT simulator and CBCT images acquired on three Varian CBCTs for the imaging protocols most used clinically. Images were then acquired quarterly over a period of two years. Images were exported via DICOM and analyzed manually using OsiriX. Baseline values were used to ensure that image quality remained consistent quarterly, and baselines were reset at any major maintenance or recalibration. Analysis of CT simulator images showed that image quality was within ACR guidelines for all tested scanning protocols. All three CBCT systems were unable to distinguish the low-contrast resolution plugs and had the same high-contrast resolution over all imaging protocols. Analysis of CBCT results over time determined a range of values that could be used to establish quantitative tolerance levels for image quality deterioration. While appropriate for the helical CT, the ACR phantom and guidelines could be modified to be more useful in evaluating CBCT systems. In addition, the observed values for the CT simulator were well within ACR tolerances.

Purpose: To compare the delivery accuracy of helical tomotherapy for stereotactic radiosurgery using two treatment planning system (TPS) versions, v4.0 and the recent upgrade that includes increased sampling in the dose calculation algorithm v4.2. Methods: Two spherical targets of diameter 6 mm and 10 mm were contoured on the CT scan of a Lucite phantom. Three sets of plans (targets positioned at the machine isocenter and 10 cm to both anterior and lateral to the machine isocenter) were generated, with both TPS v4.0 and v4.2 using Fine dose grid resolution of 2×2×1 mm3. Radiochromic film was used to measure the dose profiles and two ionization chambers (A1SL and MicroLion) were used to measure the point dose. Results: The agreement between delivered and calculated dose was inferior for v4.0 compared to v4.2. For both planning versions, the measured dose for the targets placed at the isocenter was higher than the planned dose whereas for off‐axis targets, v4.2 showed higher measured dose and v4.0 showed lower measured dose compared to the planned dose. For v4.0, the percentage dose difference varies 13%to 24%(depending upon the detector) when the targets were placed at 10 cm away from the machine isocenter instead of at the isocenter, whereas for v4.2 this range was 3% to 6%. Conclusion: The agreement between delivered and calculated dose depends upon the position of the target inside the treatment bore. The higher dose delivered at isocenter with both planning versions could be due to the relatively large size of smallest dose grid available at tomotherapy TPS. A further investigation to determine the cause of higher dose at isocenter is required and is ongoing.

This study includes planning and delivery comparison of three stereotactic radiosurgery techniques : Helical Tomotherapy (HT), circular collimator‐based Linear‐accelerator and robotic‐radiosurgery. Plans were generated for two spherical targets of diameter 6 mm and 10 mm contoured at the center of a Lucite phantom, using similar planning constrains. Planning comparison showed that average conformality (0–1best) for Linear‐accelerator, robotic‐radiosurgery and HT was 1.43, 1.24, and 1.77 and gradient index (less is better) was 2.72, 4.50 and 13.56 respectively. For delivery comparison, plans were delivered to radiochromic film and measured dose was compared with the planned dose. For Linear‐accelerator and robotic‐radiosurgery more than 99% pixels‐passing a gamma criteria of 3% dose difference and 1 mm distance to agreement where as for HT this value was as low as 40% for off‐axis targets. Further investigation of the delivery accuracy as a function of the location of the target with in the bore was initiated using small volume A1SL (0.057 cm3) and MicroLion liquid ion chamber (0.0017 cm3). Point dose measurements for targets located at the center and 10 cm away from the center of the bore showed that delivered dose varied by more than 15% for targets placed away from the center of the bore as opposed to at the center. In conclusion, Linear‐accelerator and the robotic‐radiosurgery techniques showed preferable gradient and conformality. For HT, point dose measurements were significantly lower than predicted by the TPS when the target was positioned away from the isocenter, while they were found to be higher at isocenter.

