Neck Treatment Plans Research Articles

Artificial intelligence based auto-contouring solutions for use in radiotherapy treatment planning of head and neck cancer

BackgroundManual contouring is time-consuming and subjective. Thus, auto-segmentation methods, which can be deployed in the existing workflow, are needed. The objective of this study was to assess the feasibility of Limbus AI and AI Rad Companion auto-contours for head and neck treatment planning. MethodsHead and neck patients treated with RapidArc were selected retrospectively. The manual contours on the planning CT were used as reference. Geometric analysis of the auto-contours was performed using several evaluation metrics such as the Dice Similarity Coefficient (DSC) and the Mean Distance to Conformity (MDC). Dosimetric analysis was performed by recalculating the original plan on the auto-contours and comparing Dose Volume Histogram (DVH) metrics to the original plan. Results and discussionBoth AI tools tend to underestimate the volumes of brainstem and cord. For brainstem and parotids, median DSC values were ≥ 0.8. For all auto-contours, median MDC values were ∼ 3–6 mm. Median differences were found of up to ±7 % in DVH points on the auto-contours relative to the planning CT contours, but these were not statistically-significant. ConclusionThe auto-contours could be used as a starting point to assist the clinician with the manual contouring of structures on the planning and re-scanning planning CT.

Sensitivity and specificity of secondary dose calculation for head and neck treatment plans.

Secondary dose calculation (SDC) with an independent algorithm is one option to perform plan-specific quality assurance (QA). While measurement-based QA can potentially detect errors in plan delivery, the dose values are typically only compared to calculations on homogeneous phantom geometries instead of patient CT data. We analyzed the sensitivity and specificity of an SDC software by purposely introducing different errors and determined thresholds for optimal decisions. Thirty head and neck VMAT plans and 30 modifications of those plans, including errors related to density and beam modelling, were recalculated using RadCalc with a Monte Carlo algorithm. Decision thresholds were obtained by receiver operating characteristics (ROC) analysis. For comparison, measurement-based QA using the ArcCHECK phantom was carried out and evaluated in the same way. Despite optimized decision thresholds, none of the systems was able to reliably detect all errors. ArcCHECK analysis using a 2%/2mm criterion with a threshold of 91.1% had an area under the curve (AUC) of 0.80. Evaluating differences in recalculated and planned DVH parameter of the target structures in RadCalc with a 2% threshold an AUC of 0.86 was achieved. Out-of-field deviations could be attributed to weaknesses in the beam model. Secondary dose calculation with RadCalc is an alternative to established measurement-based phantom QA. Different tools catch different errors; therefore, a combination of approaches should be preferred.

Investigating the feasibility of using Ethos generated treatment plans for head and neck cancer patients.

The Varian Ethos treatment platform is designed to automatically create complex RT treatment plans, reducing both workload and operator variability in plan quality. The aim of this study is to evaluate the quality of Ethos-generated head and neck (H&N) treatment plans. Ethos plans were created for ten previous H&N patients and these were compared with the original clinical plans generated in Eclipse. Ethos automatically creates several plans with different field arrangements for each patient. All plans were compared quantitatively using: dose-volume metrics; dose conformity; dose heterogeneity and monitor units (MU). In addition, two H&N Oncologists assessed the clinical acceptability of the Ethos plans. Consultant 1 judged there to be at least three clinically acceptable Ethos plans for 9 out of 10 patients reviewed. Consultant 2 approved of at least two Ethos plans for 5 out of 5 patients reviewed. The Ethos plans' average dose metrics were comparable to the clinical plans. The average plan MU was similar for Eclipse and Ethos VMAT plans. The average plan MU for Ethos IMRT plans was larger with respect to all VMAT plans. The Ethos Treatment Planning system is capable of automatically creating good quality treatment plans for a range of H&N cancer patients.

