Effect of low-density lateral interfaces on soft-tissue doses.

PURPOSE: Doses at the interface between tissue and low-density inhomogeneities with the interface positioned perpendicular to the beam direction have been well studied. When the inhomogeneity lies parallel to the beam direction (i.e., a lateral interface), the resulting dose distribution is not as well known. Lateral lung-soft-tissue interfaces are common in many fields used to treat malignancies in the thorax region including tangential breast fields and anteroposterior fields for lung and esophageal cancer. The purpose of this study was to evaluate the dose distribution along lateral interfaces and to determine the implications for treatment. METHODS AND MATERIALS: A polystyrene and cork slab phantom was irradiated from the side to simulate treatment fields with lateral lung-soft-tissue interfaces. The beam was positioned with the isocenter in polystyrene and the field edge in cork. Cork slabs (0.6-2.5 cm) were used to simulate different thicknesses of lung between the field edge and the target volume. Measurements were made using a parallel plate ionization chamber. With the chamber position held constant, polystyrene slabs were added between the cork and the chamber to study the dose distribution in the interface region. Interface doses were studied as a function of the amount of cork in the field, field size, beam energy (6-18 MV), and depth. RESULTS: Doses in the interface region were lower by as much as 10% compared to doses in a homogeneous phantom. For a given cork width and field size, the magnitude of the underdose increased by several percent as the x-ray energy increased from 6 to 18 MV. The underdose at the interface was 5% for 6 MV and 8% for 18 MV X-rays with a 1-cm cork width. For a 2.5-cm cork width, underdoses of 2.5% and 3% at distances up to 2.5 and 4 mm lateral to the interface were observed for 6- and 18-MV X-rays, respectively. However, doses right at the interface were 1% greater for 6 MV and 3% less for 18 MV than doses in a homogeneous phantom. For a given cork width, the interface doses were not significantly dependent on field width but decreased by an additional 2-3% as the length decreased to 4 cm. Additional decreases were also observed when the measurement depth decreased to 3 cm. With a 1-cm width of cork in the field, a lateral distance of 3-4 mm from the interface was necessary to ensure doses of at least 98% of the homogenous dose with 6-MV X-rays. A lateral distance of 6-7 mm was necessary for 10- and 18-MV X-rays. CONCLUSION: Underdosing will occur in the soft tissues adjacent to low-density inhomogeneities. The magnitude depends primarily on the width of the inhomogeneity seen in the treatment field, but also on field size, depth, and beam energy. For treatment fields with a lateral lung interface, a segment of tissue approximately 3-4 mm thick for 6 MV and 6-7 mm thick for higher-energy beams may be underdosed. Lung widths of > or = 1.75 cm as observed on film will generally guarantee doses of at least 96% of those calculated with no inhomogeneity corrections. High-energy beams are often used to treat sites in the thorax or breast to improve dose homogeneity throughout the treatment volume. Potential underdosing due to the presence of lung should be considered and may require a decrease in beam energy or an increase in the margin between the target volume and the field edge to ensure adequate treatment.

An evaluation of setup uncertainties for patients treated to pelvic sites.

PURPOSE: Successful delivery of conformal fields requires stringent immobilization and treatment verification, as well as knowledge of the setup reproducibility. The purpose of this study was to compare the three-dimensional distribution of setup variations for patients treated to pelvic sites with electronic portal imaging devices (EPID) and portal film. METHODS AND MATERIALS: Nine patients with genitourinary and gynecological cancers immobilized with custom casts and treated with a four-field whole-pelvis technique were imaged daily using an EPID and filmed once every five to seven treatments. The three-dimensional translational and rotational setup errors were determined using a technique that relies on anatomical landmarks identified on simulation and treatment images. The distributions of the translational and rotational variations in each dimension as well as the total displacement of the treatment isocenter from the simulation isocenter were determined. RESULTS: Grouped analysis of all patients revealed average unidirectional translational deviations of less than 2 mm and a standard deviation of 5.3 mm. The average total undirected distance between the treatment and simulated isocenters was 8.3 mm with a standard deviation of 5 mm. Individual patient analysis revealed eight of nine patients had statistically significant nonzero mean translational variations (p < 0.05). Translational variations measured with film were an average of 1.4 mm less than those measured with EPID, but this difference was not statistically significant. CONCLUSION: Translational variations measured in this study are in general agreement with previous studies. The use of the EPID in this study was less intrusive and may have resulted in less additional attention being given each imaging setup. This may explain the slightly larger average translational variations observed with EPID vs. film, and suggests that the use of EPIDs is a superior method for assessing the true extent of setup displacements. Although no statistically significant translational variations for the patient group overall were observed, 90% of patients had significant translational variations in at least one direction when analyzed separately. The margin to be added to the clinical target volume (CTV) to account for setup uncertainties will depend on whether it is possible to identify patients with significant translational variations, and to eliminate these displacements from routine treatments. The choice to eliminate these variations and to use a smaller CTV margin will have to be accompanied by stringent frequent position verification methods and repositioning.

