RESUMO
PURPOSE: The implementation of an accurate beam model is an integral part of the commissioning of any planning system. This process is especially challenging in the case of IMRT beam models owing to the complexity of small field sizes and MLC leaf-end and tongue-and-groove effects. The question of how to judge the quality of an IMRT beam model in comparison with other versions of the same model is central to this work. METHODS: We make an important distinction between evaluation of the beam model and evaluation of the optimization routine that is a part of any IMRT planning system. The H-shaped target used in this work has several important features: it can only be covered by segments with small field size, for which all leaf design effects are important, and it has the overall dimensions of a common IMRT target. The procedure for inter-comparison of two IMRT beam models (old and new) involves the generation of two plans optimized with each beam model using identical IMRT prescriptions. Both plans are subsequently delivered on a solid water phantom with film located in two parallel planes with a small-volume ionization chamber inserted in the center. RESULTS: Four dose calculations are performed, such that each plan is calculated with either of the two beam models. The four dose distributions are subsequently compared with the two film measurements using gamma analysis. In addition, the absolute dose measured in the center of the dose distribution is compared with the calculated value. A score is assigned to each beam model based on the results. CONCLUSIONS: Using the procedure outlined in this presentation, different versions of an IMRT beam model can be compared and scored for quality. Adoption of a unified strategy for beam model inter-comparison can greatly facilitate the evaluation and commissioning of IMRT beam models.
RESUMO
PURPOSE: Modulated electron radiation therapy (MERT) can offer significant advantages for breast treatments over conventional radiotherapy in terms of sparing distal critical structures. While intensity modulated radiation therapy (IMRT) has the advantage of achieving better dose homogeneity inside the target combining both MERT and IMRT will be the ideal scenario. The Aim of the present study is to investigate the possibility of further improving breast radiation therapy using combined MERT/IMRT treatment technique. METHODS: Accurate modeling of a prototype motorized electron multileaf collimator was verified in a separate study. In this work treatment planning was performed by an in house Monte Carlo based inverse planning system. Dose deposition coefficients were calculated using MCPLAN and utilizing real patients CTs. Optimization is then conducted based on an equivalent uniform dose objective function. MERT and IMRT plans were created for different patients. RESULTS: The clinical beneficial outcome for MERT either alone or combined with IMRT was investigated based on isodose distributions and dose volume histograms. It is shown that MERT can give similar dose distributions as IMRT in some cases. For some cases, MERT could be advantageous whenever more skin dose was required. In some cases MERT can be identified as the best option. It was found that MERT compared to IMRT could introduce hot spots inside the target. However this was resolved in combined MERT/IMRT treatment. Dose uniformity can be restored with a reduction in the maximum lung and heart received dose. CONCLUSION: MERT can improve treatment plan quality for many breast patients. In some cases better results can be obtained with a combined MERT/IMRT treatment, where a homogeneous dose in the target can be achieved with an improvement in the DVH of critical structures. This work has been supported by a UICC American Cancer Society Beginning Investigators Fellowship funded by the American Cancer Society.
RESUMO
PURPOSE: To investigate the dosimetric effect of intrafraction tumor motion during gated RapidArc Stereotactic Body Radiotherapy (SBRT) delivery. METHOD: The realtime tumor motion data were retrieved from 6 lung patients. Each of them received 3 fractions of stereotactic radiotherapy treatments with Cyberknife Synchrony. Phase gating through an external surrogate was simulated with a gating window of 5 mm. The resulting residual tumor motion curves during gating (beam-on) were retrieved. RapidArc SBRT was planned on the platform of Varian Truebeam at 6 MV with 1400 MU/min. Planning target volume (PTV) was defined as physician-contoured clinical target volume (CTV) surrounded by an isotropic 5 mm margin. Each patient was prescribed with 60Gy/3 fractions. The RA plan typically consists of 2 arcs; each contains 90-120 control points. An algorithm was developed to reconstruct the delivered dose with tumor motion. The MLC segment is assumed to move relatively to a static tumor. Each MLC control point, mainly the leaf position were modified according to the probability density function of tumor motion. The newly created MLC control points were written back to the treatment file in the dicom format which was subsequently imported to treatment planning system (Varian Eclipse) for dose recalculation. RESULTS: The magnitude of dose deviation with motion is consistent with the excursion of the residual tumor movement. Overall CTV coverage of the study group is barely affected owing to the 5 mm margin. The fractional PTV dose coverage dropped by 4% at most and that from all fractions by 3%. An examination in the point dose shows an increase of 4% in the maximum dose and decrease of 10% for the minimum dose. CONCLUSION: With effective gating, interplay effect does not change the target coverage much during gated RapidArc SBRT. However it increases the dose nonuniformity inside target.
