Prostate cancer treatment with radiotherapy: maturing methods that minimize morbidity.

Technological advances in treatment delivery and planning have provided the backdrop for an unprecedented number of options in the treatment of prostate cancer with radiotherapy. The more common choices include classical external-beam radiotherapy, external-beam radiotherapy using three-dimensional treatment planning and conformal radiotherapy (3DCRT), ultrasound-guided transperineal implant monotherapy alone or in combination with external-beam radiotherapy, and intensity-modulated radiotherapy (IMRT) techniques. This chapter reviews the data from these methods with an emphasis on dose escalation, provides comparisons with prostate-specific antigen (PSA)-era radical prostatectomy series where appropriate, and highlights future initiatives designed to further improve outcome.

Planning target volumes for radiotherapy: how much margin is needed?

PURPOSE: The radiotherapy planning target volume (PTV) encloses the clinical target volume (CTV) with anisotropic margins to account for possible uncertainties in beam alignment, patient positioning, organ motion, and organ deformation. Ideally, the CTV-PTV margin should be determined solely by the magnitudes of the uncertainties involved. In practice, the clinician usually also considers doses to abutting healthy tissues when deciding on the size of the CTV-PTV margin. This study calculates the ideal size of the CTV-PTV margin when only physical position uncertainties are considered. METHODS AND MATERIALS: The position of the CTV for any treatment is assumed to be described by independent Gaussian distributions in each of the three Cartesian directions. Three strategies for choosing a CTV-PTV margin are analyzed. The CTV-PTV margin can be based on: 1. the probability that the CTV is completely enclosed by the PTV; 2. the probability that the projection of the CTV in the beam's eye view (BEV) is completely enclosed by the projection of the PTV in the BEV; and 3. the probability that a point on the edge of the CTV is within the PTV. Cumulative probability distributions are derived for each of the above strategies. RESULTS: Expansion of the CTV by 1 standard deviation (SD) in each direction results in the CTV being entirely enclosed within the PTV 24% of the time; the BEV projection of the CTV is enclosed within the BEV projection of the PTV 39% of the time; and a point on the edge of the CTV is within the PTV 84% of the time. To have the CTV enclosed entirely within the PTV 95% of the time requires a margin of 2.8 SD. For the BEV projection of the CTV to be within the BEV projection of the PTV 95% of the time requires a margin of 2.45 SD. To have any point on the surface of the CTV be within the PTV 95% of the time requires a margin of 1.65 SD. CONCLUSION: In the first two strategies for selecting a margin, the probability of finding the CTV within the PTV is unrelated to dose variations in the CTV. In the third strategy, the specified confidence limit is correlated with the minimum target dose. We recommend that the PTV be calculated from the CTV using a margin of 1.65 SD in each direction. This gives a minimum CTV dose that is greater than 95% of the minimum PTV dose. Additional sparing of adjoining healthy structures should be accomplished by modifying beam portals, rather than adjusting the PTV. Then, the dose distributions more accurately reflect the clinical compromise between treating the tumor and sparing the patient.

Uncertainty in dose estimation for gynecological implants.

One source of uncertainty in doses computed for intracavitary gynecological applications is the imprecision inherent in localizing the sources and the points of interest on radiographs of the implant and in transferring that data into the treatment planning computer. To quantify the effect of these activities on the accuracy of computed doses, five physicists and two dosimetrists performed computerized dose calculations on five applications chosen randomly from our patient files. For each of these applications, doses were computed at the traditional points A and B and at points in the bladder and rectum. Using identical sets of films, each planner located both the radioactive sources and points of interest, or only the sources, or only the points of interest. Another set of films was used to measure the accuracy of digitizing alone. Planners received no instructions on either the definition or the placement of the points of interest. Overall uncertainties in computed doses to points A and B and bladder were found to be about 7%. Uncertainty in dose to the rectum was on the order of 50%. Analysis of the results showed that about 1% of the error was due to digitization and about 2% to identification of source locations. Among the individual planners, almost all of the dose variation was from differences in placement of the points of interest on the implant radiographs. The results demonstrate the need for standard definitions and locations for points of calculation so that meaningful comparisons can be made among institutions.

