Index of content:
Volume 19, Issue 1, January 1992

A review of electronic portal imaging devices (EPIDs)
View Description Hide DescriptionOn‐line electronic portal imaging devices are beginning to come into clinical service in support of radiotherapy. A variety of technologies are being explored to provide real‐time or near real‐time images of patient anatomy within x‐ray fields during treatment on linear accelerators. The availability of these devices makes it feasible to verify treatment portals with much greater frequency and clarity than with film. This article reviews the physics of high‐energy imaging and describes the operation principles of the electronic portal imaging devices that are under development or are beginning to be used clinically.

Investigation of the nonlinear aspects of imaging through a highly scattering medium
View Description Hide DescriptionThe edge‐spread function that is expected when imaging a sharp edge in a highly scattering medium using a time‐of‐flight imaging system has been investigated. Experimental results have been compared with computer simulations. The effect of scattering is to broaden the image of the edge, accompanied by an offset of the midpoint (50% transmission) of the edge. Small random errors in the measurement of the time of flight results in an increase in the offset of the edge position at very short times of flight. The offset in the midpoint of the edge is evidence of the nonlinear nature imaging process, in view of which we conclude that nonlinear algorithms will be necessary to maximize the image information available from measured intensities.

Systematic bias in basis material decomposition applied to quantitative dual‐energy x‐ray imaging
View Description Hide DescriptionBasis material decomposition represents dual‐energy x‐ray attenuation measurements in terms of the attenuation coefficients or thicknesses of two standard materials which, when combined, produce attenuation equivalent to the object being measured. In tomographic imaging, the reconstructed attenuation coefficient is calculated in terms of the attenuation coefficients of the basis materials, while in projection imaging, the thicknesses of two materials can be specified in terms of the basis materials. This analysis shows that basis material decomposition is exact in a dual‐monoenergetic system, but for broad spectra, x‐ray beam hardening introduces a bias into quantitative measurements. The error is small enough that it can be ignored when dual‐energy imaging is used primarily to enhance the contrast of one material over another. The magnitude of the error in quantitative measurements depends on the details of the specific application including the energy of the x‐ray beam, and the composition and thickness of the materials included in the object. The magnitude of the error for dual‐energy bone densitometry has been analyzed using a first‐order propagation of error analysis and the calculations verified by computer simulation. This analysis shows that the magnitude of the systematic error can be as high as 3% for 1 g/cm^{2} of bone mineral when aluminum and acrylic basis materials are used for the calibration. This systematic error is eliminated when the basis materials are the same as the materials that are being quantified (i.e., bone mineral and water).

Evaluation of a new set of calibration standards for the measurement of fat content via DPA and DXA
View Description Hide DescriptionA simulation study was performed to evaluate a new set of calibration standards for estimating the fat content of the body via dual‐photon absorptiometry (DPA) and dual‐energy x‐ray absorptiometry (DXA). The standards, proposed by Nord and Payne [presented at the 2nd meeting of The Bath Conference on Bone Mineral Measurement (1990)] consist of stearic acid (100% fat) and 0.6% NaCl in water (100% lean). They were compared with other standards consisting of average composition adipose/muscle tissues and fatty adipose / lean muscle tissues. Source and detector properties of a Gd‐153 DPA system and three commercial DXA systems were modeled. For each system and calibration set, rms errors in the calculated fat contents of simulated tissues having fat mass percentages that ranged from about 4%–44% and thicknesses that ranged from 5–20 cm were determined. Beam hardening errors for the systems were evaluated as was a calibration technique employed by one of the manufacturers to correct for such errors. In general, the smallest rms errors (2% or less when the calibration standards and tissues were of equal thickness) were obtained with the average adipose/muscle standards. Equivalent results were obtained with standards consisting of stearic acid and 0.8% NaCl. The latter is a higher salt content than proposed by Nord and Payne and results from differences in the x‐ray attenuation coefficients that were employed in calculating the fat equivalence of water. Other, more convenient standards, such as lucite and water may be employed by using appropriate fat equivalences (∼69% for lucite and ∼10% for water). Beam hardening errors for the DXA systems are considerable, and the simulated correction technique was shown to be effective.

