A 'learning-by-doing' treatment planning tutorial for medical physicists.

A framework for a tutorial for treatment planning in radiation oncology physics was developed, based on the University of Washington treatment planning system Prism. The tutorial is aimed at students in Medical Physics to accompany the lectures on treatment planning to enhance their theoretical knowledge. A web-based layout was chosen to allow independent work of the students. The tutorial guides the students through three different learning modules, designed mainly to enhance their understanding of the processes involved in treatment planning but also to learn the specific features of a modern treatment planning system. Each of the modules contains four units, with the aim to introduce the relevant Prism features, practice skills in different tasks and finally check the learning outcomes with a challenge and a self-scoring quiz. A survey for students' feedback completes the tutorial. Various tools and learning methods help to create an interactive, appealing learning environment, in which the emphasis is shifted from teacher-centred to student-centred learning paradigms. In summary, Prism lends itself well for educational purposes. The tutorial covers all main aspects of treatment planning. In its current form the tutorial is self-contained but still adjustable and expandable. The tutorial can be made available upon request to the authors.

A comparison of two radiological path length algorithms.

Most radiation therapy dose calculation methods require the determination of the effective path length of the primary radiation from the radiation source to the point at which the dose is calculated. This usually involves representing the patient anatomy as a set of polygons (contours) as approximations to plane curves. Several algorithms are known for determining the length of a segment or segments on a ray through a planar contour, that are interior to the contour. We have implemented two of these algorithms in a test program to benchmark their relative efficiency. One algorithm uses a linear search over all the contour segments, and the other method represents the contour as a binary tree of "strips," of successively increasing resolution. In general, the tree search should give times proportional to log(n) where n is the number of contour segments, and the linear search time should be proportional to n. Thus, one might expect the tree search to run faster once the number of segments reaches some sufficiently large value. We found that this value is a number of contour points far in excess of that typical for contours representing radiation therapy patient anatomy. Therefore, for this application the linear search method is more efficient.

Integration of radiotherapy planning systems and radiotherapy treatment equipment: 11 years experience.

PURPOSE: We have investigated the requirements, design, implementation, and operation of a computer-controlled medical accelerator with multileaf collimator (MLC), integrated with a radiation treatment-planning system (RTPS), and we report on the performance, benefits, and lessons learned from this experience. METHODS AND MATERIALS: In 1984 the University of Washington installed a computer-controlled radiation therapy machine (the Clinical Neutron Therapy System, or CNTS) with a multileaf collimator. Since the beginning of operation the control system computer has been connected by commercially available network hardware and software to three generations of radiation treatment-planning systems. Semiautomated setup and completely computerized check and confirm were incorporated into the system from the beginning of clinical operation in 1984. The system cannot deliver a patient treatment without a computer-prepared treatment plan. RESULTS: The CNTS has been in use for routine patient treatments for over 11 years. The cost of the network connection and software was an insignificant fraction of the facility cost. Operation has been efficient and reliable. Of the 441 machine-related session reschedulings (out of 18,432 sessions total) during the past 9 years, only 20 were due to problems with data transfer between the RTPS and CNTS, associated primarily with two incidents. Close integration with the treatment-planning system allows complex treatments to be delivered. Dramatic evolution of the departmental treatment-planning system has not required any changes or redesign of either the accelerator control system or the network connection. CONCLUSIONS: Our experience shows that a large degree of automation is possible with reasonable effort, by using well-known software and hardware design strategies. The lessons we have learned from this can be carried over into photon therapy now that photon accelerators with MLC facilities are commercially available.

Consistency of three-dimensional planning target volumes across physicians and institutions.

PURPOSE: Three-dimensional treatment planning depends upon exact and consistent delineation of target volumes. This study tested whether different physicians from different institutions vary significantly in their creation of planning target volumes (PTVs). METHODS AND MATERIALS: Eight physicians from three different institutions created partial planning target volumes for nine clinical test cases. Their target volumes were evaluated qualitatively and quantitatively. Quantitative results were tested for significant differences. RESULTS: Qualitative analysis showed the physicians to vary in (a) the margin placed around the clinical target volume, (b) the margin used near critical structures, and (c) handling of concavities in the clinical target volume. Quantitative analysis showed these variations to result in statistically significant differences in the measured volume of the physicians' planning target volumes. CONCLUSIONS: Individual physicians and institutions differ significantly in their creation of planning target volumes, suggesting individual and institutional differences in the working definition for the PTV. Implications of this fact are discussed, along with areas where standardization can be improved.