Evaluation and scoring of radiotherapy treatment plans using an artificial neural network.

PURPOSE: The objective of this work was to demonstrate the feasibility of using an artificial neural network to predict the clinical evaluation of radiotherapy treatment plans. METHODS AND MATERIALS: Approximately 150 treatment plans were developed for 16 patients who received external-beam radiotherapy for soft-tissue sarcomas of the lower extremity. Plans were assigned a figure of merit by a radiation oncologist using a five-point rating scale. Plan scoring was performed by a single physician to ensure consistency in rating. Dose-volume information extracted from a training set of 511 treatment plans on 14 patients was correlated to the physician-generated figure of merit using an artificial neural network. The neural network was tested with a test set of 19 treatment plans on two patients whose plans were not used in the training of the neural net. RESULTS: Physician scoring of treatment plans was consistent to within one point on the rating scale 88% of the time. The neural net reproduced the physician scores in the training set to within one point approximately 90% of the time. It reproduced the physician scores in the test set to within one point approximately 83% of the time. CONCLUSIONS: An artificial neural network can be trained to generate a score for a treatment plan that can be correlated to a clinically-based figure of merit. The accuracy of the neural net in scoring plans compares well with the reproducibility of the clinical scoring. The system of radiotherapy treatment plan evaluation using an artificial neural network demonstrates promise as a method for generating a clinically relevant figure of merit.

Very fast simulated reannealing in radiation therapy treatment plan optimization.

PURPOSE: Very Fast Simulated Reannealing is a relatively new (1989) and sophisticated algorithm for simulated annealing applications. It offers the advantages of annealing methods while requiring shorter execution times. The purpose of this investigation was to adapt Very Fast Simulated Reannealing to conformal treatment planning optimization. METHODS AND MATERIALS: We used Very Fast Simulated Reannealing to optimize treatments for three clinical cases with two different cost functions. The first cost function was linear (minimum target dose) with nonlinear dose-volume normal tissue constraints. The second cost function (probability of uncomplicated local control) was a weighted product of normal tissue complication probabilities and the tumor control probability. RESULTS: For the cost functions used in this study, the Very Fast Simulated Reannealing algorithm achieved results within 5-10% of the final solution (100,000 iterations) after 1000 iterations and within 3-5% of the final solution after 5000-10000 iterations. These solutions were superior to those produced by a conventional treatment plan based on an analysis of the resulting dose-volume histograms. However, this technique is a stochastic method and results vary in a statistical manner. Successive solutions may differ by up to 10%. CONCLUSION: Very Fast Simulated Reannealing, with modifications, is suitable for radiation therapy treatment planning optimization. It produced results within 3-10% of the optimal solution, produced using another optimization algorithm (Mixed Integer Programming), in clinically useful execution times.

Tissue heterogeneity effects in treatment plan optimization.

PURPOSE: There is general agreement that tissue density correction factors improve the accuracy of dose calculations. However, there is disagreement over the proper heterogeneity correction algorithm and a lack of clinical experience in using them. Therefore, there has not been widespread implementation of density correction factors into clinical practice. Furthermore, the introduction of optimized conformal therapy leads to new and radically different treatment techniques outside the clinical experience of the physician. It is essential that the effects of tissue density corrections are understood so that these types of treatments can be safely delivered. METHODS AND MATERIALS: In this paper, we investigate the effect of tissue density corrections on optimized conformal type treatment planning in the thorax region. Specifically, we study the effects on treatment plans optimized without type treatment planning in the thorax region. Specifically, we study the effects on treatment plans optimized without tissue density corrections, when those corrections are applied to the resulting dose distributions. These effects are compared for two different conformal techniques. RESULTS: This study indicates that failure to include tissue density correction factors results in an increased dose of approximately 5-15%. This is consistent with published studies using conventional treatment techniques. Additionally, the high-dose region of the dose distribution expands laterally into the uninvolved lung and other normal structures. The use of dose-volume histograms to compare these distributions demonstrates that treatment plans optimized without tissue density corrections lead to an increased dose to uninvolved normal structures. This increase in dose often violates the constraints used to determine the optimal solution. CONCLUSIONS: The neglect of tissue density correction factors can result in a 5-15% increase in the delivered dose. In addition, suboptimal dose distributions are produced. To benefit from the advantages of optimized conformal therapy in the thorax, tissue density correction factors should be used.

