S Brame

Technical Note: DIRART- A software suite for deformable image registration and adaptive radiotherapy research

Purpose: Recent years have witnessed tremendous progress in image guide radiotherapy technology and a growing interest in the possibilities for adapting treatment planning and delivery over the course of treatment. One obstacle faced by the research community has been the lack of a comprehensive open-source software toolkit dedicated for adaptive radiotherapy (ART). To address this need, the authors have developed a software suite called the Deformable Image Registration and Adaptive Radiotherapy Toolkit (DIRART). Methods: DIRART is an open-source toolkit developed in MATLAB. It is designed in an object-oriented style with focus on user-friendliness, features, and flexibility. It contains four classes of DIR algorithms, including the newer inverse consistency algorithms to provide consistent displacement vector field in both directions. It also contains common ART functions, an integrated graphical user interface, a variety of visualization and image-processing features, dose metric analysis functions, and interface routines. These interface routines make DIRART a powerful complement to the Computational Environment for Radiotherapy Research (CERR) and popular image-processing toolkits such as ITK. Results :DIRART provides a set of image processing/registration algorithms and postprocessing functions to facilitate the development and testing of DIR algorithms. It also offers a good amount of options for DIR results visualization, evaluation, and validation. Conclusions : By exchanging data with treatment planning systems via DICOM-RT files andCERR, and by bringing image registration algorithms closer to radiotherapy applications, DIRART is potentially a convenient and flexible platform that may facilitate ART and DIR research.

Standardizing Naming Conventions in Radiation Oncology

PURPOSE The aim of this study was to report on the development of a standardized target and organ-at-risk naming convention for use in radiation therapy and to present the nomenclature for structure naming for interinstitutional data sharing, clinical trial repositories, integrated multi-institutional collaborative databases, and quality control centers. This taxonomy should also enable improved plan benchmarking between clinical institutions and vendors and facilitation of automated treatment plan quality control. MATERIALS AND METHODS The Advanced Technology Consortium, Washington University in St. Louis, Radiation Therapy Oncology Group, Dutch Radiation Oncology Society, and the Clinical Trials RT QA Harmonization Group collaborated in creating this new naming convention. The International Commission on Radiation Units and Measurements guidelines have been used to create standardized nomenclature for target volumes (clinical target volume, internal target volume, planning target volume, etc.), organs at risk, and planning organ-at-risk volumes in radiation therapy. The nomenclature also includes rules for specifying laterality and margins for various structures. The naming rules distinguish tumor and nodal planning target volumes, with correspondence to their respective tumor/nodal clinical target volumes. It also provides rules for basic structure naming, as well as an option for more detailed names. Names of nonstandard structures used mainly for plan optimization or evaluation (rings, islands of dose avoidance, islands where additional dose is needed [dose painting]) are identified separately. RESULTS In addition to its use in 16 ongoing Radiation Therapy Oncology Group advanced technology clinical trial protocols and several new European Organization for Research and Treatment of Cancer protocols, a pilot version of this naming convention has been evaluated using patient data sets with varying treatment sites. All structures in these data sets were satisfactorily identified using this nomenclature. CONCLUSIONS Use of standardized naming conventions is important to facilitate comparison of dosimetry across patient datasets. The guidelines presented here will facilitate international acceptance across a wide range of efforts, including groups organizing clinical trials, Radiation Oncology Institute, Dutch Radiation Oncology Society, Integrating the Healthcare Enterprise, Radiation Oncology domain (IHE-RO), and Digital Imaging and Communication in Medicine (DICOM).

WE‐C‐214‐04: ADQ — A Software Tool That Automatically, Autonomously, Intelligently and Instantly Verify Patient Radiation Therapy Beam Deliveries