Bremsstrahlung review: an analysis of the Schiff spectrum.

A concise approximate formula computed by Schiff for the intensity spectrum of bremsstrahlung photons has been a valuable starting point for many medical physics applications, including the Task Group 21 protocols. This paper provides a brief review of the literature related to determination of the bremsstrahlung spectrum and to the Schiff formula in particular. It describes the approximations Schiff made to obtain this formula, including the Born approximation, and the exponential nuclear screening potential, the infinite-mass nucleus approximation, and the "extreme relativistic" approximation. A derivation of a more exact formula that avoids the last of these approximations is presented. This provides a check on the accuracy of the Schiff spectrum for linear accelerator energies used clinically. Comparisons with the Schiff thin-target result are presented. A thick-target bremsstrahlung spectrum is calculated and compared with the forward spectrum obtained from Monte Carlo simulations of the x-ray production in two linear accelerator treatment heads.

An analytic calculation of the energy fluence spectrum of a linear accelerator.

Calculation of photon dose by convolution methods requires a knowledge of the fluence spectrum of photons produced by the linear accelerator treatment head. But this spectrum is very difficult to measure accurately, and is often derived by Monte Carlo calculations modeling the elements of the treatment head. In this paper an analytic calculation technique and the corresponding computer tool for calculating the photon energy fluence spectral distribution at any point in a bremsstrahlung field are presented. Primary bremsstrahlung photon distributions are computed by modeling electron dispersion in layers in a thick target and using the thin target bremsstrahlung cross section formulas of Schiff. The first Compton scatter from all materials in a linear accelerator treatment head is computed analytically. Higher-order Compton scatter events and pair production annihilation photons are ignored, but the attenuation of both primary and first scattered photon fluence is computed. Predictions of the computer implementation of the model are compared to measurements of bremsstrahlung production in a thick target and to Monte Carlo calculations of the energy fluence emerging from a linear accelerator. Finally, the computer tool is used to investigate the source of collimator-dependent fluence fluctuations in air. In agreement with other measurements, the principal contribution to fluence outside the geometric field is found to be from scatter in the flattening filter.

Rotational kernels for conformal therapy.

Conformal therapy treatment planning involves determining an irradiation strategy to deliver a dose distribution which is optimized for a given tumor volume. A technique proposed by Brahme restricts the search for the optimal treatment strategy to x-ray distributions in which points within the target volume are irradiated uniformly under rotation of the beam source around the patient. The dose distribution in this case may be calculated as a convolution of an irradiation weighting distribution with an invariant kernel. A procedure is described for calculating three-dimensional kernels to be used for clinical treatment planning with x rays produced by an electron accelerator. The convolution kernel is calculated as the sum of pencil beams irradiating the center of a cylindrical phantom uniformly from all angles. The shape of the kernel at points off the center of the phantom is investigated by means of numerical calculations which support the assumption that the kernel is invariant with respect to position within the phantom. The calculated kernels are verified by comparison with experimentally measured rotational arc dose distributions.

Potential and limitations of invariant kernel conformal therapy.

A treatment planning methodology was developed to investigate the invariant kernel form of conformal therapy proposed by Brahme. Three-dimensional dose distributions were calculated by convolving a rotationally symmetric, invariant kernel with weighting distributions. Fourier transform convolution techniques implemented on an array processor were used to achieve high calculation speeds, thereby allowing iterative techniques in the spatial and frequency domains for computing dose distributions that asymptotically approach a desired dose distribution. To use rotationally symmetric kernels, the generality of the solution is traded for a fast, deterministic, inverse planning approach. The limitations imposed on the dose distributions by this loss of generality are characterized and tentative conclusions are drawn about the potentials and limits of clinical application of this form of the methodology. Further developments of the concept are suggested.

Validation of a new virtual wedge model.