The purpose of this study was to compare delivery accuracy of helical tomotherapy (HT) to linear accelerator (equipped with high definition multileaf collimator) and robotic based radiosurgery for small targets. In addition, delivery accuracy as a function of bore location was investigated for HT. Plans were generated for small spherical targets (6 mm and 10 mm diameter) on a Lucite phantom for the three delivery methods using typical parameters for each system. For delivery, the phantom was placed using a rigid docking device and either MVCT or stereoscopic imaging. Dose profile and isodose measurements were performed using radiochromic film and two small volume ion chambers, an A1SL air-filled chamber (0.057 cm3) and a liquid ion chamber (0.0017 cm3). After noticing discrepancies in profile and isodose measurements with the HT, and the off-axis position of the target due to the positioning of the phantom in a table top docking device, and stereotactic head frame, three sets of HT plans (target at the center of the bore, 10 cm anterior and 10 cm lateral to the center of the bore) were generated. For linear accelerator and robotic SRS, comparison of relative isodose distributions dose profile, normalized to origin, agreed with, more than 99% pixels passing a gamma index of 3% dose difference and 1 mm distance to agreement, with absolute dose differences less than 1.5%. For HT, noticeable differences were seen between the two profiles, with a gamma passing rate as low as 40% when the target was off-axis. For HT, point dose measurements were significantly lower than predicted by the TPS when the phantom was positioned 10 cm away from the isocenter, while they were found to be higher than predicted by the TPS at isocenter, with the magnitude of the discrepancy dependent on the size of the target and detector. For small targets, delivered dose varied by more than 15% when targets are placed 10 cm away from the center of the bore as opposed to at the center, regardless of the volume of the detector. Although a field safety notice was released by the manufacturer on this issue, which stems from a limitation of modeling a continuous rotation in 51 static projections, the magnitude of the error was unclear. For HT, care should be taken during SRS treatment to ensure small targets are positioned near the center of the bore to avoid underdosage of the target until the issue is resolved.Poster Viewing Abstract 3622; TablePercentage difference between dose calculated by TPS and measured by A1SL and liquid ion chambers(Dose [A1SL]-Dose [TPS])*100/Dose(TPS)(Dose [liquid ion chamber]-Dose [TPS])*100/Dose(TPS)Target placed at isocenter Diameter = 6 mm2.2%14% Diameter = 10 mm5.2%14.5%Target 10 cm anterior to bore center Diameter = 6 mm-14.9%-6.9% Diameter = 10 mm-8.4%-1.4%Target 10 cm lateral to bore center Diameter = 6 mm-15.8%-5.9% Diameter = 10 mm-8.1%1.3% Open table in a new tab

A fast and accurate MC-based scatter correction algorithm was implemented on real cone-beam computed tomography (CBCT) data. An ACR CT accreditation phantom was imaged on a Varian OBI CBCT scanner using the standard-dose head protocol (100 kVp, 151 mAs, partial-angle). A fast Monte Carlo simulation developed in the EGSnrc framework was used to transport photons through the uncorrected CBCT scan. From the simulation output, the contribution from both primary and scattered photons for each projection image was estimated. Using these estimates, a subtractive scatter correction was performed on the CBCT projection data. Implementation of the scatter correction algorithm on real CBCT data was shown to help mitigate scatter-induced artifacts, such as cupping and streaking. The scatter corrected images were also shown to have improved accuracy in reconstructed attenuation coefficient values. In three regions of interest centered on material inserts in the ACR phantom, the reconstructed CT numbers agreed with clinical CT scan data to within 35 Hounsfield units after scatter correction. These results suggest that the proposed scatter correction algorithm is successful in improving image quality in real CBCT images. The accuracy of the attenuation coefficients extracted from the corrected CBCT scan renders the data suitable for adaptive on the fly dose calculations on individual fractions, as well as vastly improved image registration.

This work describes the design and use of an inexpensive phantom designed for precision measurements in radiosurgery quality assurance. The main features of this simple phantom include its solid water construction, interchangeable ion chamber holders and film registration system, thus allowing for measurement of small fields with several detectors using the same phantom. The entire phantom was constructed using one 30cm × 30cm × 3cm slab of solid water. The phantom contains a slot that allows for the placement of two small volume ion chambers (liquid and A1SL) via custom inserts near the center of the phantom. In addition, the plug can be filled for film measurements. The phantom can be split down the center to allow for the placement of a film. As opposed to registering film to room based markers, such as lasers, the phantom contains radio-opaque fiducials that puncture the film while also providing a method to register the film images to exported dose planes. In addition to the markers used for film registration, the phantom contains several external beebees that can be used to avoid ambiguity in image registration when using image guidance for setup. This simple phantom contains many features of other much more expensive phantoms designed for this purpose and has been found to be very useful clinically and in departmental research. The key elements of this phantom could be included in several other designs allowing it to be reproduced in other centers.

