Remote interstitial afterloading in cancer of the prostate: preliminary experience with the MicroSelectron.

Brachytherapy in the treatment of prostate cancer is an accepted modality. A considerable experience has been accrued using Iridium interstitial therapy. The major problem associated with this choice is that of radioprotection. We describe the first clinical use of the MicroSelectron, a remote afterloading device, in prostate brachytherapy. Twenty-six patients have now been treated using this system. Approximately 55 treatment interruptions with an average 5 hr increase in overall treatment time occurs. The initial problems associated with instituting remote afterloading to prostate brachytherapy and their solutions are discussed.

Irradiation of volunteers in nuclear medicine.

The preliminary assessment of many radiopharmaceuticals is often carried out with the help of "normal volunteers". These volunteers are drawn from the general public, are fully informed of the procedure to be performed and its attendant risks, and in many cases are compensated financially for their trouble. The cooperation of such people is of vital importance to the full understanding of the normal kinetics and metabolism of many new radiopharmaceuticals. The restrictions on the choice of normal volunteers, and the radiation dose limits which must be observed are not explicitly defined in any of the current guidelines, and in this paper we propose a rationale, based upon available information, which sets acceptable limits for volunteers, and provides a framework within which scientists and physicians can work.

Correction for out-of-focal-plane blurring in a simulated multiplane tomographic scanner.

A technique has been developed to remove out-of-focal-plane blurring from coronal and axial images made using a multiplane tomographic scanner. The technique uses a combined smoothing and differential operator that is applied to the axial images. It has been tested using computer-simulated images, with favorable results. The usefulness of the technique in a real system has yet to be determined.

Optimized tissue compensators.

The fabrication of retracted missing tissue compensators which account only for missing primary attenuation produces compensation which is less than optimum. Experimental results can be accurately reproduced with theoretical calculations based on primary and first-order scattered radiation. This method of analysis may be extended to predict the shape of a compensator which will produce the desired dose distribution at a given depth in phantom. An optimized compensator is constructed based on these theoretical considerations and the analysis of its performance is presented.

Performance characteristics of a widely used orthovoltage x-ray therapy machine.

Experimental results sufficient to provide a data base for the Philips RT 250 x-ray machine are presented. A representative selection of operating conditions covering the working range of the machine from 75 to 250 kVp, were used. Beam quality, depth dose data, field flatness, and filtration characteristics were studied, and selected results are reported.

Modeling of asymmetric compensator geometries.

Dose distributions arising from the use of near unit density retracted missing tissue compensators in symmetric geometries have been successfully modeled on the basis of primary and first-order scattered radiation. This method of analysis has been extended to both low- and high-density materials in asymmetric geometries. Good agreement is achieved between theory and experiment.

Missing tissue compensators: evaluation and optimization of a commercial system.

A commercial system for producing retracted compensators has been adapted to suit local needs, and is evaluated here. It comprises a magnetic field surface digitizer and computer-driven milling machine. Improvements in dose distributions, resulting in standard deviations of the mean dose between 2% and 3%, have been achieved for treatment fields in wax phantoms simulating the head and neck regions. Optimization of compensator shape to allow for changes in the amount of scattered radiation has resulted in a further improvement in dose uniformity, particularly near the field borders; for these compensators the standard deviation was as low as 1.6%. The system using the basic algorithm has been in clinical use since July.

Optimization of pencil beam widths for electron-beam dose calculations.

The pencil beam method of calculating dose distributions for electron-beam radiotherapy has been very useful, however, several limitations in the approach have been recognized. One such limitation is the lack of a mechanism to model range straggling of electrons. For stationary electron-beam calculations, range straggling is incorporated incompletely in the planar-fluence-to-dose conversion factor, which uses measured percentage depth dose curves to force the calculated percentage depth dose to reproduce the measurement. When calculating the dose distribution for an arced beam using a pencil beam algorithm, insufficient modeling of the pencil beams leads to larger errors than when using a stationary beam algorithm. The calculated depth of maximum dose is systematically over-estimated by the pencil beam calculations. We will show that the lack of a way to account for range straggling in the arc-electron pencil beam calculation is primarily responsible for this discrepancy. Methods of incorporating range straggling into the electron pencil beam dose calculation have been presented before, but no data have been shown to support their use for heterogeneous phantoms (patients). This paper presents a similar range-straggling modification, as well as data to show that this model can predict pencil beam width to within 20% for heterogeneous slab phantoms. For stationary electron-beam calculations, the calculated isodose lines follows the measured isodose lines to within 1 mm down to the 10% dose level.(ABSTRACT TRUNCATED AT 250 WORDS)

Acquisition and display of radiation dose distributions using microcomputer technology.