Three dimensional planning target volumes: a model and a software tool.

PURPOSE: Three dimensional (3D) target volumes are an essential component of conformal therapy because the goal is to shape the treatment volume to the target volume. The planning target volume (PTV) is defined by ICRU 50 as the clinical target volume (CTV) plus a margin to ensure that the CTV receives the prescribed dose. The margin must include all interfractional and intrafractional treatment variations. This paper describes a software tool that automatically generates 3D PTVs from CTVs for lung cancers and immobile head and neck cancers. METHODS AND MATERIALS: Values for the interfractional and intrafractional treatment variations were determined by a literature review and by targeted interviews with physicians. The software tool is written in Common LISP and conforms to the specifications for shareable software of the Radiotherapy Treatment Planning Tools Collaborative Working Group. RESULTS: The tool is a rule-based expert system in which the inputs are the CTV contours, critical structure contours, and qualitative information about the specific patient. The output is PTV contours, which are a cylindrical expansion of the CTV. A model for creating PTVs from CTVs is embedded in the tool. The interfractional variation of setup uncertainty and the intrafractional variations of movement of the CTV (e.g., respiration) and patient motion are included in the model. Measured data for the component variations is consistent with modeling the components as independent samples from 3D Gaussian distributions. The components are combined using multivariate normal statistics to yield the cylindrical expansion factors. Rules are used to represent the values of the components for certain patient conditions (e.g., setup uncertainty for a head and neck patient immobilized in a mask). The tool uses a rule interpreter to combine qualitative information about a specific patient with rules representing the value of the components and to enter the appropriate component values for that patient into the cylindrical expansion formula. CONCLUSION: The portable software tool allows the rapid, consistent, and automatic generation of 3D PTVs from CTVs.

Prism: a new approach to radiotherapy planning software.

PURPOSE: We describe the capabilities and performance of Prism, an innovative new radiotherapy planning system with unusual features and design. The design and implementation strategies are intended to assure high quality and clinical acceptability. The features include Artificial Intelligence tools and special support for multileaf collimator (MLC) systems. The design provides unusual flexibility of operation and ease of expansion. METHODS AND MATERIALS: We have implemented Prism, a three-dimensional (3D) radiotherapy treatment-planning system on standard commercial workstations with the widely available X window system. The design and implementation use ideas taken from recent software engineering research, for example, the use of behavioral entity-relationship modeling and the "Mediator Method" instead of ad-hoc programming. The Prism system includes the usual features of a 3D planning system, including Beam's Eye View and the ability to simulate any treatment geometry possible with any standard radiotherapy accelerator. It includes a rule-based expert system for automated generation of the planning target volume as defined in ICRU Report 50. In addition, it provides special support for planning treatments with a multileaf collimator (MLC). We also implemented a Radiotherapy Treatment Planning Tools Foundation for Prism, so that we are able to use software tools form other institutions without any source code modification. RESULTS: The Prism system has been in clinical operation at the University of Washington since July 1994 and has been installed at several other clinics. The system is run simultaneously by several users, each with their own workstation operating from a common networked database and software. In addition to the dosimetrists, the system is used by radiation oncologists to define tumor and target volumes and by radiation therapists to select treatment setups to load into a computer controlled accelerator. CONCLUSIONS: Experience with the installation and operation has shown the design to be effective as both a clinical and research tool. Integration of software tools has eased the development and significantly enhanced the clinical usability of the system. The design has been shown to be a sound basis for further innovation in radiation treatment planning software and for research in the treatment planning process.

Portable software tools for 3D radiation therapy planning.