The Auto-Lindberg Project: Standardized Target Nomenclature in Radiation Oncology Enables Real-World Data Extraction From Radiation Treatment Plans

To automatically interrogate treatment plans with standardized target nomenclatures to generate real world data on lymph node (LN) involvement in patients with head and neck cancer. Using a large dataset of head and neck treatment plans (TPs), the full dataset including structure names was extracted for analysis and combined with relevant clinical variables. For each patient, the standardized nomenclature describing the LN targets was interrogated to determine the LN levels which were intended for gross nodal/high dose treatment. The dataset was further stratified based on primary site and T-category. The entire dataset and sub-stratified datasets were visualized using a traditional format for ipsilateral and contralateral cervical lymph node regions as well as with summary statistics related to the levels of cervical LN involvement. Of the dataset, 2030 patients had a clinically assigned laterality including 1053 oropharynx (OPC), 552 larynx (LC), 162 nasopharynx (NPC), and 123 hypopharynx (HPC). Patients with OPC had 80% ipsilateral nodal involvement (INI), 33% contralateral nodal involvement (CNI), and 18% retropharyngeal nodal involvement (RNI). Patients with LC had 18% INI, 7% CNI, and 2% retropharyngeal LN. Patients with NPC had 73% INI and 55% CNI, and 64% RNI. Patients with HPC had 72% INI, 28% CNI, and 11% RNI. When stratified by level, patients with OPC, NPC, larynx, and HPC had substantially different proportions of each neck level involved. OPC LN involvement also varied significantly in HPV-positive vs. HPV-negative patients. Routine clinical use of standardized detailed target volume nomenclature including separate LN targets specifying laterality and specific levels of neck involvement allows for automated real world data extraction and visualization. This simple strategy can be easily adopted in any radiation oncology center to develop real world data for any clinical disease site with minimal adaptation.

Variations in Head and Neck Treatment Plan Quality Assessment Among Radiation Oncologists and Medical Physicists in a Single Radiotherapy Department.

BackgroundAgreement between planners and treating radiation oncologists (ROs) on plan quality criteria is essential for consistent planning. Differences between ROs and planning medical physicists (MPs) in perceived quality of head and neck cancer plans were assessed.Materials and MethodsFive ROs and four MPs scored 65 plans for in total 15 patients. For each patient, the clinical (CLIN) plan and two or four alternative plans, generated with automated multi-criteria optimization (MCO), were included. There was always one MCO plan aiming at maximally adhering to clinical plan requirements, while the other MCO plans had a lower aimed quality. Scores were given as follows: 1–7 and 1–2, not acceptable; 3–5, acceptable if further planning would not resolve perceived weaknesses; and 6–7, straightway acceptable. One MP and one RO repeated plan scoring for intra-observer variation assessment.ResultsFor the 36 unique observer pairs, the median percentage of plans for which the two observers agreed on a plan score (100% = 65 plans) was 27.7% [6.2, 40.0]. In the repeat scoring, agreements between first and second scoring were 52.3% and 40.0%, respectively. With a binary division between unacceptable (scores 1 and 2) and acceptable (3–7) plans, the median inter-observer agreement percentage was 78.5% [63.1, 86.2], while intra-observer agreements were 96.9% and 86.2%. There were no differences in observed agreements between RO–RO, MP–MP, and RO–MP pairs. Agreements for the highest-quality, automatically generated MCO plans were higher than for the CLIN plans.ConclusionsInter-observer differences in plan quality scores were substantial and could result in inconsistencies in generated treatment plans. Agreements among ROs were not better than between ROs and MPs, despite large differences in training and clinical role. High-quality automatically generated plans showed the best score agreements.

Dosimetric Comparison of Two Volume and Three Volume Head and Neck Treatment Plans; is an Intermediate Risk CTV Necessary?

Dosimetric Comparison of Two Volume and Three Volume Head and Neck Treatment Plans; is an Intermediate Risk CTV Necessary?