Fluoroscopic performance tests using a portable computer/frame grabber: Wiener spectra measurements
View Description Hide DescriptionCurrently, routine tests of fluoroscopic image quality in common use are highly subjective. As part of an effort to develop more quantitative routine tests of fluoroscopic image quality, a method was developed to quickly and easily measure Wiener spectra (WS) of TV‐viewed fluoroscopic systems that considers both spatial and temporal noise correlations. A PC‐mounted frame grabber captures images at the TV frame rate to form a three‐dimensional (3‐D) array of pixels. Scans of a ‘‘two‐dimensional slit’’ are then synthesized from which a one‐dimensional central section of a 3‐D WS is calculated. To avoid errors due to coarse (8‐bit) quantization, a video amplifier is used to expand a portion of the signal to the full digitizer range. A reference signal (2 mm of aluminum) is then used to normalize image contrast. Ensemble averages of 250 spectra were obtained in ∼1 min, including all processing. Results are presented to demonstrate reproducibility, sensitivity, and behavior of the WS. The eventual goal of this work is to use this method in conjunction with measurements of an MTF to calculate fundamental descriptors of image quality, such as SNR and NEQ.

Describing the signal‐transfer characteristics of asymmetrical radiographic screen‐film systems
View Description Hide DescriptionThe measurement of modulation transfer function for radiographic screen‐film systems depends critically upon a proper linearization of the measured line spread function. This is normally done by photographic photometry (i.e., using the measured density versus log exposure relationship to transform the density line spread function into an exposure line spread function). It has been long appreciated that this procedure may fail for asymmetrical dual screen systems that use film with emulsion coated on both sides of the support. The advent of asymmetrical and near‐zero crossover films that can be used with highly asymmetric screen pairs has prompted a reinvestigation of these concerns about the definition and measurement of modulation transfer function. For such cases, it is useful to define the contrast transfer function, which is a function of exposure and spatial frequency. When normalized by its zero frequency value the contrast transfer function can serve as the ‘‘effective MTF’’ for low‐contrast input signals in such systems. In the limit of symmetrical systems this quantity approaches the conventionally measuredMTF. The utility of this approach is demonstrated by applying it to a commercially available asymmetrical screen‐film combination.

Density related errors encountered with an electrostatically coupled film digitizer tablet
View Description Hide Description

Broad line quantitative chemical shift spectroscopy
View Description Hide DescriptionProton NMRspectroscopy was applied to quantitate the measurement of total body water/fat distributions i n v i v o. A special MR protocol was developed to excite a thick slab of tissue and display the magnitude NMR spectrum of the collected response signal. Very short echo time TE (8 ms) and long repetition time TR (4000 ms) were used to minimize relaxation damping of the signal intensities. The spectrum was then decomposed into individual lines and proton densities of different species were calculated. Proton density information was converted into weight percentage data using stoichiometrical and physiological information. The technique was validated using phantoms that contained different mixtures of water/maize oil. A high‐resolution NMR spectrum of maize oil samples was used to determine the stoichiometric information. The test results showed good agreement with the known composition of the phantom within the whole range of water content (0–100%). This method is very fast since no phase encoding of data is required. Preliminary results for monkeys show promising potential in clinical applications.

Automated lesion detection and lesion quantitation in MR images using autoassociative memory
View Description Hide DescriptionPrevious efforts concerning lesion extraction in radiologic images indicated that autoassociative memory models can be a valuable tool in automated lesion detection. Preliminary results are expanded to resolve the technical problems of image registration and magnification. Instead of operating on selected portions of the MR images, each entire image matrix is operated upon as image vector comprising all stacked columns of the matrix. Spin density weighted images (TR=3000 ms and TE=40 ms) of 42 normal subjects were remapped and standardized with respect to location and magnification. All image vectors were orthonormalized to span a linear manifold. Standardized abnormal image vectors were then tested by the stored autoassociative memory and the abnormalities (novelties) were extracted by application of an autocorrelation matrix to the input vector. The autocorrelation matrix is computed using image vectors from normal subjects. The lesions (multiple sclerosis and tumors) are then identified as the orthogonal component to the linear manifold spanned by the basis vectors of the normal brain scans. Lesion extraction has been achieved with the intention of quantitating and staging diseased parenchyma after automated edge detection.