RESUMO
PURPOSE: RapidArc is routinely used for stereotactic radiotherapy for lung cancer. While treatment dose is optimized and calculated on a static CT image, the motion of the target in conjunction with the motion of the MLC may Result in a delivered dose deviating from the planed dose. In this study, we investigate the dosimetric consequences of the inter-play effect by simulating dynamic dose delivery on a dynamic CT dataset of real patients. METHODS: The target motion in 20 patients was analyzed and 5 patients with >10 mm motion were chosen for this study. The RapidArc plan for eachpatient is optimized on a free-breathing CT using 2 arcs. Inherent in each plan is data on the associated parameters such as timestamp, MLC leave position, gantry angle and delivered beam MUs for each control point. Simulated dynamic delivery is performed by associating these parameters with each of the breathing phases of the 4D-CT. The starting breathing phase is selected randomly for each of the two arcs. Dose from the derived partial plans associated with each phase of the 4D-CT dose is recalculated in Eclipse. Accumulation of dose is performed using deformable image registration from each phase of the 4D-CT to the exhale phase of the 4D-CT. RESULTS: The coverage of the GTV and PTV shows negligible variations from the interplay effect. But the Homogeneity Index is affected by the motion. The prescription isodose volume is smaller than what was from the treatment plan dose. There were both intra- and inter-fraction effects seen inthe OARs dose in some patients. CONCLUSIONS: We investigated the motioneffect in RapidArc Lung SBRT delivery in 5 patients. Negligible variations were shown for target coverage. However the motion effects were observed in high dose distribution and volume. Some OARs dose distributions were affected by the motion.
RESUMO
PURPOSE: Fabrication of electron beam cutouts not only is a time consuming process but also involves the handling of cerrobend which is a toxic material. Hospital workers involved in cutout construction can actually be exposed to toxic fumes that are usually generated during the process. The aim of this work is to study the feasibility of replacing electron cutouts with our prototype motorized electron multileaf collimator (eMLC). METHODS: Electron beams collimated by an eMLC have very similar penumbra to those collimated by applicators and cutouts as we already demonstrated in a previous study. However undulation of the isodose curves is expected due to the finite size of the eMLC. This may be a problem when the field edge is close to critical structure. Thus ten different breast cases that were previously treated with an electron boost were selected from our database. An inhouse Monte Carlo based treatment planning system were used for dose calculation using the patients CTs. For each patient two plans were generated one with electron beams collimated using the applicator/cutout combination and the other plan with beams collimated only by the eMLC. Treatment plan quality was compared for each patient based on dose distribution and dose volume histogram. In order to determine the optimal position of the leaves, the impact of the different leaf positioning strategies were investigated. RESULTS: Results have shown that target coverage and critical structure sparing can be effectively achieved by electron beams collimated by eMLC. Preliminary results have shown that the out-of-field strategy is most conservative and would be the recommended method to define the actual leaf position for the eMLC defined field. CONCLUSION: The eMLC represents an effective time saving and pollution free device that can completely eliminate the need for patient specific cutouts. This work has been supported by a UICC American Cancer Society Beginning Investigators Fellowship funded by the American Cancer Society.