PURPOSE:To improve image quality in cone-beam computed tomography (CBCT) scans by implementation of a fast and accurate MC-based scatter correction algorithm.METHODS:A Solid WaterTM phantom was imaged on a Varian OBI CBCT scanner using the standard-dose head protocol (100 kVp, 151 mAs, partial-angle). A fast Monte Carlo simulation developed in the EGSnrc framework was used to transport photons through the uncorrected CBCT scan. From the simulation output, the contribution from both primary and scattered photons for each projection image was estimated. Using these estimates, a subtractive scatter correction was performed on the CBCT projection data. This correction procedure was repeated iteratively, using the previous scatter corrected scan as input to the Monte Carlo simulation.RESULTS:Implementation of the scatter correction algorithm on real CBCT data was shown to help mitigate scatter-induced artifacts, such as cupping and streaking. The scatter corrected images were also shown to have improved accuracy in reconstructed attenuation coefficient values. In a region of interest centered on the Solid Water phantom, the number of voxels agreeing to within 10% of the theoretical attenuation coefficient increased from 46% to 97% after two iterations of the scatter correction.CONCLUSIONS:These results suggest that the proposed scatter correction algorithm is successful in improving image quality in real CBCT images. The accuracy of the attenuation coefficients extracted from the corrected CBCT scan renders the data suitable for on-the-fly dose recalculations, as well as vastly improved image registration.

To evaluate the dosimetry of volumetric modulated arc therapy (VMAT) and three-dimensional conformal electron radiotherapy (3D-ERT) for the tumor bed boost in breast cancer patients and to report the acute toxicity of a series of patients treated with a VMAT tumor bed boost. Fifteen patients with breast cancer treated by lumpectomy and post-operative whole breast radiotherapy and requiring tumor bed boost were planned using both VMAT and 3D-ERT. The tumor bed dose evaluation volume (DEV) was prescribed 10 Gy in 4 fractions and dosimetry for the 2 techniques were compared. The acute toxicity in the first 25 patients treated with VMAT tumor bed boost was analyzed. Coverage of the tumor bed DEV was adequate in all VMAT plans but only in 13 of 15 3D-ERT plans: V95% = 99.1% vs. 97.6%, with VMAT and 3D-ERT respectively; p = 0.42. High dose to the breast outside the tumor bed DEV was significantly lower with VMAT (V107% = 0.01% vs. 2.13%; p<0.01). Ipsilateral lung V2Gy (2.13% vs. 19.21%; p<0.0001), V7Gy (0.00% vs. 2.03%; p = 0.0001) and mean dose (0.33 Gy vs. 1.29 Gy; p<0.001) were significantly lower with VMAT. In patients with left-sided tumors, heart V2Gy (1.11% vs. 6.17%; p<0.05) was significantly lower with VMAT, but V5Gy and mean dose were comparable. Contralateral lung and breast doses were minimal and not significantly different for the 2 techniques. Maximal acute dermatitis due to whole breast RT followed by VMAT tumor bed boost was grade 2 in 72% of patients and grade 1 in 28% of patients. For the tumor bed boost for patients with breast cancer, both VMAT and 3D-ERT provide adequate target volume coverage and low heart doses, but VMAT avoids unnecessary breast overdosage while improving ipsilateral lung dosimetry. The VMAT tumor bed boost was well tolerated, with an acceptable early toxicity profile.