Application of Tumor Control (TCP) and Normal Tissue-Complication Probabilities (NTCP) to Determine the Best Robust Optimization (RO) Approaches for Proton Head and Neck Radiotherapy

Implementation of RO for intensity modulated proton therapy (IMPT) can be performed with coupled or uncoupled geometric and stopping power (range) uncertainties. Currently, there is no clinical consensus regarding the optimal approach. TCP and NTCP models can be used as a relative metric to help determine the approach that may lead to better clinical outcomes. IMPT plans were created for ten anonymized head and neck (HN) patients previously treated with volumetric modulated arc therapy (VMAT). Comparison plans were created utilizing RO with both coupled and uncoupled geometric and stopping power uncertainties using equivalent optimization parameters. Equivalent-uniform dose (EUD), logistic-response modeling of multi-volume TCPs and NTCPs of the parotids (≥ Grade 2 xerostomia), constrictors (≥ Grade 2 dysphagia), and larynx (≥ Grade 2 edema) were used to compare plans under 20 separate uncertainty, coupled and uncoupled, evaluation scenarios (geometric uncertainties of ±3mm and stopping power uncertainties of ±3%). Although there are patient-by-patient variations, mean total NTCP, for the three OARs analyzed, were comparable under both coupled and uncoupled RO scenarios. Compared to nominal uncoupled plans, the total mean NTCP increased by a maximum of 10.7±4.8% (uncoupled RO) and 9.9±6.3% (coupled RO). Both maximums occurred for an evaluation with 3mm superior shifts coupled with underestimation (-3%) of stopping powers. Compared to the same nominal plans, multiplicative TCP rates varied precipitously under various evaluation scenarios, ranging from a loss of TCP of 72.7±26.6% to a TCP gain of 10.2±9.8%. Regardless of the magnitude of the TCP change, the coupled and uncoupled plans for each scenario were within ±16.1% of each other with a mean difference of 0.2±3.2%. Plans with uncoupled RO had 5.2% greater relative TCP than coupled RO plans when stopping powers were underestimated. Conversely, plans with coupled RO had 4.5% greater relative TCP than uncoupled RO plans when stopping powers were overestimated. The ideal RO implementation for head and neck IMPT is primarily dependent on the accuracy of the planned stopping power ratio to the true stopping power ratio of each beam. When planned stopping power is overestimated, RO with uncoupled uncertainty parameters outperform coupled uncertainties in terms of overall TCP and NTCP. The opposite is observed when planned stopping power is underestimated. High HU streaking from metal implants is the most common artifact that affects head and neck proton planning, likely leading to an overestimation of stopping powers, making uncoupled RO preferred for head and neck treatment planning.

Classification of the source of treatment deviation in proton therapy using prompt-gamma imaging information.

Prompt-gamma imaging (PGI)-based range verification has been successfully implemented in clinical proton therapy recently and its sensitivity to detect treatment deviations is currently investigated. The cause of treatment deviations can be multiple - for example, computed tomography (CT)-based range prediction, patient setup, and anatomical changes. Hence, it would be beneficial, if PGI-based verification would not only detect a treatment deviation but would also be able to directly identify its most probable source. Here, we propose a heuristically derived decision tree approach to automatically classify the sources of range deviation in proton pencil-beam scanning (PBS) treatments of head and neck tumors based on range information obtained with a PGI slit camera. The decision tree model was iteratively generated on a training dataset of different anatomical complexities, for an anthropomorphic head phantom and patient CT data (planning and control CTs) from five patients. Different range prediction errors, setup changes and relevant and nonrelevant anatomical changes were introduced or derived from control CTs, summing up to a total of 98 training scenarios. Independent validation was performed for another 98 scenarios, derived solely from patient CT data of another seven patients. PBS head and neck treatment plans were generated for the nominal scenario. For all PBS spots in the investigated field, PGI profiles were simulated using a dedicated analytical model of the slit camera for the nominal as well as the different error scenarios. From comparison of PGI profiles for nominal and error scenarios, a spot-wise range shift after spot aggregation with a kernel of 7mm sigma was determined for each error scenario. The heuristic approach includes a prefiltering of the most suitable PBS spots for PGI treatment verification. From the validation, the accuracy, sensitivity, and specificity of the model were determined. A five-step consecutive filter was developed to preselect PBS spots. On average, 25% of spots (1044 spots) remained as input for the classification model. The derived heuristic decision tree model is based on five parameters: The coefficient of determination (R2 ), the slope and intercept of the linear regression between PGI-derived range shifts and the respectively predicted proton ranges for the investigated PBS spots, as well as the average and standard deviation of the PGI-derived shifts. With this approach, 94 of 98 error scenarios could be classified correctly in validation (accuracy of 96%). A sensitivity and specificity of 100% and 86% were reached. In this simulation study it was demonstrated that the source of a treatment deviation can be identified from simulated noiseless PGI information in head and neck tumor treatments with high sensitivity and specificity. The application, refinement, and evaluation of the approach on measured PGI data will be the next step to show the clinical feasibility of PGI-based error source classification.