PURPOSE: Produce a collection of software tools (computer programs) that support three-dimensional (3D) radiation therapy planning. The tools are not a complete 3D planning system. Instead, they work with any 3D planning system that meets certain minimal specifications. The tools assist in deriving anatomic data from images, generating target volume contours, evaluating treatment plans, and verifying accurate treatment delivery. The tools are portable: they can run without source code changes in any computing environment that provides a library of functions and data definitions called the Foundation. The Foundation couples the portable tools to the (usually nonportable) file system and dose calculation associated with a particular 3D planning system. METHODS AND MATERIALS: Tools were written at three different (geographically separated) institutions. Software developers from all three sites specified the Foundation. The programmers' interface to the Foundation is portable, but a Foundation implementation need not be portable. Each group implemented a Foundation adapted to the (different) 3D planning system used at their site. RESULTS: All tools run at all three sites without source code changes. Each Foundation was implemented in a few person-months of programming effort. The program text and documentation for the tools have been placed in the public domain. CONCLUSIONS: It is practical and economical to produce portable radiotherapy treatment planning tools. Providers of 3D planning programs should offer Foundations for their systems, so they can be used with tools. Researchers considering new computer programs should write them as tools, so they can work with any 3D planning system.

Automated planning target volume generation: an evaluation pitting a computer-based tool against human experts.

PURPOSE: Software tools are seeing increased use in three-dimensional treatment planning. However, the development of these tools frequently omits careful evaluation before placing them in clinical use. This study demonstrates the application of a rigorous evaluation methodology using blinded peer review to an automated software tool that produces ICRU-50 planning target volumes (PTVs). METHODS AND MATERIALS: Seven physicians from three different institutions involved in three-dimensional treatment planning participated in the evaluation. Four physicians drew partial PTVs on nine test cases, consisting of four nasopharynx and five lung primaries. Using the same information provided to the human experts, the computer tool generated PTVs for comparison. The remaining three physicians, designated evaluators, individually reviewed the PTVs for acceptability. To exclude bias, the evaluators were blinded to the source (human or computer) of the PTVs they reviewed. Their scorings of the PTVs were statistically examined to determine if the computer tool performed as well as the human experts. RESULTS: The computer tool was as successful as the human experts in generating PTVs. Failures were primarily attributable to insufficient margins around the clinical target volume and to encroachment upon critical structures. In a qualitative analysis, the human and computer experts displayed similar types and distributions of errors. CONCLUSIONS: Rigorous evaluation of computer-based radiotherapy tools requires comparison to current practice and can reveal areas for improvement before the tool enters clinical practice.

The use of medical images in planning and delivery of radiation therapy.

The authors provide a survey of how images are used in radiation therapy to improve the precision of radiation therapy plans, and delivery of radiation treatment. In contrast to diagnostic radiology, where the focus is on interpretation of the images to decide if disease is present, radiation therapy quantifies the extent of the region to be treated, and relates it to the proposed treatment using a quantitative modeling system called a radiation treatment planning (RTP) system. This necessitates several requirements of image display and manipulation in radiation therapy that are not usually important in diagnosis. The images must have uniform spatial fidelity: i.e., the pixel size must be known and consistent throughout individual images, and between spatially related sets. The exact spatial relation of images in a set must be known. Radiation oncologists draw on images to define target volumes; dosimetrists use RTP systems to superimpose quantitative models of radiation beams and radiation dose distributions on the images and on the sets of organ and target contours derived from them. While this mainly uses transverse cross-sectional images, projected images are also important, both those produced by the radiation treatment simulator and the treatment machines, and so-called "digital reconstructed radiographs," computed from spatially related sets of cross-sectional images. These requirements are not typically met by software produced for radiologists but are addressed by RTP systems. This review briefly summarizes ongoing work on software development in this area at the University of Washington Department of Radiation Oncology.

125I brachytherapy k-edge dose enhancement with AgTPPS4.