Prostate target volume variations during a course of radiotherapy.

PURPOSE: The purpose of this study was to measure the mobility of the clinical target volume (CTV) in prostate radiotherapy with respect to the pelvic anatomy during a course of therapy. These data are needed to properly design the planning target volume (PTV). METHODS AND MATERIALS: Seventeen patients were studied. Each patient underwent computed tomography (CT) scanning for treatment planning purposes. Subsequently, three CT scans were obtained at approximately 2-week intervals during treatment. The prostate, seminal vesicles, bladder, and rectum were outlined on each CT study. The second through the fourth CT studies were aligned with the first study using a rigid body transformation based on the bony anatomy. The transformation was used to compute the center of mass position and bounding box of each organ in the subsequent studies relative to the first study. Differences in the bounding box limits and center of mass positions between the first and subsequent studies were tabulated and correlated with bladder and rectal volume and positional parameters. RESULTS: The mobility of the CTV was characterized by standard deviations of 0.09 cm (left-right), 0.36 cm (cranial-caudal), and 0.41cm (anterior-posterior). Prostate mobility was not significantly correlated with bladder volume. However, the mobility of both the prostate and seminal vesicles was very significantly correlated with rectal volume. Bladder and rectal volumes decreased between the pretreatment CT scan and the first on-treatment CT scan, but were constant for all on-treatment CT scans. CONCLUSION: Margins between the CTV and PTV based on the simple geometric requirement that a point on the edge of the CTV is enclosed by the PTV 95% of the time are 0.7 cm in the lateral and cranial-caudal directions, and 1.1 cm in the anterior-posterior direction. However, minimum dose to the CTV and avoidance of organs at risk are more important considerations when drawing beam apertures. More consistent methods for reproducing prostate position (e.g., empty rectum) and more sophisticated beam aperture optimization are needed to guarantee consistent coverage of the CTV while avoiding organs at risk.

Comparison of simulated annealing algorithms for conformal therapy treatment planning.

PURPOSE: The efficiency of four fast simulated annealing algorithms for optimizing conformal radiation therapy treatment plans was studied and the resulting plans were compared with each other and to optimized conventional plans. METHODS AND MATERIALS: Four algorithms were selected on the basis of their reported successes in solving other minimization problems: fast simulated annealing with a Cauchy generating function, fast simulated annealing with a Lorentzian generating function, variable step size generalized simulated annealing (VSGSA), and very fast simulated reannealing (VFSR). They were tested on six clinical cases using a multiple beam coplanar conformal treatment technique. Relative beam weights were computed that maximized the minimum tumor dose subject to dose-volume constraints on normal organ doses. Following some initial tuning of the annealing parameters, each algorithm was applied identically to each test case. Optimization tests were run using different random number sequences and different numbers of iterations. RESULTS: The VSGSA algorithm consistently produced the best results. Using long run times, it generated plans with the highest minimum tumor dose in five of the six cases. For the short run times, the VSGSA solutions averaged larger minimum tumor doses than those of the other algorithms for all six patients, with increases ranging from 0.4 to 5.9 Gy. For three of the patients, the conformal plan gave a clinically significant increase in the minimum tumor dose over the conventional plan, ranging from 8.2 to 13.0 Gy. In two other cases, there was little difference between the two treatment approaches. For one case, the optimized conventional plan was much better than the conformal plan because the conventional beam arrangement included wedges, which offset the multiple beam advantage of the conformal plans. CONCLUSIONS: For equal computing times of both long and short duration, the VSGSA algorithm consistently produced conformal plans that were superior to those produced by the other algorithms. The simple conformal technique used in this study showed a significant potential advantage in the treatment of abdominal tumors. In three of the cases, the conformal plans showed clinically important increases in tumor dose over optimized conventional plans.

The influence of dose constraint point placement on optimized radiation therapy treatment planning.