Purpose: To effectively mitigate errors in IMRTradiation therapydelivery, all beams, all fractions, for all patients should be checked in vivo, immediately, automatically, autonomously and intelligently for integrity, quality and safety. For this purpose, ADQ or Automatic Dynalog QA, is implemented for instant and automatic patient delivery beam verification. Methods: ADQ contains multiple functional modules. DICOM receiver program is developed in C++ to receive DICOM‐RT Plans from treatment planning systems, process data and save to database. The verification tool is implemented in MATLAB that automatically validates beam parameters (gantry angle, collimator angle and positions, MLC positions, fluence maps, etc.) between the treatment plans and recorded dynamic MLC log files, generate reports for each treatment session, and send out alert emails for detected urgent problems. Report reviewer is implemented in C++ that enables physicists to review, comment and confirm reports. ADQ programs use own database and Mosaiq R&V database. Simple, automatic and no human intervention is needed unless an error is detected. Results: ADQ is running to generate near real‐time QA reports for every treatment date. DICOM receiver is running 24 hours to collect plans. Report reviewer deployed through network facilitates easy access to reports. All IMRT beams delivered to the patients are checked for a period of four months to study the reliability, MLC performance, false positive rate and importantly identify true positive. More than 80000 thousand beams from 4 different Linacs were analyzed up‐to‐date. Conclusions: We developed new software tools to improve the RT treatment QA by automatic checking patient treatment beam delivery records for each patient and each treatment session. Report data achieved in database can be easily used for further studies, for example, analysis of MLC leaf failures.

WE‐E‐BRB‐10: DosCheck ‐ an Electronic Chart Checking Tool for Dosimetrists

PURPOSE In addition to treatment planning, dosimetrists have to prepare documentation and manually enter data in treatment management system (TMS) which did not transfer or setup automatically. The required documents and data are dependent on the disease site, treatment machine and clinical workflow. Errors and inconsistencies can cause redundant work, treatment delays and potentially treatment errors. To address these issues, an electronic checking software tool, DosCheck was clinically implemented to check the existence of necessary documentations and the integrity of manually-entered data. The purpose of this software is to reduce the frequency of human errors and to improve efficiency. METHODS DosCheck reads data and documents from 1) TMS, 2) Pinnacle TPS, and 3) DICOM plan files stored in a DICOM-RT PACS. It processes documents in Word and PDF format, treatment plan data in Pinnacle native format and DICOM format, and Mosaiq data in database records. The software cross-checks data accuracy and consistency by following rules that are pre-defined according to the clinical requirements and treatment sties. It interacts with dosimetrists and presents instantaneous results via graphical user interface. RESULTS DosCheck has been implemented in C#. It performs a full check for a patient with 20 seconds. It has been clinically commissioned and is used daily by all dosimetrists at our institution. Retrospective analysis shows that DosCheck identifies 30% to 40% of previously reported dosimetrist human errors. Additional ∼30% errors are checked by other tools that could be integrated DosCheck in the near future. CONCLUSIONS As an electronic data checking tool, DosCheck can obtain and process data and documents from multiple clinical computer systems in the radiation oncology department, and perform checks according to clinical rules. It is able to improve the accuracy and efficiency of clinical data and document process, and therefore to reduce any potential inconsistencies and errors.

SU-E-J-68: Plan Rejections during Treatment Planning: A Part of Normal Process Or An Opportunity for Improvement

Purpose: To describe a process for learning from radiation treatment plan(RTP) rejections by physicians and physicists for improvement of quality, safety, and efficiency of treatment planning. In our clinic IMRTtreatment plans are pre‐checked by physicists for technical and quality validation before review by the physician. Any problems identified during physics or physician reviews of RTPs offer an opportunity to identify systematic and random deviations and to develop tools, processes, and training to eliminate or reduce such problems in the future. We describe a process wherein the collection of plan rejections and related reasons is considered a part of normal process and not as additional work. Methods: In our clinic we use the DMAIC formalism for management of our operations, whereby processes are understood as comprising of define, measure, analyze, improve and control activities. Systematically quantifying plan rejections and the corresponding reasons offers an opportunity for improvement of treatment planning process and monitoring the effectiveness of corrective measures. This was accomplished through a set of features in our electronic whiteboards that seamlessly and effectively collect rejection information that were further analyzed for identification of improvement opportunities. Efficacy of any implemented improvement measures can then also be evaluated by quantifying future rejections. In this study we concentrate on plan rejections by physicists. Results: There were 110 plan rejections by physicists from September 2010 to Mar 2011. The rejections were due to plan‐quality, plan technical integrity and in some cases data‐entry errors. These rejections were used to identify systematic opportunities for training and better education and for design of in‐house developed auto‐planning and quality‐control scripts. Conclusions: Systematic and sustainable collection of treatment plan rejections for purposes of process improvement is possible. We have demonstrated that this data collection can be seamlessly integrated in normal treatment planning process through the use of electronic white boards.