Photon activation is a radiotherapy technique in which an element is added to the absorbing medium to raise the probability that a photoelectric interaction will occur, thus causing an increase in the absorption of ionizing radiation. Binding energies of key elements within an absorbing medium are closely matched with the incident photon energies to maximize the production of free electrons and subsequent absorption of their kinetic energies. The purpose of this research was to quantify potential dose enhancement using a silver tetraphenyl sulfonato porphyrin (AgTPPS4) in tumors as a photon activator for use with interstitial 125I brachytherapy. A three-dimensional Monte Carlo dosimetry model was developed using the EGS4 coding system. The photon source was modeled using spectral gamma emissions from models 6702 or 6711 brachytherapy seeds for comparison. Absorbed dose within the tumor volume was calculated for AgTPPS4 concentrations ranging between 0 and 20 mmol/kg tumor weight. These theoretical studies demonstrated linear increases in dose absorbed by the tumor with corresponding increases in AgTPPS4 concentration. The required AgTPPS4 concentration (RSC) to achieve at least a ten percent absorbed dose increase is approximately 6.5 mmol/kg tumor weight for model 6702 seeds. In vivo biodistribution and in vitro toxicity studies were conducted to determine if the theoretically derived RSC could be achieved biologically. Cell toxicity studies showed that TPPS4 porphyrin derivatives were cytotoxic at concentrations required to provide significant brachytherapy dose enhancement. Reverse phase HPLC confirmed that toxicity was due to intrinsic properties of the TPPS4 molecule, not the presence of free silver, drug impurities, or metabolites. Further research is necessary to develop a nontoxic molecular carrier for delivering silver to the DNA of tumor cells.

Knowledge-based computer systems for radiotherapy planning.

Radiation therapy is one of the first areas of clinical medicine to utilize computers in support of routine clinical decision making. The role of the computer has evolved from simple dose calculations to elaborate interactive graphic three-dimensional simulations. These simulations can combine external irradiation from megavoltage photons, electrons, and particle beams with interstitial and intracavitary sources. With the flexibility and power of modern radiotherapy equipment and the ability of computer programs that simulate anything the machinery can do, we now face a challenge to utilize this capability to design more effective radiation treatments. How can we manage the increased complexity of sophisticated treatment planning? A promising approach will be to use artificial intelligence techniques to systematize our present knowledge about design of treatment plans, and to provide a framework for developing new treatment strategies. Far from replacing the physician, physicist, or dosimetrist, artificial intelligence-based software tools can assist the treatment planning team in producing more powerful and effective treatment plans. Research in progress using knowledge-based (AI) programming in treatment planning already has indicated the usefulness of such concepts as rule-based reasoning, hierarchical organization of knowledge, and reasoning from prototypes. Problems to be solved include how to handle continuously varying parameters and how to evaluate plans in order to direct improvements.

How to draw irregular radiation beams in 3-D treatment plans.

We describe a general method for computing the outline which an irregular field originating from some arbitrary angle makes on a plane which may be oriented obliquely within the patient. We describe the mathematical theory of the method, which is based on coordinate transformations expressed as matrix multiplications. Then we describe the implementation of the method in the Pascal programming language, emphasizing language-independent optimizations which ensure fast interactive response. Finally, we describe a systematic program testing procedure that is derived from the mathematical theory, which improves our confidence that the method is coded correctly.

Radiation therapy treatment planning using concurrent programming.

Concurrent programming can be applied to the problem of computer graphic simulation of radiation treatment of tumors (radiation treatment planning). Running several tasks or programs simultaneously on behalf of a single user provides a big improvement over the traditional sequential approach, in which editing a treatment plan and computing and displaying dose distributions are separate operations which must be invoked by explicit commands. With our system, the user sees isodose contours being updated automatically and continuously as the plan is edited; this greatly facilitates plan optimization. The complexity of parallel processing has resulted in a 'conventional wisdom' which discourages this technique. The usual approach is to have parallel processes share a common global data structure, which makes interaction hard to control and discourages modularity and data abstraction. We have developed an alternative approach based on message streams which instead enhances modularity and data abstraction while still providing the advantages of parallel processing. The system is very reliable and is used routinely in a practical clinical environment.

Predicting the efficacy of radiotherapy in individual glioblastoma patients in vivo: a mathematical modeling approach.

