Y Na

Accuracy of surface registration compared to conventional volumetric registration in patient positioning for head-and-neck radiotherapy: A simulation study using patient data

PURPOSE 3D optical surface imaging has been applied to patient positioning in radiation therapy (RT). The optical patient positioning system is advantageous over conventional method using cone-beam computed tomography (CBCT) in that it is radiation free, frameless, and is capable of real-time monitoring. While the conventional radiographic method uses volumetric registration, the optical system uses surface matching for patient alignment. The relative accuracy of these two methods has not yet been sufficiently investigated. This study aims to investigate the theoretical accuracy of the surface registration based on a simulation study using patient data. METHODS This study compares the relative accuracy of surface and volumetric registration in head-and-neck RT. The authors examined 26 patient data sets, each consisting of planning CT data acquired before treatment and patient setup CBCT data acquired at the time of treatment. As input data of surface registration, patients skin surfaces were created by contouring patient skin from planning CT and treatment CBCT. Surface registration was performed using the iterative closest points algorithm by point-plane closest, which minimizes the normal distance between source points and target surfaces. Six degrees of freedom (three translations and three rotations) were used in both surface and volumetric registrations and the results were compared. The accuracy of each method was estimated by digital phantom tests. RESULTS Based on the results of 26 patients, the authors found that the average and maximum root-mean-square translation deviation between the surface and volumetric registrations were 2.7 and 5.2 mm, respectively. The residual error of the surface registration was calculated to have an average of 0.9 mm and a maximum of 1.7 mm. CONCLUSIONS Surface registration may lead to results different from those of the conventional volumetric registration. Only limited accuracy can be achieved for patient positioning with an approach based solely on surface information.

Automatic deformable surface registration for medical applications by radial basis function-based robust point-matching

Deformable surface mesh registration is a useful technique for various medical applications, such as intra-operative treatment guidance and intra- or inter-patient study. In this paper, we propose an automatic deformable mesh registration technique. The proposed method iteratively deforms a source mesh to a target mesh without manual feature extraction. Each iteration of the registration consists of two steps, automatic correspondence finding using robust point-matching (RPM) and local deformation using a radial basis function (RBF). The proposed RBF-based RPM algorithm solves the interlocking problems of correspondence and deformation using a deterministic annealing framework with fuzzy correspondence and RBF interpolation. Simulation tests showed promising results, with the average deviations decreasing by factors of 21.2 and 11.9, respectively. In the human model test, the average deviation decreased from 1.72±1.88mm to 0.57±0.66mm. We demonstrate the effectiveness of the proposed method by presenting some medical applications.

SU-D-BRD-01: Cloud-Based Radiation Treatment Planning: Performance Evaluation of Dose Calculation and Plan Optimization.

PURPOSE To report the first experience on the development of a cloud-based treatment planning system and investigate the performance improvement of dose calculation and treatment plan optimization of the cloud computing platform. METHODS A cloud computing-based radiation treatment planning system (cc-TPS) was developed for clinical treatment planning. Three de-identified clinical head and neck, lung, and prostate cases were used to evaluate the cloud computing platform. The de-identified clinical data were encrypted with 256-bit Advanced Encryption Standard (AES) algorithm. VMAT and IMRT plans were generated for the three de-identified clinical cases to determine the quality of the treatment plans and computational efficiency. All plans generated from the cc-TPS were compared to those obtained with the PC-based TPS (pc-TPS). The performance evaluation of the cc-TPS was quantified as the speedup factors for Monte Carlo (MC) dose calculations and large-scale plan optimizations, as well as the performance ratios (PRs) of the amount of performance improvement compared to the pc-TPS. RESULTS Speedup factors were improved up to 14.0-fold dependent on the clinical cases and plan types. The computation times for VMAT and IMRT plans with the cc-TPS were reduced by 91.1% and 89.4%, respectively, on average of the clinical cases compared to those with pc-TPS. The PRs were mostly better for VMAT plans (1.0 ≤ PRs ≤ 10.6 for the head and neck case, 1.2 ≤ PRs ≤ 13.3 for lung case, and 1.0 ≤ PRs ≤ 10.3 for prostate cancer cases) than for IMRT plans. The isodose curves of plans on both cc-TPS and pc-TPS were identical for each of the clinical cases. CONCLUSION A cloud-based treatment planning has been setup and our results demonstrate the computation efficiency of treatment planning with the cc-TPS can be dramatically improved while maintaining the same plan quality to that obtained with the pc-TPS. This work was supported in part by the National Cancer Institute (1R01 CA133474) and by Leading Foreign Research Institute Recruitment Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (MSIP) (Grant No.2009-00420).