The Rainbow Scale for the Assessment of the Cervicomental Angle: A Validated Scale.

Aging of the neck results in an increased cervicomental angle, which can be treated by various surgical and nonsurgical procedures. To measure the success of these procedures, standardized validated objective photographic measurement tools are needed. However, no online standardized photographic measurement tools exist for the assessment of the cervicomental angle. The purpose of this study was to establish a validated and reliable measurement tool for the assessment of the cervicomental angle based on the Rainbow Scale. A 5-point photographic rating scale was developed and created from 1 photograph with Adobe Photoshop. Fifteen reference photographs of women, 3 photographs per grade, were included for validation. Seven panelists (ie, plastic and maxillofacial surgeons) assessed the reference photographs 3 times with a minimal interval of 3 days in an online survey. Intra- and inter-observer agreements were calculated utilizing the weighted kappa coefficient. Mean intra-observer agreement was 0.93 (0.78-1.00). Mean interobserver agreement was 0.796 (0.574-0.961) for survey 1, 0.868 (0.690-0.960) for survey 2, and 0.820 (0.676-0.959) for survey 3. The Rainbow Scale for the assessment of the cervicomental angle has been validated in an online fashion. The scale is reproducible and reliable and requires no learning curve. Potential applications include objective assessment of neck treatment planning and surgical outcome.

Knowledge-Based Statistical Inference Method for Plan Quality Quantification.

Aim:The aim of the study is to develop a geometrically adaptive and statistically robust plan quality inference method.Methods and Materials:We propose a knowledge-based plan quality inference method that references to similar plans in the historical database for patient-specific plan quality evaluation. First, a novel plan similarity metric with high-dimension geometrical difference quantification is utilized to retrieve similar plans. Subsequently, dosimetric statistical inferences are obtained from the selected similar plans. Two plan quality metrics—dosimetric result probability and dose deviation index—are proposed to quantify plan quality among prior similar plans. To evaluate the performance of the proposed method, we exported 927 clinically approved head and neck treatment plans. Eight organs at risk, including brain stem, cord, larynx, mandible, pharynx, oral cavity, left parotid and right parotid, were analyzed. Twelve suboptimal plans identified by dosimetric result probability were replanned to validate the capability of the proposed methods in identifying inferior plans.Results:After replanning, left and right parotid median doses are reduced by 31.7% and 18.2%, respectively; 83% of these cases would not be identified as suboptimal without the proposed similarity plan selection. Analysis of population plan quality reveals that average parotid sparing has been improving significantly over time (21.7% dosimetric result probability reduction from year 2006-2007 to year 2016-2017). Notably, the increasing dose sparing over time in retrospective plan quality analysis is strongly correlated with the increasing dose prescription ratios to the 2 planning targets, revealing the collective trend in planning conventions.Conclusions:The proposed similar plan retrieval and analysis methodology has been proven to be predictive of the current plan quality. Therefore, the proposed workflow can potentially be applied in the clinics as a real-time plan quality assurance tool. The proposed metrics can also serve the purpose of plan quality analytics in finding connections and historical trends in the clinical treatment planning workflow.