To efficiently use linear and quadratic programming for treatment planning optimization on a routine basis, automated methods are needed for placing dose constraint points. We have investigated, for linear programming optimization, the minimum number of constraint points needed to achieve an acceptable approximation to the desired (ideal) solution. Seven different constraint point placement algorithms were evaluated for a given objective function. One of these algorithms was chosen for routine clinical use at our institution. This algorithm places constraint points on the perimeter of the target volume and on the perimeter and in the interior of each normal structure. Additional points are placed on the perimeter of a constant thickness buffer region surrounding the target volume. Excellent optimization results are obtained with 40-70 constraint points per treatment planning slice.

Improved dose homogeneity in the head and neck using computer controlled radiation therapy.

Computer-controlled radiation therapy techniques are demonstrated which improve dose homogeneity throughout the nasopharynx when compared to conventional treatment techniques. The typical approach using a heavily weighted anterior field and opposed wedged lateral fields results in a dose gradient from 95% to 110% or greater. All three of the computer-controlled techniques investigated improved the dose uniformity to a range from 95% to 105% or less. Multiple overlapping fields are used to compensate for patient anatomy and treatment beam characteristics. Treatment planning and monitor unit calculations are quite time-consuming at this stage of development. Actual treatment time is not unreasonably long and can be improved in future releases of the therapy machine control software.

Custom beam profiles in computer-controlled radiation therapy.

A computer-controlled radiation therapy technique is demonstrated which uses multiple concurrent boost fields to modify the beam profile of a conventional treatment beam. A principal field, identical to that of a corresponding conventional treatment plan, delivers the major component of the prescribed dose. Dose increments given from boost fields placed within this principal field compensate for variations in patient anatomy, for variations in target volume shape, and/or for imperfect beam characteristics, such as excessive off-axis dose or inadequate beam wedge angle. This concurrent boost field technique is demonstrated for several treatment sites. It produces significant improvement in uniformity of dose delivered to the target compared to conventional treatment. Implementation of these treatments requires a computer-controlled linear accelerator with independently-movable collimator jaws, an automatic beam set-up procedure, and a patient prescription database. Since all fields are delivered under computer control, concurrent boost technique treatment times are not much longer than those of conventional treatments.

Conventional vs. conformal radiotherapy for prostate cancer: preliminary results of dosimetry and acute toxicity.

PURPOSE: To compare conformal radiotherapy using three dimensional treatment planning (3D-CRT) to conventional radiotherapy (Conven-RT) for patients with Stages T2-T4 adenocarcinoma of the prostate. METHODS AND MATERIALS: A Phase III randomized study was activated in May 1993, to compare treatment toxicity and patient outcome after 78 Gy in 39 fractions using 3D-CRT to that after 70 Gy in 35 fractions using Conven-RT. The first 46 Gy were administered using the same nonconformal field arrangement (four field) in both arms. The boost was given nonconformally using four fields in the Conven-RT arm and conformally using six fields in the 3D-CRT arm. The dose was specific to the isocenter. The first 60 patients, 29 in the 3D-CRT arm and 31 in the Conven-RT arm, are the subject of this preliminary analysis. RESULTS: The two treatment arms were first compared in terms of dosimetry by dose-volume histogram analysis. Using a subgroup of patients in the 3D-CRT arm (n=15), both Conven-RT and 3D-CRT plans were generated and the dose-volume histogram data compared. The mean volumes treated to doses above 60 Gy for the bladder and rectum were 28 and 36% for the 3D-CRT plans, and 43 and 38% for the Conven-RT plans, respectively (p < 0.05 for the bladder volumes). The mean clinical target volume (prostate and seminal vesicles) treated to 95% of the prescribed dose was 97.5% for the 3D-CRT arm, and 95.6% for the Conven-RT arm (p < 0.05). There were no significant differences in the acute reactions between the two arms, with the majority experiencing Grade 2 or less toxicity (92%). Moreover, no relationship was seen between acute toxicity and the volume of bladder and rectum receiving in excess of 60 Gy for those in the 3D-CRT arm. There was also no difference between the groups in terms of early biochemical response. Prostate-specific antigen levels at 3 and 6 months after completion of radiotherapy were similar in the two treatment arms. There was only one biochemical failure in the study population at the time of the analysis. CONCLUSIONS: Comparison of the Conven-RT and 3D-RT treatment plans revealed that significantly less bladder was in the high dose volume in the 3D-CRT plans, while the volume of rectum receiving doses over 60 Gy was equivalent. There were no differences between the two treatment arms in terms of acute toxicity or early biochemical response. Longer follow-up is needed to determine the impact of 3D-CRT on long-term patient outcome and late reactions.