In the commissioning or quality assurance of a medical linear accelerator or a computerized radiotherapy planning system, the traditional approach usually consists of acquiring and comparing one-dimensional dose profiles. This methodology is tedious and incomplete since only a portion of the radiation field can realistically be sampled. We have developed an automated measurement system which allows efficient measurement and display of complete two-dimensional dose distributions. The general purpose microcomputer used (IBM PC/XT compatible) can be interfaced economically to any water phantom dosimetry system equipped with a three axis scan controller, and can also communicate data to the treatment planning system. This allows for direct comparison of measured with computed dose distributions, thus revealing discrepancies in the dose computation algorithms used. In this paper, we describe the interface between the microcomputer, a conventional water dosimetry system (Therados RFA-3), and a treatment planning computer. We report our early experience with acquiring dose distributions and show sample comparisons with computed results for megavoltage electron beams incident on homogeneous and heterogeneous systems.

A convolution method of calculating dose for 15-MV x rays.

Arrays were generated using the Monte Carlo method representing the energy absorbed throughout waterlike phantoms from charged particles and scatter radiation set in motion by primary interactions at one location. The resulting "dose spread arrays" were normalized to the collision fraction of the kinetic energy released by the primary photons. These arrays are convolved with the relative primary fluence interacting in a phantom to obtain three-dimensional dose distributions. The method gives good agreement for the 15-MV x-ray dose in electronic disequilibrium situations, such as the buildup region, near beam boundaries, and near low-density heterogeneities.

Verification of a two-dimensional pencil beam arc electron dose calculation algorithm.

The dosimetry of arced electron beams is of increasing importance because of the increased capabilities of modern linear accelerators. A practical pencil beam algorithm has been developed for arc electron beams and is capable of using computed tomography information for heterogeneity corrections. For homogeneous phantoms, the maximum dose and bremsstrahlung components are predicted very accurately, that is, within 1% of the maximum dose. However, the depth of maximum dose (treatment depth) is predicted to be deeper than measurement, as much as 0.7 cm deeper. For a heterogeneous lung phantom, the discrepancies are as high as 30%, but the accuracy of dose calculation is consistent with conventional stationary pencil beam algorithms. It was concluded that improvements in the dose prediction are possible with more accurate calculations of the pencil beam widths and the incorporation of range straggling into the algorithm.

Shielding considerations for tomotherapy.

Tomotherapy presents an evolutionary modality that holds forth the promise of better dose conformation to tumor volumes with a concomitant reduction in radiation-induced damage to surrounding normal structures. This delivery technique also presents a new set of radiation protection challenges that impact upon the design of the shielding vault required to house such a unit. A formalism is presented to determine the requisite amounts of shielding for both the primary beam and leakage radiation associated with a generic tomotherapy unit. A comparison is made with the shielding requirements for a conventional linear accelerator operated in a standard manner. Substantial differences in the amount of both primary and secondary shielding are indicated. A tomotherapy primary beam shield is both reduced in width by a factor of almost 10 and increased in thickness by more than a tenth value layer in comparison to a conventional accelerator. Furthermore, the secondary shielding requirements are enhanced by more than two tenth value layers with respect to conventional shielding demands.

An assessment of the physical properties of a multiplane tomographic imager.

The Pho/Con is a dual Anger camera multiplane tomographic scanner, capable of producing longitudinal tomograms through the subject. The physical characteristics of this device have been investigated. The resolution capabilities in three dimensions have been determined, and sensitivity, linearity, and uniformity of each detector have been studied with a view to establishing the influence of each of the subcomponents of the detector on overall performance. Experimental measurements have been made using both analogue and computerised reconstruction techniques.

Iterative deblurring algorithm for a multiplane tomographic scanner.

An iterative algorithm has been developed to reduce blur artefacts in images obtained from the Pho/Con multiplane tomographic scanner. In this technique estimates of the activity distribution are convolved with the detector response function and compared with the original blurred image planes. Successive estimates of the activity distribution are modified so as to minimise the square of the difference in pixel count between the two sets of blurred images. The technique has been tested with images containing non-correlated noise provided by computer simulation and images with correlated noise acquired in experimental phantom studies.

Electron dose distributions in experimental phantoms: a comparison with 2D pencil beam calculations.