P148] Evaluation of delta4 phantom’s spatial sensitivity

Purpose The aim of this study is to evaluate the spatial sensitivity of the Delta4 diode array phantom (Scandidos, Uppsala, Sweden) investigating how set up errors affect the measurement’s results during head and neck treatment plan verification. Method The Delta4 phantom is used to verify ten IMRT head and neck plans, created on Oncentra v4.5.2 (Nucletron, Elekta) treatment planning system. Patient plans were transferred to the Delta4 geometry and were delivered on a 6MV photon beam SIEMENS Oncor Impression linac. Each reference (no-shift) plan was then delivered on the linac with a ± 3 mm shift of the phantom at lateral, longitudinal and vertical direction. The measurements’ results for all delivered plans were evaluated using the gamma index criteria (3%/3 mm). The application of the treatment plan is considered successful with an over 90% gamma index passing rate. Results The gamma index passing rate was measured at no-shift position (96.1 ± 0.7) %. Mean gamma index for each direction (lateral, longitudinal, vertical) was measured as 85.5% (SD = 2.6), 85.7% (SD = 4.5) and 85.2% (SD = 4.5) respectively. For all shifts, in all directions, the mean gamma index passing rate was found enough below the passing rate limit of 90% and more specifically 10.6% less than the no-shift measured gamma index. Conclusion Delta4 is an efficient QA tool for individual patient treatment plan verification exhibiting high reproducibility of results for γ -index. Shifts of 3 mm reveal significant deviation of measured 3%/3 mm gamma index passing rates, indicating high spatial sensitivity and efficiency of this QA tool. Future work with 3DVH software analysis is needed to report more statistics apart from γ -index values.

When and how to treat an IMRT patient on a second accelerator without replanning?

When a linear accelerator is unavailable for treatment, a clinical decision is imminent regarding whether a patient should be treated on a linear accelerator other than the machine the patient was scheduled on, or whether treatment should be postponed until the original Linac becomes available. This work investigates the feasibility of switching patients to different accelerators for intensity-modulated radiation therapy (IMRT). We have performed Monte Carlo simulations of photon beams from different Linac models and vendors. Prostate and head and neck (H&N) treatment plans for Siemens Primus, Primart, and Varian 21EX 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 dose-volume histograms (DVHs) and the maximum, the minimum, and the mean doses to the target and critical structures because of switching accelerators. In the process of switching a treatment plan to a different accelerator, issues exist, including optimum penumbra compensation, dose distribution at the boundary of target and critical structures, and multileaf collimator (MLC) leaf-width effects, which need to be considered and verified with measurements. Our Monte Carlo simulation results confirm that, for the cases we tested, the dose received by 95% of the planning target volume differs by 0.2% to 1.5% between Siemens Primus and Varian 21EX Linacs. The discrepancy is within our clinical acceptance criteria of 3% for IMRT treatments. In making the final decision on whether to switch machines or not, the tumor control probabilities (TCPs) based on a linear-quadratic model are compared. Based on the analyses performed 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 H&N cancers.

SU‐D‐BRB‐06: Toward Big‐Data Driven Treatment Planning: VMAT/IMRT Inverse Planning Autopiloted by Population‐Based Prior Data