SU-E-J-217: Accuracy Comparison Between Surface and Volumetric Registrations for Patient Setup of Head and Neck Radiation Therapy

PURPOSE Optical surface imaging has been applied to radiation therapy patient setup. This study aims to investigate the accuracy of the surface registration of the optical surface imaging compared with that of the conventional method of volumetric registration for patient setup in head and neck radiation therapy. METHODS Clinical datasets of planning CT and treatment Cone Beam CT (CBCT) were used to compare the surface and volumetric registrations in radiation therapy patient setup. The Iterative Closest Points based on point-plane closest method was implemented for surface registration. We employed 3D Slicer for rigid volumetric registration of planning CT and treatment CBCT. 6 parameters of registration results (3 rotations and 3 translations) were obtained by the two registration methods, and the results were compared. Digital simulation tests in ideal cases were also performed to validate each registration method. RESULTS Digital simulation tests showed that both of the registration methods were accurate and robust enough to compare the registration results. In experiments with the actual clinical data, the results showed considerable deviation between the surface and volumetric registrations. The average root mean squared translational error was 2.7 mm and the maximum translational error was 5.2 mm. CONCLUSION The deviation between the surface and volumetric registrations was considerable. Special caution should be taken in using an optical surface imaging. To ensure the accuracy of optical surface imaging in radiation therapy patient setup, additional measures are required. This research was supported in part by the KIST institutional program (2E24551), the Industrial Strategic technology development program (10035495) funded by the Ministry of Trade, Industry and Energy (MOTIE, KOREA), and the Radiation Safety Research Programs (1305033) through the Nuclear Safety and Security Commission, and the NIH (R01EB016777).

SU‐E‐T‐651: A Novel Non‐Uniformly Distributed Control Points‐Based Algorithm for VMAT Treatment Plan Optimization

PURPOSE To develop a new method for adding control points to volumetric modulated arc therapy (VMAT) optimization in non-uniformly distributed scheme that improves the delivery efficiency and dose distribution. METHODS The algorithm generates piecewise constant fluence maps using total-variation regulation (TVR) for each beam direction. The optimized fluence maps of one fluence profile per the initial arc spacing of 6° are distributed as coarse sampling of control points over a single arc, 360°. From the initial control points having an aperture sequence {A(0), …, A(N)}, each association {A(0)-A(1), …, A(N-1)-A(N)} of apertures is assigned a mutual information(MI) score. A new control point (B0) is added to the mid-point of two apertures {A(0)-B(0)-A(1)} in which the {A(0)-A(1)} association has the lowest MI score having the biggest aperture shape difference. A new shape of aperture at B(0) is created by the interpolation of the adjacent apertures. The aperture shape is then rectified to be a sequencing of arc using the manufacture constraints. The implemented plan optimization is evaluated using clinical cases of head neck, lung, and prostate previously treated with a Varian TrueBeam™STX linac. RESULTS The propose method with non-uniformly distributed 120 control points reduced the dose delivery time approximately 32%, 26%, and 30% compared to the uniformly distributed 180 control points for head neck, lung, and prostate plans, respectively. The target dose coverage and critical structure sparing of plans obtained using the proposed method with non-uniformly distributed 120 control points are almost identical to those obtained using the conventional method with 180 control points uniformly distributed. CONCLUSION The results demonstrate that non-uniformly distributed control points-based VMAT plan optimization reduced the dose delivery time and slightly improved dose conformity. This new algorithm can substantially provide future station mediated planning parameters. This work was supported in part by Korean grants from the National Research Foundation of Korea (NRFK) of the Korea government (MEST) (No.K20901000001-09E0100-00110) and the National Cancer Institute (1R01 CA133474).

SU‐E‐T‐609: Improving the Efficiency of VMAT Plan Optimization by Using Sparse Decomposition Method