Quantitative phase‐velocity MR imaging of in‐plane laminar flow: Effect of fluid velocity, vessel diameter, and slice thickness
View Description Hide DescriptionQuantitative MR phase imaging is frequently used to measure spin velocities. A potential difficulty may arise, however, when in‐plane phase images are acquired of a vessel carrying laminar flow, for which the fluid velocity profile is parabolic. In that case, depending on the flowvelocity (v), the vessel diameter (D), and the chosen MR slice thickness (ST), a spin velocity gradient will be present to some extent within each intraluminal voxel. The resulting intravoxel phase dispersion may be expected to affect the net pixel phase value, and hence compromise the assumed linear correlation between phase shift and velocity. In this study, the effects of alterations of v, D, and ST on the apparent image phase are investigated for the case of laminar flow directed parallel to the sequence read gradient. A theoretical model is developed and the conclusions experimentally tested using a flow phantom. The data demonstrate that when quantitating in‐plane phase‐flow images, significant velocity underestimations may occur when the net flow‐induced phase shifts are small and the MR slice thickness is an appreciable fraction of the vessel diameter.

Nuclear magnetic relaxation characterization of irradiated Fricke solution
View Description Hide DescriptionThe spin‐lattice relaxation rate of irradiated Fricke solution was studied as a function of the absorbed dose D. The R _{1} increases linearly with dose up to D∼400 Gy after which the response saturates. A model describing the R _{1} of a solution of either ferrous (Fe^{2} ^{+}) or ferric (Fe^{3} ^{+}) ions is presented; it is based on the spin relaxation of protons on water molecules in the bulk and protons on water molecules in the coordination shells of the ions with fast exchange occurring between the two water environments. All inherent relaxation parameters of the different proton groups are determined empirically at NMR frequencies of 9 and 25 MHz. An extension of the model is made to describe the spin‐lattice relaxation behavior of irradiated Fricke solution. Good agreement between model predictions and experimental results is observed. The model relates the spin‐lattice relaxation rate of a Fricke dosimeter to the chemical yield of ferric ion, thus potentially creating an absolute NMRdosimetry technique. Various practical aspects of the NMR‐Fricke system, such as the optimal initial ferrous concentration and the NMR frequency dependence of the sensitivity, are described.

Localized beta dosimetry of ^{131}I ‐labeled antibodies in follicular lymphoma
View Description Hide DescriptionThe purpose of this study is to assess the multicellular dosimetry of ^{131}I ‐labeled antibody in follicular lymphoma based on histological measurements on human tumor biopsy tissue. Photomicrographs of lymph node specimens were analyzed by first‐order treatment to determine the mean values and statistical variations of the radii of follicles (260±90 μm), interfollicular distances (740±160 μm), and the number density of follicles [60±18 in a volume of (2×1480 μm)^{3}]. Based on these measurements, two geometrical models were developed for localized beta dosimetry. The first, a regular cubic lattice model, assumes no variation in follicular radius of follicles and interfollicular distance. The second, a randomized distribution model, is a more complicated but more realistic representation of observed histological specimens. In this model,Monte Carlo methods were used to reconstruct the spatial distribution of follicles by simulating the distribution of the radii of follicles, interfollicular distances, and the number density of follicles. Dose calculations were performed using Berger’s point kernels for absorbed‐dose distribution for beta particles in water, assuming the ^{131}I ‐labeled antibodies as point sources. It was assumed that the activity concentration of the labeled antibody within the follicles was ten times the activity concentration in the interfollicular spaces. The spatial distribution of localized dose was calculated for a tumor having an average dose of 40 Gy. The localized dose was found to be highly nonuniform, ranging from 20 to 90 Gy, and varying by a factor of about 2 from the average tumor dose. Most of the tissue (70%–80% by volume) was found to receive a lower dose than the average tumor dose. No significant difference in localized dose distribution was found between the regular cubic lattice model and the randomized distribution model, suggesting that sometimes a simple geometrical model may be adequate as a starting point for dosimetry of more complex situations. Finally, the localized dose distribution is discussed in terms of its usefulness for extracting information such as the mean dose, the spatial variation of dose, and the fraction of tissue receiving various absorbed doses. An approach for applying localized dosimetry to improve the estimated cell‐killing efficiency is suggested.