Results of a validation study of a commercial virtual wedge device recently installed at our institution are presented. The wedge simulation produces an energy fluence from the treatment head that is equivalent to the primary energy fluence attenuated through a wedge-shaped slab of water with the central axis fluence set to unity. A simple exponential formula used to compute off-axis wedge factors is compared to beam profiles measured in a water phantom. A fast Fourier transform (FFT) convolution dose calculation is compared to measured dose profiles. Measured and calculated central axis wedged/open field ratios as a function of depth are also compared.

Beam characteristics of a retrofitted double-focused multileaf collimator.

Multileaf collimators (MLCs) are generally believed to be convenient and cost-effective tools for intensity modulation and conformal therapy. They are becoming a standard feature on new accelerators; however, the older units can be retrofitted with modern MLCs. Before such a unit can be clinically used, the beam characteristics must be verified. In this study the beam characteristics of a Siemens double-focused MLC retrofitted to an MD2 linear accelerator are presented. The head leakage along with inter- and intra-leaf radiation transmission were measured using film. The collimator (Sc), phantom (Sp), total (Scp) scatter factors, central axis depth dose, beam profiles for off-axis ratios, penumbra, and surface dose were evaluated for square, rectangular, and irregularly shaped fields. The maximum head leakage was estimated to be < 0.05% in any plane at a distance of 1 m and maximum transmission through the MLC leaves was estimated to be < 1.4% and < 1.1% for the 10 MV and 6 MV beams, respectively. The maximum differences between pre- and post-MLC installation data for the Sc and Scp were < or = 0.7% and < or = 1.4%, respectively. Similarly, the percent depth dose data for all fields and both beam energies were within 1.5% of the original data. The beam profiles measured at various depths were also in agreement with those of the pre-MLC installation data. The measured beam penumbra (20%-80%) showed a range of 7.8 mm-11.0 mm for the 6 MV and 8.4 mm-11.1 mm for the 10 MV beams from smallest to largest fields. These ranges differ by less than a millimeter from those of the old data. The surface dose measurements were slightly lower than the conventional jaw values suggesting that MLC does not produce significant electron contamination. It is concluded that the retrofitted MLC maintains the integrity of the original beam and may provide a cost-effective conformal therapy.

Dose distributions of x-ray fields as shaped with multileaf collimators.

Multileaf collimators (MLC) with various blade widths were simulated using standard cerrobend blocks, and three-dimensional dose computations were carried out to study the resultant radiation field edges. Film measurements made with 6 and 18 MV x-ray beams were compared with calculations that employed a three-dimensional Fourier convolution. A spatial accuracy of better than 3 mm was found in the 50% isodose line of the penumbral region with a calculation voxel size of 5 mm x 5 mm x 5 mm. The computer simulation was used to study the deviation of the calculated 50% isodose line from the desired geometric field edge using various MLC blade positions. The study suggests that multileaf collimation to the outside of the desired field edge will lead to overdose outside the field, whereas multileaf collimation to the inside of the desired field edge will lead to underdose inside the field. When the direction of travel of the leaves with respect to the field edge is near 45 degrees, the 50% isodose of a multileaf-collimated beam will fall close to the desired edge with no underdose when the leaf corners are allowed to insert into the desired field edge by 1.2 mm for 6 MV x-rays and 1.4 mm for 18 MV x-rays using a 1 cm wide leaf. These blade offsets account for the scattering of photons and electrons in the medium within the penumbral region.

Radiographic technique for acquisition of external patient contour.

Despite advances in techniques for contour acquisition, most patients are still contoured manually using lead wire. The lead wire method has unpredictable inherent inaccuracies due to the fact that the wire must be removed from the patient to obtain the contour. A simple radiographic method for a quick and accurate contour is presented that eliminates such an operator error. The technique requires that a thin lead wire be placed on the patient's skin in the desired plane as in the manual contouring process. With the wire still in place, the simulator table is rotated 90 degrees, and then the gantry is rotated to any reasonable angle (30-45 degrees) for a simple radiograph. The source-to-film distance (SFD) and gantry angle are noted. The radiograph of the lead wire contour is digitized on a personal computer (PC) and transformed from the film coordinates to the patient coordinates using the SFD and the gantry angle as input parameters. The accuracy and efficacy of this method were tested in four test cases: a square box, and contours in the head and neck, thoracic, and pelvic regions. The accuracy of the calculated contour was within +/- 4 mm for the large-gradient contours and +/- 3 mm in the other cases. A root mean square deviation of < 2 mm was observed between this method and the CT contour. This method can be used in any clinical setting where a simulator and a digitizer with a PC are available. The transformation code is simple and could be placed on any treatment planning computer system.