RESUMO
PURPOSE: If the Linac is unavailable during the IMRT treatment schedule, the patient can be switched to a different Linac or prostpond treatment until the origonal Linac is available. The resulting dosimetric difference is estimated and the compromise in the TCP is estimated for both scenarios. This work investigates the feasibility and rationale of switching patients between different accelerators for IMRT in contrast to prostponing the treatment. METHODS: We performed Monte Carlo simulations of photon beams from different Linac models and vendors. Prostate and head and neck treatment plans for Siemens Primus, Primart, Artiste and Varian-21Ex/IX accelerators are studied in this work. Dose distributions for given plans are recalculated using different beam data with the same nominal energy from different Linacs. We have compared DVHs, the maximum, the minimum and the mean dose to the target and critical structures due to switching accelerators. In the process of switching a treatment plan to a different accelerator, there are issues, such as optimum penumbra compensation, dose distribution at the boundary of target and critical structures and multileaf collimator (MLC) leaf width effects, needed to be considered and verified with measurements. In making the final decision whether to switch machines, the TCP based on a linear-quadratic model with time factor is considered. RESULTS: Two DVHs of two plans from Varian and Siemens models are delivered on different machines. Slight dose coverage differences have been observed. TCP estimation with both delayed and without delayed treatments is calculated. Undesired drop of TCP is observed with treatment gap. CONCLUSIONS: Based on the analyses done in this work, it is therapeutically more beneficial to switch a patient to a different machine than to postpone a treatment until the original machine is available, especially for fast growing tumors such as head and neck cancers.
RESUMO
PURPOSE: To compare measured and calculated doses using Pencil Beam (PB) and Monte Carlo (MC) algorithm on a CIRS thorax phantom for SBRT lung treatments. METHODS: A 6MV photon beam generated by a Primus linac with an Optifocus MLC (Siemens) was used. Dose calculation was done using iPlan v4.1.2 TPS (BrainLAB) by PB and MC (dose to water and dose to medium) algorithms. The commissioning of both algorithms was done reproducing experimental measurements in water. A CIRS thorax phantom was used to compare doses using a Farmer type ion chamber (PTW) and EDR2 radiographic films (KODAK). The ionization chamber, into a tissue equivalent insert, was placed in two position of lung tissue and was irradiated using three treatments plans. Axial dose distributions were measured for four treatments plans using conformal and IMRT technique. Dose distribution comparisons were done by dose profiles and gamma index (3%/3mm). RESULTS: For the studied beam configurations, ion chamber measurements shows that PB overestimate the dose up to 8.5%, whereas MC has a maximum variation of 1.6%. Dosimetric analysis using dose profiles shows that PB overestimates the dose in the region corresponding to the lung up to 16%. For axial dose distribution comparison the percentage of pixels with gamma index bigger than one for MC and PB was, plan 1: 95.6% versus 87.4%, plan 2: 91.2% versus 77.6%, plan 3: 99.7% versus 93.1% and for plan 4: 98.8% versus 91.7%. It was confirmed that the lower dosimetric errors calculated applying MC algorithm appears when the spatial resolution and variance decrease at the expense of increased computation time. CONCLUSIONS: The agreement between measured and calculated doses, in a phantom with lung heterogeneities, is better with MC algorithm. PB algorithm overestimates the doses in lung tissue, which could have a clinical impact in SBRT lung treatments.
RESUMO
PURPOSE: To investigate the feasibility of RapidArc technique on intracranial radiosurgery for multiple lesions. METHODS: Six patients who were previously treated using cone-based technique, Cyberknife were included in this study. These patients have multiple lesions (6-9, mostly metastasis). In our current clinical practice, each lesion was planned and treated individually. The prescription was 15-21 Gy at 80% with single fraction. These cases were replanned with RapidArc on the platform of Varian Truebeam STx equipped with high resolution MLC leaves of 2.5mm at center. The maximum dose rate is 1400 MU/min at 6 MV for flattering filter free mode. Because of long span of multiple lesions, the targets were divided into two groups with two isocenters. Each plan with one isocenter contains 4 non-coplanar arcs, and dose optimization was performed with the two plans combined. Critical organs, such as eyes, brainstem and brain were constrained. The individual Cyberknife plans were summed to compare with the RapidArc plan. Scenarios of setup error were simulated during RapidArc treatment. RESULTS: RapidArc plans can achieve comparable target coverage and normal tissue avoidance to Cyberknife plans. The brain dose volume histogram (DVH) curves of the two techniques are similar in spite of different appearance of their 3D dose distributions. MU is much higher for summed Cyberknife plan. Because RapidArc can treat several lesions together, the complete treatment time for all lesions is significantly reduced. However RapidArc treatment is susceptible to setup error, which may cause increase in normal tissue dose and decrease in target dose coverage. The level of discrepancy depends on the magnitude of setup error, location and dose distribution of the target. CONCLUSION: Multiple brain lesions treatment with RapidArc radiosurgery is clinically feasible with setup error fully accounted. It can provide dose performance comparable to cone-based Cyberknife treatment.