Writing software for the clinic.

Medical physicists often write computer programs to support scientific, educational, and clinical endeavors. Errors in scientific and educational software can waste time and effort by producing meaningless results, but errors in clinical software can contribute to patient injuries. Although the ultimate goal of error-free software is impossible to achieve except in very small programs, there are many good design, implementation, and testing practices that can be used by small development groups to significantly reduce errors, improve quality, and reduce maintenance. The software development process should include four basic steps: specifications, design, implementation, and testing. A specifications document defining what the software is intended to do is valuable for clearly delimiting the scope of the project and providing a benchmark for evaluating the final product. Keep the software design simple and straightforward. Document assumptions, and check them. Emphasize maintainability, portability, and reliability rather than speed. Use layers to isolate the application from hardware and the operating system. Plan for upgrades. Expect the software to be used in unplanned ways. Whenever possible, be generous with RAM and disk storage; hardware is cheaper than development and maintenance. During implementation, use well-known algorithms whenever possible. Use prototypes to try out ideas. Use generic modules, version numbering, unique file names, defensive programming, and operating system and language/compiler defaults. Avoid binary data files and clever tricks. Remember that real numbers are not exact in a computer. Get it right before making it faster. Document the software extensively. Test continuously during development; the later a problem is found, the more it costs to fix. Use a written procedure to test the final product exactly as a typical user would run it. Allow no changes after clinical release. Expect to spend at least an additional 50% of the initial development effort on testing, fixing errors, and getting the software into routine operation.

Isodose plotting for pen plotters.

A general algorithm for treatment plan isodose line plotting is described which is particularly useful for pen plotters. Unlike other methods of plotting isodose lines, this algorithm is designed specifically to reduce pen motion, thereby reducing plotting time and wear on the transport mechanism. Points with the desired dose value are extracted from the dose matrix and stored, sorted into continuous contours, and then plotted. This algorithm has been implemented on DEC PDP-11/60 and VAX-11/780 computers for use with two models of Houston Instrument pen plotters, two models of Tektronix vector graphics terminals, a DEC VT125 raster graphics terminal, and a DEC VS11 color raster graphics terminal. Its execution time is similar to simpler direct-plotting methods.

A three-film technique for reconstruction of radioactive seed implants.

Computer dose calculations for interstitial implants of radioactive seeds require a knowledge of the spatial coordinates of each seed in the implant. These coordinates are usually constructed from the seed images on either an orthogonal or a stereo pair of radiographs. Such a procedure, however, requires that each seed can be identified on each film. We have developed a computer algorithm which computes the locations of the seeds from a set of two stereo and one AP radiographs, even when seed identities on all films are indeterminable. This is done by means of a ray tracing technique. Under ideal conditions the projections from the three films uniquely define all seed locations. Under clinical conditions, however, where digitizing uncertainties and patient movement are inevitable, the algorithm must determine the most likely set of seed locations from the larger set of all possible seed locations. The criteria for making this selection, as well as clinical examples, are presented.

Assessment of a linear accelerator for segmented conformal radiation therapy.

Segmented conformal radiation therapy is a new computer-controlled treatment technique under investigation in which the target volume is subdivided into thick transverse segments each of which is then treated individually by rectangular transverse abutting fields. In order to obtain uniform dose at abutments, the machine isocenter remains fixed in the patient and field edges are defined by independently moving focused collimator jaws to give matching geometric divergence. Mechanical variation in jaw and gantry positioning will create some dose variation at field abutments. Film dosimetry was used to study the radiation field positioning accuracy and precision of a commercial linear accelerator. A method of field position calibration was developed using multiple nonabutting fields exposed on the same radiograph. Verification of collimator jaw calibration measurements was performed using multiple abutting fields exposed on a single radiograph. Measurements taken over 5 months of clinical accelerator operation studied the effects of simple jaw motion, simple gantry motion, and combined jaw/gantry motion on jaw position precision and accuracy. The inherent precision and accuracy of radiation field positioning was found to be better than +/- 0.3 mm for both jaws with all types of motions except for the Y2 jaw under combined jaw/gantry motion. When the ability to deliver abutting beams was verified in clinical mode, the average dose variation at abutments was less than 6% at all gantry angles except for one. However, due to accelerator software limitations in clinical mode, the settings for collimator positions could not take advantage of the maximum accuracy of which the hardware is capable.(ABSTRACT TRUNCATED AT 250 WORDS)