The spatial and energy dependence of bremsstrahlung production about beta point sources in H_{2}O
View Description Hide DescriptionA Monte Carlo simulation was performed to characterize the spatial and energy distribution of bremsstrahlung radiation from β point sources important to radioimmunotherapy (RIT). Using the EGS4 Monte Carlo code, the isotropic emission and transport of monoenergetic 0.1‐, 0.5‐, 1.0‐, 2.0‐, and 3.0‐MeV electrons and ^{3} ^{2}P and ^{9} ^{0}Y β particles was simulated in an infinite, homogeneous H_{2}O phantom. The probability of bremsstrahlung production for each Monte Carlo‐simulated electron step was accumulated in energy intervals not exceeding 5 keV and stored as a function of radial position. To validate this scheme, the EGS4 code was tested in the continuous slowing down approximation (csda) mode, with resulting radiation yields seen to agree with values in ICRU Report No. 37 (ICRU, Bethesda, MD, 1984) to better than 1.6%. The radiation yield calculated with the simulation of secondary particles is seen to be 3%–5% greater than the csda yield. The photon energy distributions are characterized by a typically broad bremsstrahlung spectrum with the probability of photon generation decreasing with radial distance. In the energy range 0.05–0.511 MeV, the probability for bremsstrahlung production from ^{9} ^{0}Y (2.76×10^{−} ^{2} decay^{−} ^{1}) is twice that from ^{3} ^{2}P (1.35×10^{−} ^{2} decay^{−} ^{1}). When passed through 10 cm of H_{2}O and put upon a standard NaI scintillation camera, count rates of 2.3×10^{−} ^{6} and 1.2×10^{−} ^{6} counts s^{−} ^{1} Bq^{−} ^{1} are estimated from point sources of ^{9} ^{0}Y and ^{3} ^{2}P. These results predict the inherent spatial resolution limitation and provide the initial data required for modeling and analyzing the scatter, attenuation, and image formation processes in quantitative imaging of bremsstrahlung for RIT dosimetry.

Electron bolus design for radiotherapy treatment planning: Bolus design algorithms
View Description Hide DescriptionComputer algorithms to design bolus for electron beamradiotherapytreatment planning were investigated. Because of the significant electron multiple scatter, there is no unique solution to the problem of bolus design. However, using a sequence of operators, a bolus can be designed that attempts to meet three important criteria: adequate dose delivery to the target volume, avoidance of critical structures, and dose homogeneity within the target volume. Initial calculation of bolus shape was based upon creation operators forcing either the physical or the effective depths of the distal surface of the target volume to a specified value. Modification operators were then applied to the bolus to alter the shape to better meet the design criteria. Because the operators each address a single dosimetric issue, they can often adversely affect some other attribute of the dose distribution. In addition, an extension operator is used to design the bolus thickness outside the target volume. Application of these operators is therefore carried out in certain sequences and each may be used more than once in the design of a particular bolus. The effects of these operators on both the bolus and the resulting dose distribution are investigated for test geometries and patient geometries in the nose, parotid, and paraspinal region.

Optimal electron‐beam treatment planning for retinoblastoma using a new three‐dimensional Monte Carlo‐based treatment planning system
View Description Hide DescriptionElectron‐beam treatment planning for retinoblastoma was investigated and an optimal treatment plan was devised for a particular case using a new three‐dimensional Monte Carlo‐based treatment planning system known to be capable of correctly predicting dose perturbations caused by body surface obliquities and tissue heterogeneities. Computed tomography(CT) data files were used to construct a three‐dimensional eye phantom representing the anatomy of a child’s orbit. Dose distributions in sagittal, transverse, and coronal planes were predicted with 1‐mm resolution. Study of these distributions led to an optimal treatment plan consisting of an anterior‐lateral pair, with the anterior field being a 10‐MeV, 30‐mm‐diam circular field, centrally blocked by a 10‐mm‐diam lucite lens shield and the lateral field being a 16‐MeV, 30×25‐mm D‐shaped field. The anterior field delivers a therapeutic dose to the ora serrata, but it underdoses the posterior retinal surface behind the lens shield; the lateral field provides the necessary boost dose to the posterior retinal surface. An equally weighted combination of the two fields produces a dose distribution in which the entire retinal surface receives a therapeutic dose, with less than 10% of that dose being delivered to the lens, brain, and the contralateral orbit.