Purpose:Current treatment planning remains a costly and labor intensive procedure and requires multiple trial‐and‐error adjustments of system parameters such as weighting factors and prescriptions. The purpose of this work is to develop an autonomous treatment planning technique with effective use of population‐based prior knowledge.Methods:Our autopiloted planning tool consists of three major components: (i) a commercial treatment planning system (TPS) (EclipseTM, Varian Medical Systems, CA); (ii) a statistical formulation of prior knowledge, which takes an assemble of prior treatment plans similar to the patient under planning and provides a priori variation range of final solution with an assigned preference level; and (iii) a decision‐making or outer‐loop optimization that assesses plan generated by Eclipse and drives the search toward a solution consistent with population‐based prior knowledge. To query and interact with TPS, Microsoft (MS) Visual Studio Coded UI is applied to record some common planner‐TPS interactions as subroutines. These subroutines are called back in the autopilot program written in C# to navigate through the solution space with an iterative algorithm. The utility of the approach is demonstrated by using a prostate IMRT case and three head and neck VMAT case.Results:An autonomous inverse planning formalism of using population‐based prior knowledge for autopiloting the VMAT/IMRT treatment planning process is implemented successfully in platform of a commercial TPS. The process mimics the decision‐making process of autopiloted automobile and provides a clinically sensible treatment plan automatically, thus effectively eliminating the tedious manual trial‐and‐errors process of treatment planning. It is found that the prostate and head and neck treatment plans compare favorably with that used for patients’ actual treatments.Conclusion:Clinical inverse treatment planning process can be autopiloted effectively with the use of population‐based prior knowledge. The strategy lays foundation for future development of big data‐driven inverse planning of various disease sites.Research grant and speak honoraria from Varian Medical Systems

Automatic planning of head and neck treatment plans.

Treatment planning is time‐consuming and the outcome depends on the person performing the optimization. A system that automates treatment planning could potentially reduce the manual time required for optimization and could also provide a method to reduce the variation between persons performing radiation dose planning (dosimetrist) and potentially improve the overall plan quality. This study evaluates the performance of the Auto‐Planning module that has recently become clinically available in the Pinnacle3 radiation therapy treatment planning system. Twenty‐six clinically delivered head and neck treatment plans were reoptimized with the Auto‐Planning module. Comparison of the two types of treatment plans were performed using DVH metrics and a blinded clinical evaluation by two senior radiation oncologists using a scale from one to six. Both evaluations investigated dose coverage of target and dose to healthy tissues. Auto‐Planning was able to produce clinically acceptable treatment plans in all 26 cases. Target coverages in the two types of plans were similar, but automatically generated plans had less irradiation of healthy tissue. In 94% of the evaluations, the autoplans scored at least as high as the previously delivered clinical plans. For all patients, the Auto‐Planning tool produced clinically acceptable head and neck treatment plans without any manual intervention, except for the initial target and OAR delineations. The main benefit of the method is the likely improvement in the overall treatment quality since consistent, high‐quality plans are generated which even can be further optimized, if necessary. This makes it possible for the dosimetrist to focus more time on difficult dose planning goals and to spend less time on the more tedious parts of the planning process.PACS number: 87.55.de

Effects of Proton Pencil Beam Scanning Spot Size on Head and Neck Treatment Plan Quality

Effects of Proton Pencil Beam Scanning Spot Size on Head and Neck Treatment Plan Quality

SU‐E‐T‐547: Modeling Biological Response to Proton Irradiation and Evaluating Its Potential Clinical Consequences

Purpose:In addition to physical uncertainties, proton therapy may also be associated with biologic uncertainties. Currently a generic RBE value of 1.1 is used for treatment planning. In this work the effects of variable RBE, in comparison to a fixed RBE, were evaluated by calculating the effective dose for proton treatments.Methods:The repair misrepair fixation (RMF) model was used to calculate variable proton RBEs. The RBE weighted spread‐out Bragg peak (SOBP) dose in water phantom was calculated using Monte Carlo simulation and compared to 1.1 weighted SOBP dose. A head and neck proton treatment was used to evaluate the potential effects, by comparing the head and neck treatment plan computed with a commercial treatment planning system that incorporates fixed RBE of 1.1 and a Monte Carlo treatment planning system that incorporates variable RBE.Results:RBE calculations along the depth of SOBP showed that the RBE at the entrance is approximately 1 and reaches 1.1 near the center of the SOBP. However, in distal regions the RBE rises to higher values (up to 3.5 depending on the cell type). Comparison of commercial treatment plans using a fixed RBE of 1.1 and Monte Carlo using variable RBE showed noticeable differences in the effective dose distributions.Conclusion:The comparison of the treatment planning with fixed and variable RBE shows that using commercial treatment planning systems that incorporate fixed RBE (1.1) could Result in overestimation of the effective dose to part of head and neck target volumes, while underestimating the effective dose to the normal tissue beyond the tumor. The accurate variable RBE as a function of proton beam energy in patient should be incorporated in treatment planning to improve the accuracy of effective dose calculation.