WE-C-214-03: ECCK - A Computer System of Automate Patient Chart Checking

Purpose: ECCK (Electronic Chart ChecKing) is a clinical computer system(software and database) designed to improve quality, efficiency, and frequency of patient chart checking. The goal is to use computer programs to automatically review patient treatment data and to highlight discrepancies and potential concerns through color‐coded reports and email messages. Methods: ECCK contains a dedicated PC workstation, a separate database server and a data storage server. It integrates many different technologies, including SQL database, data structure design, DICOM, PACS, dynamic HTML, javascript, MATLAB, C++ and image processing. ECCK contains 200+ MATLAB programs, many C++ programs, DHTML design templates, style sheets, javascript programs and design documents. Most ECCK programs run on the ECCK workstation. Patient data are obtained from the treatment planning system and oncology information system (OIS) server in both native format and DICOM format, and analyzed according to predefined rules with logical conditions. Reports, generated in DHTML format, are stored on the storage server and indexed in database. A standalone report‐browser program is run by users on any computer in the department. Other features include OIS database querying and offline document browsing. Results: ECCK functionality is purpose dependent and varies for therapists, physicians, and physicists. For physics, ECCK checks patient prescriptions, beam parameters, setup parameters, fractionation schedules, treatment calendar, treatment records, beam delivery history, existing of mandatory documents and images, etc. Conclusions: ECCK is useful to reduce the repetitive routine work of users, and to improve the speed and accuracy of chart checking. ECCK enables increased frequency of patient chart audits, and allows verification tasks which are not easily performed by humans. While human is the ultimate expert, ECCK ensures a consistent amount of quality in the chart checking process which is otherwise impossible to maintain.

WE‐C‐214‐08: Benefit of an Industrial Engineering Tool (Statistical Process Control) in Radiation Oncology: Clinical Experience

Purpose: 1. Develop a tool for data collection, retrieval, visualization, real‐time feedback and advanced statistical analysis of IMRT QA process. 2. Evaluate the need for site‐specific tolerances for IMRT QA. 3. Timely identify systematic drifts in the IMRT QA measurements. Methods: Statistical process control (SPC) is a statistical tool that can be used to study process variations and evaluate process capability. This tool is applicable to IMRT QA processes. There are two causes of variation present in a process — common cause, which is the reason for random variation in the process, and special (assignable) cause, which is the reason for identifiable variation in the process. Process capability and control charts (S chart for process variability and EWMA chart to study the process mean) are two of the several tools available to study the process of interest. We use SPC and paperless environment for collection and processing of clinical IMRT QA data to create a tool for improved management of IMRT QA. Results: Automatic data collection, data validation and visualization are facilitated with the help of the developed tool. Data visualization and control charts are used to study the process variation and identify drifts and outliers. Process capability studies identify systematic variations present in the process and evaluate the need for site‐specific tolerances. For instance, process capability studies of the QA results for treatment plans using 6 MV X‐rays revealed a 1.1% variation in calibration and 0.5% variation in beam model and MLC leaf model. Conclusion: We have built a tool that collects data, analyzes them, plots S chart for process variability and EWMA charts. Process capability studies show that using a global tolerance for all different sites is not sufficient. Hence, we propose the use of 2 standard deviations for a 95% confidence level in the IMRT QA measurements.

TH‐D‐BRD‐03: Experience with Error Reporting and Tracking Database Tool for Process Improvement in Radiation Oncology

Purpose: To present long‐term results of systematic near‐miss and actual error reporting and analysis system based on a web‐based tool and effects of formal process improvement structure on error rates and safety culture in radiation therapy (RT). Materials and Methods: An internally developed web‐based system was used to report, track, and analyze errors and near‐misses in a large RT department for almost two years. The system was designed as an efficient and effective process for collecting, storing, and analyzing the failure rate data in individual RT facilities. The aim of the system was to support process improvement in patient care and safety. The reporting tool was designed so individual events could be reported in as little as two minutes. Events were categorized based on functional area, type, and severity of failure. The events were processed and analyzed by a formal process improvement group which used the data and statistics collected through the web‐based tool for guidance in reengineering clinical processes. The results for the first nineteen months of clinical use of the system are presented. Results: The collected data and the process improvement structure resulted in measureable safety and error rate improvements in several clinical areas. The collected data was also very effective in identifying ineffective measures and efforts which did not produce improvements in clinical processes. The overall process demonstrated that it was possible to establish and maintain a high functioning safety culture in radiation oncology. The error reporting compliance, though voluntary, was very high and consistent from the inception of the process through the date of this report. Conclusions: Near‐miss and actual error collection process in RT can result in quantifiable safety and error rate improvements and more importantly it can result in a sustainable safety culture.