Purpose: To evaluate the positioning accuracy of the BrainLAB ExacTrac Image Guidance System under gating conditions. Methods: Two types of phantoms were used in measurements: an anthropomorphic RANDO head phantom and a BrainLAB ExacTrac Gating Phantom. Our setup included a Varian Novalis Tx radiosurgery system equipped with the ExacTrac 6D IGRT. This system consists of an infrared positioning system for the initial patient positioning and patient tracking, and a stereoscopic kV X‐ray imaging system for final localization using internal markers or anatomy. Uncertainties were broken down into individual components, and the different BrainLAB fusion modalities (internal markers and bony fusion) were used to compare the effect of slice thickness on positional accuracy. Gating uncertainties were deduced with varying tumor motion amplitudes and window sizes in conjunction with a hidden target test. Results: Our results of CT slice thickness dependence for both fusion algorithms with the hidden target test gave similar deviation (<0.7mm), and were reasonably consistent up to a 5 mm slice width. Tumor motion and gating window size yielded an uncertainty of up to 1 mm for the parameters tested. Combining a non‐gating uncertainty of 0.9 mm with the gating uncertainty resulted in a geometrical accuracy of 1.6 ± 0.7 mm for 2.25 cm tumor amplitude and a 30% window size. For tumor motions up to 3 cm and gating window sizes up to 30%, the localization accuracy remained within 2 mm. Conclusions: We have tested the gating window and tumor amplitude effects on the spatial accuracy of the ExacTrac System equipped Novalis Tx linac for stereotactic body radiation therapy. While the CT slice thickness, mechanical deviation of the linac and gating window size contribute to uncertainty, the system provides an external modality that allows for localization accuracy of less than 2 mm for gated delivery.

Purpose: To asses the feasibility and benefits of replacing patient specific QA measurements and back‐up MU calculations with Monte Carlo based dose calculations for improved plan verification in radiation therapy. Methods: An in‐house Monte Carlo based planning system (MMCTP) has been clinically implemented at McGill University. Over 50 patient plans have been recalculated on the system by both dosimetrists and physicists. Initial plans chosen for recalculation were head and neck IMRT and lung SBRT plans, where simple monitor unit calculators and measurements in water equivalent phantoms are often insufficient to show agreement of the planned and delivered doses within 5%. Results: MMCTP was found to be easy to use by both dosimetrists and physicists. Eighteen head and neck IMRT plans were recalculated by a dosimetrist and over 35 lung SBRT plans have been calculated by the physics staff. As no effort was spent in implementing fast MC engines, the calculation time is long but all mechanical operations, i.e., the transfer of the plans from the clinical treatment planning system to the MMCTP system and starting the calculation, take only a few minutes. Differences between MMCTP and the clinical TPS (Varian Eclipse, AAA) were small in terms of dose delivered to the PTV for the head and neck IMRT group but differences of greater the 5% were seen among the lung SBRT group. For both groups, the largest differences were seen for anatomy close to the skin and near air cavities. Conclusions: Verifying the delivery accuracy of complex treatment plans can be challenging. Current techniques requiring recalculation and measurement of plans in uniform phantoms are cumbersome and do not take into account the patient specific heterogeneity effects. The MMCTP treatment planning system requires minimal physics resources and likely introduces a more comprehensive method of evaluating calculation accuracy throughout the treatment volume.

Purpose: To determine secondary cancer risk in paediatric patients treated with intensity modulated proton radiation therapy (IMPT) compared to IMRT. Methods: Proton therapy plans were created for fifteen patients previously treated with photon beam IMRT. IMPT plans were planned using the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, CA) with a scanned proton beam model. The proton plans were planned to the same prescription dose as the photon plans. Each proton plan consisted of one to three fields, depending on tumour location, and photon plan constraints were used as a guide for IMPT constraints. Proton and photon plans were compared for dose conformity, homogeneity, volumes of tissue receiving low doses, integral dose, and second cancer induction risk. Second cancer risk was determined using two methods. The relative risk of secondary cancer was found by applying a linear relationship between integral dose and relative risk of secondary cancer. The second approach used the organ equivalent dose concept to describe the dose in the body and then calculate the excess absolute risk (EAR) for solid cancers. Results: IMPT and IMRT plans had similar target conformity, homogeneity, near minimum, near maximum and median doses however IMPT plans had reduced integral dose and volumes of the body receiving low dose. IMPT plans resulted in a 0.313±0.098 smaller relative risk of secondary cancer than IMRT plans. The EAR of secondary cancer in the body 30 years after treatment was reduced by 20.89±9.56, 24.30±8.46 and 22.91±7.91 patients per 10000 patients per year for the linear, linear exponential and plateau dose‐response models respectively in IMPT compared to IMRT plans. Conclusions: Two methods were used to determine the risk of secondary cancers following radiation therapy in paediatric patients. Both methods indicated that IMPT results in a lower risk of secondary cancer than photon beam IMRT.