A fast GPU-based Monte Carlo simulation of proton transport with detailed modeling of nonelastic interactions.

Very fast Monte Carlo (MC) simulations of proton transport have been implemented recently on graphics processing units (GPUs). However, these MCs usually use simplified models for nonelastic proton-nucleus interactions. Our primary goal is to build a GPU-based proton transport MC with detailed modeling of elastic and nonelastic proton-nucleus collisions. Using the cuda framework, the authors implemented GPU kernels for the following tasks: (1) simulation of beam spots from our possible scanning nozzle configurations, (2) proton propagation through CT geometry, taking into account nuclear elastic scattering, multiple scattering, and energy loss straggling, (3) modeling of the intranuclear cascade stage of nonelastic interactions when they occur, (4) simulation of nuclear evaporation, and (5) statistical error estimates on the dose. To validate our MC, the authors performed (1) secondary particle yield calculations in proton collisions with therapeutically relevant nuclei, (2) dose calculations in homogeneous phantoms, (3) recalculations of complex head and neck treatment plans from a commercially available treatment planning system, and compared with (GEANT)4.9.6p2/TOPAS. Yields, energy, and angular distributions of secondaries from nonelastic collisions on various nuclei are in good agreement with the (GEANT)4.9.6p2 Bertini and Binary cascade models. The 3D-gamma pass rate at 2%-2 mm for treatment plan simulations is typically 98%. The net computational time on a NVIDIA GTX680 card, including all CPU-GPU data transfers, is ∼ 20 s for 1 × 10(7) proton histories. Our GPU-based MC is the first of its kind to include a detailed nuclear model to handle nonelastic interactions of protons with any nucleus. Dosimetric calculations are in very good agreement with (GEANT)4.9.6p2/TOPAS. Our MC is being integrated into a framework to perform fast routine clinical QA of pencil-beam based treatment plans, and is being used as the dose calculation engine in a clinically applicable MC-based IMPT treatment planning system. The detailed nuclear modeling will allow us to perform very fast linear energy transfer and neutron dose estimates on the GPU.

SU‐E‐T‐21: A Novel Sampling Algorithm to Reduce Intensity‐Modulated Radiation Therapy (IMRT) Optimization Time

Purpose:The IMRT optimization problem requires substantial computer time to find optimal dose distributions because of the large number of variables and constraints. Voxel sampling reduces the number of constraints and accelerates the optimization process, but usually deteriorates the quality of the dose distributions to the organs. We propose a novel sampling algorithm that accelerates the IMRT optimization process without significantly deteriorating the quality of the dose distribution.Methods:We included all boundary voxels, as well as a sampled fraction of interior voxels of organs in the optimization. We selected a fraction of interior voxels using a clustering algorithm, that creates clusters of voxels that have similar influence matrix signatures. A few voxels are selected from each cluster based on the pre‐set sampling rate.Results:We ran sampling and no‐sampling IMRT plans for de‐identified head and neck treatment plans. Testing with the different sampling rates, we found that including 10% of inner voxels produced the good dose distributions. For this optimal sampling rate, the algorithm accelerated IMRT optimization by a factor of 2–3 times with a negligible loss of accuracy that was, on average, 0.3% for common dosimetric planning criteria.Conclusion:We demonstrated that a sampling could be developed that reduces optimization time by more than a factor of 2, without significantly degrading the dose quality.

