R. Yaparpalvi

Application of AAPM TG 119 to volumetric arc therapy (VMAT)

The purpose of this study was to create AAPM TG 119 benchmark plans for volumetric arc therapy (VMAT) and to compare VMAT plans with IMRT plan data. AAPM TG 119 proposes a set of test clinical cases for testing the accuracy of IMRT planning and delivery system. For these test cases, we generated two treatment plans, the first plan using 7–9 static dMLC IMRT fields and a second plan utilizing one‐ or two‐arc VMAT technique. Dose optimization and calculations performed using 6 MV photons and Eclipse treatment planning system. Dose prescription and planning objectives were set according to the TG 119 goals. Plans were scored based on TG 119 planning objectives. Treatment plans were compared using conformity index (CI) for reference dose and homogeneity index (HI) (for D5‐D95). F or test cases prostate, head‐and‐neck, C‐shape and multitarget prescription dose are 75.6 Gy, 50.4 Gy, 50 Gy and 50 Gy, respectively. VMAT dose distributions were comparable to dMLC IMRT plans. Our planning results matched TG 119 planning results. For treatment plans studied, conformity indices ranged from 1.05–1.23 (IMRT) and 1.04–1.23 (VMAT). Homogeneity indices ranged from 4.6%–11.0% (IMRT) and 4.6%–10.5% (VMAT). The ratio of total monitor units necessary for dMLC IMRT to that of VMAT was in the range of 1.1–2.0. AAPM TG 119 test cases are useful to generate VMAT benchmark plans. At preclinical implementation stage, plan comparison of VMAT and IMRT plans of AAPM TG 119 test case allowed us to understand basic capabilities of VMAT technique. PACS number: 87.55.Qr

A method incorporating 4DCT data for evaluating the dosimetric effects of respiratory motion in single-arc IMAT.

This study introduces a method incorporating 4DCT data to determine the impact of respiratory motion in single-arc intensity-modulated arc therapy (IMAT). Simulation was done by re-warping the static dose distribution of all phases of a 4DCT image set with a 3D deformation map to reference CT images at end-inspiration and end-expiration. To calculate the dose received during respiration under IMAT, the control points were interpolated and re-distributed into separate IMAT plans corresponding to each respiratory phase. This study also investigated the role that plan complexity may play in the dosimetric impact of the respiratory motion in the delivery of IMAT. The dosimetric impact of organ motion was evaluated by analyzing the degradation of D(95,) D(50) and D(05) of the CTV and PTV. From the results shown for the patients in this study who had maximum organ motion displacement approximately 15 mm, the dosimetric impact is rather small. Therefore, our preliminary results suggest that respiratory motion of less than 1.5 cm may be ignored for both moderately and highly modulated IMAT, irrespective of the number of fractions. Specifically, highly modulated plans only increased the degradation of D(95) of the DVH curves for a single fraction by 2% in the CTV and 9% in the PTV compared to the expected value of the multi-fraction plan.

SU‐FF‐T‐288: Initial Experience with a Commercial Monte Carlo Electron Treatment Planning System

Introduction:Monte Carlo modeling of clinical electron beams has the potential to substantially improve accuracy and quality of treatment planning, as excessive compute time and lack of commercial availability has hindered its application. Here we present a preliminary evaluation of a commercial electron Monte Carlo algorithm.Methods and Materials: Percent depth dose and profiles of 6‐20 MeV electrons and 6×6 – 25×25 cm cones were measured in a water tank at 100 cm SSD using a Farmer chamber for electrons. Absolute output was measured at 110 and 100 cm SSDs. Outputs and distributions of two extreme test cases were measured: a 2.1 cm × 3 cm insert in a 6×6 cone and a 2.8 cm × 15.7 cm long slit on a 25×25 cone. The algorithms ability to accurately model relative and absolute dose of an obliquely (30°) oriented beam was evaluated ionization chamber measurements. Clinical cases with were checked using Mosfet dosimeters in vivo. Monte Carlo calculations were performed with a 2 mm grid, and smoothing filters provided with the algorithm were applied to minimize noise in the data. Results: Agreement of 2% of measured and modeled doses was found over the evaluated range of energies, cones, obliquities and SSDs. Compute times of 1–5 minutes were a function of increasing field size. Visual comparison of the shapes of the profiles was in agreement with measurement. Profiles of the eccentric geometry test cases appeared to be to be physically unrealistic (e.g. an inverted V) in the smaller dimension suggesting that the beam model was not valid. Conclusions: The Monte Carlo electron algorithm provides accurate distributions for most clinical cases. For extreme cases, measurements should be made to test the accuracy of the system, and further development of the algorithm is recommended.