Purpose: This work explains the development of an advanced treatment planning system for the generation of intensity and energy modulated electron radiotherapy (MERT) plans. The quality of MERT plans for the treatment of tumour bed boost in breast cancer is compared to direct electrons (DE) and volumetric modulated photon arc therapy (MAT). Method and Materials: The MERT treatment planning and delivery system at McGill University consists of MMCTP, an inverse optimization toolkit and the few leaf electron collimator. The stepwise planning process consists of: 1) generating a series of field openings, 2) Monte Carlo dose calculation for each field, 3) planning constraints, 4) iterative direct‐aperture optimization. For evaluation purposes, fourteen patients with breast cancer treated by lumpectomy and requiring post‐operative whole breast radiotherapy with tumour bed boost were planned using conventional DE, MAT and MERT. The planning goal was to deliver 10 Gy to at least 95% of the target volume. Dosimetry parameters for all techniques were compared. Results: Target coverage and homogeneity was best for MERT (D98=9.77 Gy, D2=11.03 Gy) followed by MAT (D98=9.56 Gy, D2=11.07 Gy) and DE (D98=9.81 Gy, D2=11.52 Gy). Relative to the DE plans, the MERT plans predicted a reduction of 35% in mean breast dose (p<0.05), 54% in mean lung dose (p<0.05) and 46% in mean body dose (p<0.05). Relative to the MAT plans, the MERT plans predicted a reduction of 24%, 36% and 39% in mean breast dose, heart dose and body dose respectively (p<0.05). Conclusions: MERT was a considerable improvement in dosimetry over DE. In some cases, there was a dosimetric advantage in using MERT over MAT for increased target conformity and low‐dose sparing of healthy tissue. Based on the favorable comparisons shown in this work it is reasonable to suggest that MERT could play a more significant role in breast radiotherapy.

Purpose: To investigate the clinical application of a Monte Carlotreatment planning system for verification of fixed‐field IMRT and RapidArc® treatment plans. Methods: The McGill Monte Carlotreatment planning (MMCTP) system was used to calculate dose distributions from fixed‐field IMRT and RapidArc plans that had been generated in Eclipse v8.6 using the AAA dose calculation algorithm. Planar dose measurements were performed using the MapCHECK 2 diode array in three different measurement geometries. The first measurement geometry collapsed all of the gantry angles to 0° for the treatment delivery, and was only possible for the fixed‐field IMRT plans. The second measurement geometry used an Isocentric Mounting Fixture (IMF) that rotated the MapCHECK 2 array while the gantry rotated during RapidArc delivery. The third measurement geometry placed the MapCHECK 2 array between two 6 cm Solid Water® slabs with the detector remaining stationary on the treatment couch while the gantry rotated during RapidArc delivery. Results: A benchmark head and neck fixed‐field IMRT plan demonstrated excellent agreement between the Eclipse calculations, MMCTP calculations, and the MapCHECK 2 measurements for the collapsed gantry angle geometry. The Eclipse and MMCTP calculations were generally in good agreement for the RapidArc plans, but differences of up to 7% were seen in high dose gradient regions. For the RapidArc plans, the MapCHECK 2 results agreed well with the calculated distributions for stationary measurements between the Solid Water slabs but had poorer agreement when the IMF delivery was used. Conclusions: This work demonstrates that MMCTP can be used as an independent verification of fixed‐field IMRT and RapidArc treatment plans. MapCHECK 2 measurements between the Solid Water slabs agreed well with the calculated results and were the preferred QA method since the measured distributions best simulated the intended dose distributions in the patient's treatment plan.