Glioblastoma multiforme (GBM) is the most malignant form of primary brain tumors known as gliomas. They proliferate and invade extensively and yield short life expectancies despite aggressive treatment. Response to treatment is usually measured in terms of the survival of groups of patients treated similarly, but this statistical approach misses the subgroups that may have responded to or may have been injured by treatment. Such statistics offer scant reassurance to individual patients who have suffered through these treatments. Furthermore, current imaging-based treatment response metrics in individual patients ignore patient-specific differences in tumor growth kinetics, which have been shown to vary widely across patients even within the same histological diagnosis and, unfortunately, these metrics have shown only minimal success in predicting patient outcome. We consider nine newly diagnosed GBM patients receiving diagnostic biopsy followed by standard-of-care external beam radiation therapy (XRT). We present and apply a patient-specific, biologically based mathematical model for glioma growth that quantifies response to XRT in individual patients in vivo. The mathematical model uses net rates of proliferation and migration of malignant tumor cells to characterize the tumor's growth and invasion along with the linear-quadratic model for the response to radiation therapy. Using only routinely available pre-treatment MRIs to inform the patient-specific bio-mathematical model simulations, we find that radiation response in these patients, quantified by both clinical and model-generated measures, could have been predicted prior to treatment with high accuracy. Specifically, we find that the net proliferation rate is correlated with the radiation response parameter (r = 0.89, p = 0.0007), resulting in a predictive relationship that is tested with a leave-one-out cross-validation technique. This relationship predicts the tumor size post-therapy to within inter-observer tumor volume uncertainty. The results of this study suggest that a mathematical model can create a virtual in silico tumor with the same growth kinetics as a particular patient and can not only predict treatment response in individual patients in vivo but also provide a basis for evaluation of response in each patient to any given therapy.

A research-oriented treatment planning program system.

The function of a treatment planning program is to graphically simulate radiation dose distribution from proposed radiation therapy treatments. While many such programs are available which provide this much-needed service, none addresses the question of how to compare calculation and display techniques. This paper describes a program system described for support of research efforts, particularly development and testing of new calculation algorithms. The system emphasizes a modular flexible structure, enabling programs to be developed somewhat as interchangeable parts. Thus multiple variants of a calculation algorithm can be compared without undue software overhead or additional data management. Unusual features of the system include extensive use of command procedures, logical names and a structured language (PASCAL). These features are described along with other implementation details. Obstacles, limitations and future applications are also discussed.

Biomdecial and Health Informatics Research and Education at the University of Washington.

Although an extensive medical informatics research program as well as courses and training experiences in biomedical informatics have existed at the University of Washington (UW) for many years, a formal home did not exist until 1997 when the Division of Biomedical Informatics was created in the Department of Medical Education, School of Medicine. Since that time the expansion of the research, service and teaching programs has been rapid with a key milestone being a university commitment to provide funding, space and faculty to support the development of a new graduate program in Biomedical and Health Informatics. Hallmarks of the biomedical and health informatics program at the University of Washington include: - Strong shared belief that informatics research can contribute to the improvement of healthcare and health; - Large, multidisciplinary faculty including faculty from computer science, library and information science as well as the health sciences schools (dentistry, medicine, nursing, pharmacy, and public health and community medicine); - Comprehensive research and development partnership with the University of Washington Medical Centers information systems group and the UW Primary Care Network to move research from the laboratory to operational clinical systems; - Extensive and diverse regional setting in which to study information needs and develop informatics solutions in primary care settings; - Lack of barriers to interdisciplinary research and teaching.

Radiotherapy planning: direct tumor location on simulation and port films using CT. Part I. Principles.

Although there have been great advances in cancer diagnosis in recent years, it remains difficult to transfer tumor location information from cross-sectional computed tomographic (CT) scans or magnetic resonance images to the simulation and verification films used in planning radiotherapy. A newly developed system uses radioopaque markers attached to the patient as reference points. These markers are identified on both CT scans and simulation films and their locations entered into the treatment planning computer. The tumor and any desired normal structures are then outlined manually on each CT section. Transparent overlays produced by the computer show the position of the reference markers and tumor outlines for any combination of gantry angles and source-film distance. Because the overlays are scaled to the simulation films, the reference points enable precise alignment of overlay and film. The tumor outline thus appears on the simulation or verification films exactly as it is "seen" by the therapy beam, making field verification straightforward and accurate, even on oblique films.