PO-0845: Evaluating dosimetric indices in lung SBRT for establishing treatment plan quality guidelines

Material and Methods: Treatment plans of 100 Lung SBRT patients treated were retrospectively reviewed. The targets (n=102) were evenly distributed – right lung (53) and left lung (49). Patients were prescribed to a total dose of 50-60 Gy in 3-5 fractions. Dose optimizations were accomplished with 6 MV using either 2-5 arcs VMAT (90); 8-14 IMRT fields (6) or 1016 static fields (6). Dose calculations were performed using AAA algorithm with heterogeneity correction. A literature review on dosimetric indices recommended for qualitative analyses of conventional and stereotactic dose distributions in target coverage, homogeneity, conformity and gradient categories was performed. From each patient treatment plan, the necessary parameters for calculating various indices were quantified.

SU-F-T-600: Influence of Acuros XB and AAA Dose Calculation Algorithms On Plan Quality Metrics and Normal Lung Doses in Lung SBRT

PURPOSEnTo study the influence of superposition-beam model (AAA) and determinant-photon transport-solver (Acuros XB) dose calculation algorithms on the treatment plan quality metrics and on normal lung dose in Lung SBRT.nnnMETHODSnTreatment plans of 10 Lung SBRT patients were randomly selected. Patients were prescribed to a total dose of 50-54Gy in 3-5 fractions (10?5 or 18?3). Doses were optimized accomplished with 6-MV using 2-arcs (VMAT). Doses were calculated using AAA algorithm with heterogeneity correction. For each plan, plan quality metrics in the categories- coverage, homogeneity, conformity and gradient were quantified. Repeat dosimetry for these AAA treatment plans was performed using AXB algorithm with heterogeneity correction for same beam and MU parameters. Plan quality metrics were again evaluated and compared with AAA plan metrics. For normal lung dose, V20 and V5 to (Total lung- GTV) were evaluated.nnnRESULTSnThe results are summarized in Supplemental Table 1. PTV volume was mean 11.4 (±3.3) cm3 . Comparing RTOG 0813 protocol criteria for conformality, AXB plans yielded on average, similar PITV ratio (individual PITV ratio differences varied from -9 to +15%), reduced target coverage (-1.6%) and increased R50% (+2.6%). Comparing normal lung doses, the lung V20 (+3.1%) and V5 (+1.5%) were slightly higher for AXB plans compared to AAA plans. High-dose spillage ((V105%PD - PTV)/ PTV) was slightly lower for AXB plans but the % low dose spillage (D2cm) was similar between the two calculation algorithms.nnnCONCLUSIONnAAA algorithm overestimates lung target dose. Routinely adapting to AXB for dose calculations in Lung SBRT planning may improve dose calculation accuracy, as AXB based calculations have been shown to be closer to Monte Carlo based dose predictions in accuracy and with relatively faster computational time. For clinical practice, revisiting dose-fractionation in Lung SBRT to correct for dose overestimates attributable to algorithm may very well be warranted.

SU-E-T-791: Validation of a Determinant Based Photon Transport Solver in Dose Perturbed By Diverse Media