PURPOSE Volumetric modulated arc therapy (VMAT) is capable of delivering highly conformable dose distribution efficiently. Its planning is, however, more computationally intensive and requires a huge amount of memory space for optimization. We present an efficient sparse decomposition method for VMAT plan optimization. METHODS A quadratic objective function with volumetric constraints is expressed as a function of the aperture shapes and weights of the incident beams. The algorithm generates a sequence of iterates to solve the optimization problem. Each step of iteratively reweighed method is to be updated by solving the subproblem involving a quadratic (L2) term and a sparsity-inducing regulation (L1) term. Through the sparse decomposition techniques of the given problem, the deliverable apertures are directly generated. The shape of each aperture is iteratively rectified to be a sequencing of arc using the manufacture constraints. An initial arc spacing of 8 degree creates 45 beams directions for a single arc, 360 degree. The angular separation is equispaced every 2 degree over the end of optimization cycle. The optimization is implemented for a Varian TrueBeamTM STX linac beams with and without flattening filters available. Three clinical cases, head and neck, lung, and prostate, have been studied for the purpose of evaluating the planning efficiency and quality of the plans. RESULTS The target dose coverage and critical structure sparing of VMAT plan are comparable to those of IMRT plans. The VMAT plan delivers lower doses to other OARs while keeping the similar target dose coverage to IMRT plan. The VMAT plan optimizations takes less than 3 minutes on average of the cases indicating great efficiency compared to existing methods. CONCLUSIONS The results demonstrate that the proposed method provides competent computational efficiency for optimizing VMAT plan. The method substantially improves the speed and accuracy of VMAT plan optimization and makes future on-treatment adaptive re-planning possible.

SU‐E‐T‐628: A Cloud Computing Based Multi‐Objective Optimization Method for Inverse Treatment Planning

PURPOSE Multi-objective (MO) plan optimization entails generation of an enormous number of IMRT or VMAT plans constituting the Pareto surface, which presents a computationally challenging task. The purpose of this work is to overcome the hurdle by developing an efficient MO method using emerging cloud computing platform. METHODS As a backbone of cloud computing for optimizing inverse treatment planning, Amazon Elastic Compute Cloud with a master node (17.1 GB memory, 2 virtual cores, 420 GB instance storage, 64-bit platform) is used. The master node is able to scale seamlessly a number of working group instances, called workers, based on the user-defined setting account for MO functions in clinical setting. Each worker solved the objective function with an efficient sparse decomposition method. The workers are automatically terminated if there are finished tasks. The optimized plans are archived to the master node to generate the Pareto solution set. Three clinical cases have been planned using the developed MO IMRT and VMAT planning tools to demonstrate the advantages of the proposed method. RESULTS The target dose coverage and critical structure sparing of plans are comparable obtained using the cloud computing platform are identical to that obtained using desktop PC (Intel Xeon® CPU 2.33GHz, 8GB memory). It is found that the MO planning speeds up the processing of obtaining the Pareto set substantially for both types of plans. The speedup scales approximately linearly with the number of nodes used for computing. With the use of N nodes, the computational time is reduced by the fitting model, 0.2+2.3/N, with r̂2>0.99, on average of the cases making real-time MO planning possible. CONCLUSIONS A cloud computing infrastructure is developed for MO optimization. The algorithm substantially improves the speed of inverse plan optimization. The platform is valuable for both MO planning and future off- or on-line adaptive re-planning.

SU‐E‐T‐775: Toward Web‐Based Real‐Time VMAT and IMRT Inverse Planning Using Cloud Computing

Purpose: To provide an efficient cloud computing infrastructure for VMAT and IMRT inverse planning. Methods: Amazon Elastic Compute Cloud (EC2) with a master node (17.1 GB memory, 2 virtual cores, 420 GB instance storage, 64‐bit platform) is used as the backbone of cloud computing for VMAT and IMRT dose optimization. The master node is scalable on an “on‐demand” basis. It can lunch more workers if the CPU usage exceeds the upper limit for given time (e.g., over 90% usage for 2 minutes), or terminate the workers if there are finished tasks. A quadratic objective function with volumetric constraints is expressed as a function of the aperture shapes and weights of the incident beams. A Monte Carlo simulation is employed to generate accurate dose kernel by parallel tasks. The number of parallel tasks is approximately the same as or less than the total number of beamlets in iterative simulation for all beams. The cloud optimization is implemented for a Varian TrueBeam STX linac beams with and without flattening filters available. The filed output from EC2 is sent down to the Simple Storage Service (S3). Three clinical cases have been studied for the purpose of evaluating the performance of the new planning platform. Results: A cloud computing environment leads to speedups of 150–300 times for the cases considered in this study. The speedup scales approximately linearly with the number of nodes used for computing. The resultant plans from the cloud computing are found almost identical to that obtained using a desktop PC, indicating the reliability of the cloud computing platform. Conclusions: A cloud computing infrastructure has been established for VMAT and IMRT inverse planning with different beam characteristics. The cloud computing environment substantially improves the speed of inverse planning and makes future on‐treatment adaptive replanning possible. This work was supported by grants from the National Research Foundation of Korea (NRFK) of the Korea Government (MEST) (Grant No. K20901000001‐09E0100‐ 00110)