Differential‐pencil‐beam dose calculations for charged particles
View Description Hide DescriptionThe use of a convolution or differential‐pencil‐beam (DPB) algorithm has been studied for charged‐particle dose calculations as a means of more accurately modeling the effects of multiple scattering. Such effects are not reflected in current charged‐particle dose calculations since these calculations rely on depth‐dose data measured in homogeneous water‐equivalent phantoms and use ray‐tracing techniques to calculate the water‐equivalent pathlength from patient CT data. In this study, isodose plots were generated from three‐dimensional dose calculations using Monte Carlo, DPB, and standard ray‐tracing methods for a 4‐cm modulated 150‐MeV proton beam incident on both homogeneous and heterogeneous phantoms. To simulate therapy conditions with charged particles, these studies included cases where compensating boluses were introduced to modify the particle range across the treatment field. Results indicate that multiple‐scattering effects, including increased penumbral width as a function of beam penetration and the ‘‘smearing’’ of isodose distributions downstream from complex heterogeneities, are well modeled by the DPB algorithm. The DPB algorithm may also be used to obtain more useful estimates of the dose uncertainty in regions near the end of the beam’s range downstream from complex heterogeneities than can be derived from standard ray‐tracing calculations.

Some considerations regarding w values for heavy charged‐particle radiotherapy
View Description Hide DescriptionThe AAPM recommendations regarding the w values for heavy charged‐particles radiotherapy are discussed. (AIP)

An interactive beam‐weight optimization tool for three‐dimensional radiotherapy treatment planning
View Description Hide DescriptionA computer software tool has been developed to aid the treatment planner in selecting beam weights for three‐dimensional radiotherapytreatment planning. The program consists of a feasibility search algorithm embedded in an interactive, user‐friendly driving program. The feasibility search algorithm is based on the iterative relaxation algorithm of Cimmino [L a R i c e r c a S c i e n t i f i c a, Vol. I, pp. 326–333 (1938)] as applied to the radiotherapyinverse problem by Altschuler e t a l. [Med. Phys. 1 3, 590 (1986)]. Relative importances of structures based upon clinical considerations can be incorporated into the algorithm. In order to speed convergence, the relaxation parameter is made to vary, with its value based upon a measure of deviation from feasibility. The interactive driving program is designed so that the treatment planner can make reasonable judgments regarding the acceptability of a plan in the event that the dose constraints yield no feasible solution. An example of the use of this program applied to a problem in three‐dimensional radiotherapytreatment planning is illustrated.

Generation and use of measurement‐based 3‐D dose distributions for 3‐D dose calculation verification
View Description Hide DescriptionA 3‐D radiation therapytreatment planning system calculates dose to an entire volume of points and therefore requires a 3‐D distribution of measured dose values for quality assurance and dose calculation verification. To measure such a volumetric distribution with a scanning ion chamber is prohibitively time consuming. A method is presented for the generation of a 3‐D grid of dose values based on beam’s‐eye‐view (BEV) film dosimetry. For each field configuration of interest, a set of BEV films at different depths is obtained and digitized, and the optical densities are converted to dose. To reduce inaccuracies associated with film measurement of megavoltage photon depth doses,doses on the different planes are normalized using an ion‐chamber measurement of the depth dose. A 3‐D grid of dose values is created by interpolation between BEV planes along divergent beam rays. This matrix of measurement‐based dose values can then be compared to calculations over the entire volume of interest. This method is demonstrated for three different field configurations. Accuracy of the film‐measured dose values is determined by 1‐D and 2‐D comparisons with ion chamber measurements. Film and ion chamber measurements agree within 2% in the central field regions and within 2.0 mm in the penumbral regions.

Coordinate transformation as a primary representation of radiotherapy beam geometry
View Description Hide DescriptionAn approach to both geometric specification of radiotherapy beams and computerized solution of geometric treatment planning problems using coordinate transformations is presented. It is demonstrated that the specification of the geometric relationship of a treatment beam to a patient can be uniquely given by a 4×4 coordinate transformation matrix, and that the matrix representation can be translated from (and to) the more conventional machine‐based specification of geometry. This approach enables a compact representation of the patient/beam geometry which is independent of the specific labeling conventions of the treatment machine and which can be directly exploited in the solution of treatment planning problems. Beam geometry can be easily described either in terms of the natural degrees of freedom of a therapy machine or in terms of alternative, problem‐specific frames of reference. The ability to use these various frames of reference interchangeably allows the designer of treatment design software to present appropriate task‐specific user interfaces for arbitrarily complex tasks, and thus reduce the cognitive burden on users of the software.