Purpose: To validate a determinant based photon transport solver in dose imparted within different transition zone between different medium. Methods: Thickness (.2cm,.5cm, 1cm, 3cm) from various materials (Air - density=0.0012g/cm3, Cork-0.19g/cm3, Lung-0.26g/cm3, Bone-1.85g/cm3, Aluminum (Al)-2.7g/cm3, Titanium (Ti)-4.42g/cm3, Iron (Fe)-8g/cm3) were sandwiched by 10cm solid water. 6MV were used to study the calculation difference between a superposition photon beam model (AAA) and the determinant based Boltzmann photon transport solver (XB) at the upstream (I) and downstream boarder (II) of the medium, within the medium (III), and at far distance downstream away from medium (IV). Calculation was validated with available thickness of Air, Cork, Lung, Al, Ti and Fe. Results are presented as the ratio of the dose at the point with medium perturbation to the same point dose without perturbation. Results: Zone I showed different backscatter enhancement from high-density materials within the 5mm of the upstream border. AAA showed no backscatter at all, XB showed good agreement beyond 1mm upstream (1.18 vs 1.14, 1.09 vs 1.10, and 1.04 vs 1.05 for Fe, Ti, and Fe, respectively). Zone II showed a re-buildup after exiting high-density medium and Air but no build up for density close to water in both of the measurement and XB. AAA yielded the opposite results in Zone II. XB and AAA showed in Zone III very different absorption in high density medium and the Air. XB and measurement had high concordance regarding photon attenuation in Zone IV. AAA showed less agreement especially when the medium was Air or Fe. Conclusion: XB compared well with measurement in regions 1mm away from the interface. Planning using XB should be beneficial for External Beam Planning in situations with large air cavity, very low lung density, compact bone, and any kind of metal implant.

SU-E-J-202: Is Pretreatment Imaging at Each Treatment Fraction Needed in Spine SBRT to Enable Margin Reduction?

PURPOSEnTo study translational setup errors of Spine SBRT treatments using BodyFIX immobilization system and the potential for margin reduction using pretreatment imaging.nnnMETHODSn148 online match results of 39 Spine SBRT patients treated at our institution were reviewed. All treatments performed using Varian, TrueBeam and MI™, BodyFIX immobilization device. Prescriptions (dose-fractionation schedules) for the patients in the study are 8 Gy X 3 (n=24), 6 Gy X 4 (n=10) and 16Gy X 1 (n=5). For all patients, OBI was performed on verification day and before each treatment. Recorded shifts are from adopted couch positions after physician online review of KV image-pair and CBCT. Data analyzed by computing Mean error (M), Systematic error (Σ) and Random error (σ) for three translation directional shifts and 3D Vector length. To estimate correlation between imaging sessions, Anova-test and T-Test with alpha of 0.05 performed over the data.nnnRESULTSnMean error (M), Systematic error (Σ) and Random error (σ) of Vertical direction are -1.81mm, 4.05mm and 2.01mm respectively. Similarly, Errors in Longitudinal direction are 0.76mm (M), 4.21mm (Σ) and 1.34mm (σ). Errors in Lateral direction are - 0.16mm (M), 3.22mm (Σ) and 1.83 mm (σ). Mean, SD and CI95% of 3D Vector lengths are 7.93mm, 4.41mm and 0.7mm respectively. Anova-test results show, P-Values for longitudinal and lateral shifts are > 0.3 and P-value of vertical shift is 0.054. This signifies mean of the each imaging session are not significantly different. T-Test on treatment sessions data of vertical shifts with respect to verification day shows, mean of Vertical shifts on verification day are different from Vertical shifts on treatment sessions.nnnCONCLUSIONnResults shows immobilization with BodyFIX is reproducible in the mm range. To eliminate systematic error component and to enable margin reduction, pretreatment imaging prior to each treatment fraction is indispensable.

SU-E-T-502: Biometrically Accepted Patient Records

PURPOSEnTo implement a clinically useful palm vein pattern recognition biometric system to treat the correct treatment plan to the correct patient each and every time and to check-in the patient into the department to access the correct medical record.nnnMETHODSnA commercially available hand vein scanning system was paired to Aria and utilized an ADT interface from the hospital electronic health system. Integration at two points in Aria, version 11 MR2, first at the appointment tracker screen for the front desk medical record access and second at the queue screen on the 4D treatment console took place for patient daily time-out. A test patient was utilized to check accuracy of identification as well as to check that no unintended interactions take place between the 4D treatment console and the hand vein scanning system. This system has been in clinical use since December 2013.nnnRESULTSnSince implementation, 445 patients have been enrolled into our biometric system. 95% of patients learn the correct methodology of hand placement on the scanner in the first try. We have had two instances of patient not found because of a bad initial scan. We simply erased the scanned metric and the patient enrolled again in those cases. The accuracy of the match is 100% for each patient, we have not had one patient misidentified. We can state this because we still use patient photo and date of birth as identifiers. A QA test patient is run monthly to check the integrity of the system.nnnCONCLUSIONnBy utilizing palm vein scans along with the date of birth and patient photo, another means of patient identification now exits. This work indicates the successful implementation of technology in the area of patient safety by closing the gap of treating the wrong plan to a patient in radiation oncology. FOJP Service Corporation covered some of the costs of the hardware and software of the palm vein pattern recognition biometric system.