RESUMO
PURPOSE: To develop and validate an EPID-based 4D patient dose reconstruction framework accounting for linac delivery uncertainties, interfractional and intrafractional motions, and interplay effect. METHODS: Patients with fiducial markers were scanned with 4D-CT for SBRT planning. Before treatment, in-room 4D-CT was performed. Both the MLC and the tumor movements were tracked by continuously acquiring EPID images during treatment. Instead of directly using the heterogeneous transit photon fluence measured by the EPID, this method reconstructed the incident beam fluence based on the MLC apertures measured by the EPID and the delivered MU recorded by the linac. To account for the time-dependent-geometry, the incident fluence distributions were sorted into their corresponding phases based on the tumor motion pattern detected by the EPID and accumulated as the incident fluence map for each phase. Together with 4D-CT, it was then used for Monte Carlo dose calculation. Deformable registration was performed to sum up the phase doses for treatment assessment. The feasibility of using the transit EPID images for incident fluence reconstruction was evaluated against EPID in-air measurements. The accuracy of 3D- and 4D-dose reconstruction was validated by a motordriven cylindrical diode array for six clinical SBRT plans. RESULTS: The average difference between the measured and reconstructed fluence maps is within 0.16%. The reconstructed 3D-dose shows 1.4% agreement in the CAX-dose and >98.5% gamma-passing-rate (2%/2mm) in the peripheral-dose. A distorted dose distribution is observed in the measurement for the moving ArcCheck-phantom. The comparison between the measured and the reconstructed 4D-dose without considering interplay fails the gammaevaluation (59%-88.9% gamma-passing-rate). In contrast, when the interplay is considered, the dose distortion phenomena is successfully represented in the reconstructed dose (>97.6% gamma-passing-rate). CONCLUSIONS: The experimental validation demonstrates that the proposed method provides a practical way to reconstruct the fractional 4D-doses received by the patient and enables adaptive SBRT strategy.
RESUMO
PURPOSE: To examine and facilitate the feasibility of the ArcCheck cylindrical diode array system as a patient specific QA device for CyberKnife radiosurgery delivery. METHODS: There is an obvious necessity for CyberKnife robotic radiosurgery patient QA procedures for hypofractionated treatment of larger planned treatment volumes (PTV), e.g. prostate. This need will increase when the future CyberKnife MLC is introduced. The small unflattened CyberKnife fields, along with the variation of beam-to-detector spatial angles, pose a significant detection challenge for dosimetric systems. The feasibility of the ArcCheck (Sun Nuclear Inc.) cylindrical diode array system for patient-specific QA on the CyberKnife is demonstrated using a beam-to-diode specific angular correction that was developed and has been applied. For localization and tracking, four gold seed fiducial markers were embedded in the system's central plug. We used a Monte Carlo 1% uncertainty for the dose calculation. RESULTS: By disabling the Linac based corrections and applying the custom CyberKnife correction that we developed, the passing rate increased from 39.6% to 99.8% using a 3%3mm gamma criteria for a given lung case. An additional lung case passed 98.5%. In both cases, a 10% dose threshold was used. In addition, brain, trigeminal nerve and lung cases with synchrony tracking are being investigated. CONCLUSIONS: We demonstrated the ArcCheck feasibility for CyberKnife patient specific QA performance. The custom CK angular correction that we developed and applied showed a high passing rate for the lung cases. A verification of the polar angle response should be conducted, in addition to the azimuthal angle that was verified for Linacs. Any data that is being retrieved is additional data to the current chamber point measurement procedures.