SU-E-T-379: Concave Approximations of Target Volume Dose Metrics for Intensity- Modulated Radiotherapy Treatment Planning

Purpose: The widely used treatment plan metric Dx (mimimum dose to the hottest x% by volume of the target volume) is simple to interpret and use, but is computationally poorly behaved (non-convex), this impedes its use in computationally efficient intensity-modulated radiotherapy (IMRT) treatment planning algorithms. We therefore searched for surrogate metrics that are concave, computationally efficient, and accurately correlated to Dx values in IMRT treatment plans. Methods: To find concave surrogates of D95—and more generally, Dx values with variable x values—we tested equations containing one or two generalized equivalent uniform dose (gEUD) functions. Fits were obtained by varying gEUD ‘a’ parameter values, as well as the linear equation coefficients. Fitting was performed using a dataset of dose-volume histograms from 498 de-identified head and neck IMRT treatment plans. Fit characteristics were tested using a crossvalidation process. Reported root-mean-square error values were averaged over the cross-validation shuffles. Results: As expected, the two-gEUD formula provided a superior fit, compared to the single-gEUD formula. The best approximation uses two gEUD terms: 16.25 x gEUD[a=0.45] – 15.30 x gEUD[a=1.75] – 0.69. The average root-mean-square error on repeated (70/30) cross validation was 0.94 Gy. In addition, a formula was found that reasonably approximates Dx for xmore » between 80% and 96%. Conclusion: A simple concave function using two gEUD terms was found that correlates well with PTV D95s for these head and neck treatment plans. More generally, a formula was found that represents well the Dx for x values from 80% to 96%, thus providing a computationally efficient formula for use in treatment planning optimization. The formula may need to be adjusted for other institutions with different treatment planning protocols. We conclude that the strategy of replacing Dx values with gEUD-based formulas is promising.« less

SU-E-T-155: Evaluating IMRT and Volumetric Arc Therapy Patient-Specific Quality Assurance by Using Octavius 4D System

Purpose: The goal of this study was to perform and evaluate pre-treatment patient specific verifications of IMRT and volumetric arc therapy(VMAT) plans by using a new, commercially available 2D array Octavius 4D system. Methods: The OCTAVIUS 4D system and its accompanying software (VERISOFT) by PTW were evaluated for the accuracy of the dose reconstruction for patient specific QA. (1)Commissioning of Octavius 4D system for IMRT and VMAT technique was by following the tests suggested in the AAPM TG119 protocol, in order to give a benchmark for optimization process and the calculation accuracy. (2) Verification of total 87 clinical IMRT and VMAT plans, 50 IMRT and 37 VMAT plans were used for the comparison. Comparing treatment planning system calculated versus measured dose using local gamma index analysis (3%, 3 mm criteria). Results: (1)Agreement between calculated and planar doses measured by using Octavius 4D system. It was very good for both modalities. (2)Verifications of pre-treatment clinical plans, showed an average gamma evaluation passing ratio of 97.3±1.9% for 50 IMRT cases and an average pass ratio of 94.94±2.43% for 37 VMAT cases. 2D gamma evaluation for IMRT plans showed 96.82%, 97.5%, 97.93% and 96.94% passing results for head and neck, chest, abdomen and Pelvis sites. Volumetric global gamma evaluation for VMAT plans showed 93.8%, 94.66% and 96.27% passing results for chest, abdomen and Pelvis sites respectively. Conclusion: The TG119 results suggest that it is appropriate to apply patient-specific QA action levels derived from IMRT to VMAT technique. Octavius 4D system together with 2D array 729 is an appropriate and efficient tool for IMRT and VMAT patient specific QA and can potentially improve the data collection and analysis. Further studies will focus on more complex VMAT plans (head and neck treatment plans).