SU‐E‐J‐135: Measurements of Non‐Linearity Features of Breathing Patterns Using Recurrence Quantification Analysis (RQA) and Dynamic Complexity (DC)

Purpose: To investigate if there exists a difference in breathing patterns between patients treated with SBRT and IMRT using RQA and DC (using K2: correlation entropy and D2: correlation dimension) measures. Methods: 9 patients treated with SBRT and 8 treated with IMRT were scanned with 4D CT and the breathing patterns were acquired. One of the SBRT patients was scanned with and without meditation. Each breathing signal consisted of a scalar time series and recurrence quantification analysis (RQA) was utilized to determine the following measures: Periodicity of the system as percentage of recurrence points (%RR), determinism (DET), maximum diagonal line length (L_max) whose inverse the divergence (DIV) is measure for how fast trajectories diverge from each other, the average diagonal line length (L) that can be interpreted as the mean prediction time of the signal, and the entropy (ENTR) a measure of information complexity. In addition the invariant measures of K2 and D2 were also estimated. A locally nonlinear forecast was applied to predict future breathing signals of N time step ahead for the patient with and without meditation. Results: Our results showed %RR has significant correlation with L_max and has inverse correlation with DIV. DET has significant correlation with Lmax, L, ENTR and DIV. Independent t test suggests there is no difference between the SBRT and IMRT groups in terms of the RQA measures and K2. Patient that had undergone meditation showed improvement in %RR, L_max, DIV, K2 and had an estimated correlation dimension of 1.7. Prediction showed similar results for one and three time step ahead but meditation one had better prediction horizon when time step was higher. Conclusion: RQA is a powerful tool that allows one to analyze the dynamic nature of breathing pattern. No significant difference was found in the dynamical complexity of SBRT and IMRT patients.

SU‐E‐T‐172: Portal Dosimetry of Gated VMAT; with and Without Gating

PURPOSEnTo study duty cycle and respiratory period influence on VMAT delivery using portal dosimetry.nnnMETHODSnFor this study, we selected 12 ARC fields from four different gated VMAT plans with Varian, TrueBeam 6 MV photon beam and HD120 MLC. For Planning and portal dose prediction we used Varian, Eclipse (Version 10.0) AAA and PDIP algorithms respectively. We acquired integrated portal images with and without gated arc fields delivered to EPID at SID 140cm. We simulated respiratory periods with Quasar, Motion Phantom. For gated delivery, we used planned duty cycle for patient respiratory trace, and 6 sec per breath (SPB) and duty cycles 40%, 30%, 20% and 10% with 3SPB. Eclipse, Portal dosimetry software was used for Gamma analyses of measured and predicted portal dose images. In addition, we compared with and without gated portal dose images of same arc. MLC log file analyses of delivered arcs were done with Mobius, Dose lab software. Gamma passing criteria is to have gamma<1 for greater than 90% data points.nnnRESULTSnFor all arcs, predicted vs. measured dose distributions passed with criteria ΔD = 3% & DTA = 3mm, excluding few 10% duty cycle measurements. For these, pass criteria is ΔD = 5% & DTA = 4mm. Similarly, with and without gated portal dose image comparison passed the criteria of ΔD = 1% & DTA = 1mm excluding few 10% duty cycle measurements. For these pass criteria is ΔD = 2% & DTA = 2mm. For all delivered arcs, Leaf error 95 percentile is < 0. 1mm and error RMS is below 0.05mmConclusion: Except 10 % duty cycle, the gating influence on VMAT delivery is minimal. Therefore, gated VMAT QA can be done without gating. Duty cycles larger than 20% are recommended to minimize delivery errors.