Accuracy of a two-sensor sonic digitizer.

Digitizing devices are typically used in radiotherapy computer treatment planning for entering patient anatomy, the locations of internal radioactive source, and the outlines of irregularly shaped external beams. The errors encountered in the use of a large-area two-sensor sonic digitizer for computer input have been studied. Conversion of data from triangular to Cartesian coordinates makes the precision of the digitizer nonuniform over the sensitive area. The response of each senor has been measured and found to be a nonlinear function of distance. The assumption of linearity in computing the triangular distances from the sensor readings produces errors in the computed distances of up to 0.8%. An alternative method of computing the distances using a fitted cubic function reduces the errors to less than 0.1%. For a test pattern, the maximum position error was reduced from 0.5 to 0.1 cm.

Effect of using an initial polyenergetic spectrum with the pencil-beam redefinition algorithm for electron-dose calculations in water.

This work compares the accuracy of dose distributions computed using an incident polyenergetic (PE) spectrum and a monoenergetic (ME) spectrum in the electron pencil-beam redefinition algorithm (PBRA). It also compares the times required to compute PE and ME dose distributions. This has been accomplished by comparing PBRA calculated dose distributions with measured dose distributions in water from the National Cancer Institute electron collaborative working group (ECWG) data set. Comparisons are made at 9 and 20 MeV for the 15 x 15 cm2 and 6 x 6 cm2 fields at 100- and 110-cm SSD. The incident PE spectrum is determined by a process that best matches the weighted sum of monoenergetic PBRA calculated central-axis depth doses, each calculated with the energy correction factor, C(E), equal to unity, to the ECWG measured depth dose for the 15 x 15 cm2 field at 100-cm SSD. C(E) is determined by a least square fit to central-axis depth dose for the PE PBRA. Results show that both the PE and ME PBRA accurately calculate central-axis depth dose at 100-cm SSD for the 6 x 6 cm2 and 15 x 15 cm2 field sizes and also at 110-cm SSD for the 15 x 15 cm2 field size. In the penumbral region, the PE PBRA calculation is significantly more accurate than the ME PBRA for all measurement conditions. Both the PE and ME PBRA exhibit significant dose errors (> 4%) outside the penumbra at shallow depths for the 6 x 6 cm2 and 15 x 15 cm2 fields at 100-cm SSD and inside the penumbra at shallow depths for the 6 x 6 cm2 field size at 110-cm SSD. These errors are attributed to the fact that the PBRA does not model collimator scatter in the incident beam. Calculation times for the PE PBRA are approximately 70%-140% greater than those for the ME PBRA. We conclude that the PE PBRA is significantly more accurate than the ME PBRA, and we believe that the increase in time for the PE PBRA will not significantly impact the clinical utility of the PBRA.

An equivalent squares template.

A template for calculating equivalent squares of irregularly shaped fields is described. Calculations using the template are essentially scatter summations. Comparisons with a computer program showed good agreement in equivalent squares and tumor doses (within 0.8%). Comparisons with area-perimeter ratio methods of computing equivalent squares showed consistently better accuracy using the template. For highly irregular fields, the template calculation required 3-10 min for central axis computation, comparable to that required by the other manual methods.

Satellite digital display for the Clinac-18.

A satellite digital display of the gantry and collimator positions has been mounted on the Clinac-18 console. This module provides simultaneous digital readout of the gantry angle, collimator rotation angle, upper-jaw position, and lower-jaw position. Continuous display of these parameters during the treatment is important in minimizing patient treatment errors.