Dose distributions were measured and computed within inhomogeneous phantoms irradiated with beams of electrons having initial energies of 10 and 18 MeV. The measurements were made with a small p-type silicon diode and the calculations were performed using the pencil beam algorithm developed originally at the M D Anderson Hospital (MDAH). This algorithm, which is available commercially on many radiotherapy planning computers, is based on the Fermi-Eyges theory of electron transport. The phantoms used in this work were composed of water into which two- and three-dimensional inhomogeneities of aluminum and air (embedded in wax) were introduced. This was done in order to simulate the small bones and the air cavities encountered clinically in radiation therapy of the chest wall or neck. Our intent was to test the adequacy of the two-dimensional implementation of the pencil beam approach. The agreement between measured and computed doses is very good for inhomogeneities which are essentially two-dimensional but discrepancies as large as 40% were observed for more complex three-dimensional inhomogeneities. We can only trace the discrepancies to the complex interplay of numerous approximations in the Fermi-Eyges theory of multiple scattering and its adaptation for practical computer-aided radiotherapy planning.

Monoenergetic approximately of a polyenergetic beam: a theoretical approach.

There exist numerous occasions in which it is desirable to approximate the polyenergetic beams employed in radiation therapy by a beam of photons of a single energy. In some instances, commonly used rules of thumb for the selection of an appropriate energy may be valid. A more accurate approximate energy, however, may be determined by an analysis which takes into account both the spectral qualities of the beam and the material through which it passes. The theoretical basis of this method of analysis is presented in this paper. Experimental agreement with theory for a range of materials and beam qualities is also presented and demonstrates the validity of the theoretical approach taken.

An analytic approach to optimized retracted missing tissue compensators.

The introduction of an attenuating medium into a photon beam serves both to reduce the intensity of the primary beam and to create secondary radiation due to scatter. When retracted missing tissue compensators are employed to compensate for irregular surface contour or internal inhomogeneities, they are often fabricated without regard to the scattered radiation that they introduce into the system. The study of a mathematically describable conical geometry has clearly demonstrated the need for improved compensator design. Experimental results obtained with this geometry can be reproduced with good agreement, using theoretical calculations based on primary and first order scattered radiation. This method of analysis may be extended to predict the shape of a compensator which will produce an optimized dose distribution at a given depth in a phantom, equivalent to that which would be obtained when an irregular surface is filled with unit density material (bolus), producing a flat surface. An optimized compensator was constructed based on these theoretical considerations and excellent agreement was observed between theory and experiment. Dramatic improvement in the restoration of bolus dose is obtained with this optimized compensator. Finally, an anthropomorphic phantom of the neck region has been constructed and the performance of a compensator designed according to current clinical methods for this geometry has been evaluated. The performance of an optimized compensator specific to this geometry is presented and good agreement between theoretical predictions and experimental results is observed. Dramatic improvement in bolus dose restoration over that obtained with the clinically designed compensator is realized.

Limitations of retracted missing tissue compensators: an experimental analysis.

The introduction of an attenuating medium into a photon beam serves both to reduce the intensity of the primary beam and to create secondary radiation due to scatter. When retracted missing tissue compensators are employed to compensate for irregular surface geometry or internal inhomogeneities they are almost always fabricated without regard to the scattered radiation that they introduce to the system. Analysis of such a retracted compensator designed for use with an anthropomorphic phantom reveals an inability to provide true compensation. Subsequent analysis of a mathematically describable conical geometry demonstrates the need for improved compensator design.

Experimental validation of dose distributions predicted by a treatment planning system for complex models.

The trend towards dose escalation in radiation therapy creates the need for non-coplanar beam arrangements to help minimize healthy tissue morbidity in the proximity of treatment volumes. Treatment planning software packages are able to generate external beam non-coplanar isodose distributions and this must be validated before use in a clinical setting. Kodak XV film was used to measure the relative dose within various tissue equivalent phantoms for comparison with the distributions predicted by the G.E. Target ("Target") treatment planning software. The results confirm that Target is able to predict the isodose distribution for coplanar and non-coplanar beams incident on patient equivalent phantoms containing relatively large, semi-infinite inhomogeneities well enough to warrant its implementation into routine clinical use. However, we have found that Target may not be able to adequately predict dose distributions around smaller inhomogeneous inclusions. Further work will be required to investigate this potential problem.

The use of an universal wedge for asymmetric fields.

Beam collimators on newer linear accelerators may be collimated asymmetric to the central axis. The asymmetric beam has a non-flat profile adjusted to yield fields whose half widths are not symmetric about the central axis. While some treatment planning systems modify their programs to mimic the nonuniformity, ideally it is preferred to have a flat profile under the open beam. We have developed a universal wedge that can be used to flatten the field for a variety of jaw sizes and positions and energies for the Varian 2100C. The wedge flattens the field to +/- 3% over 80% of the field.