Table of contents

Various forms of GAFChromic film have been used for several years as radiographic media for measuring dose distributions of brachytherapy sources and small radiation fields. Upon irradiation the film changes colour and darkens with time post-irradiation. The darkening is most rapid in the first 24 h, and it has been suggested that for accurate dosimetry a waiting period of 24 h should occur before any optical density (OD) measurements are taken. A more rapid colour stabilization (RCS) procedure has been developed and is evaluated. The procedure consists of heating the film post-irradiation for a period of 2 h at C. The RCS procedure is compared with a control group and the dose response is tested for linearity, stability and reproducibility using two densitometers with light sources at different wavelengths (632.8 nm and 671 nm). The rise in net optical density (NOD) for the period 3-168 h is less than 3% for the RCS group as compared with 12% for the controls. In the first 24 h, the increase in NOD for the RCS samples is less than 0.5%, as opposed to 6% for the control group.

A new method of calibrating gel dosimeters (applicable to both Fricke and polyacrylamide gels) is presented which has intrinsically higher accuracy than current methods, and requires less gel. Two test-tubes of gel (inner diameter 2.5 cm, length 20 cm) are irradiated separately with a field end-on in a water bath, such that the characteristic depth-dose curve is recorded in the gel. The calibration is then determined by fitting the depth-dose measured in water, against the measured change in relaxivity with depth in the gel. Increased accuracy is achieved in this simple depth-dose geometry by averaging the relaxivity at each depth. A large number of calibration data points, each with relatively high accuracy, are obtained. Calibration data over the full range of dose (1.6-10 Gy) is obtained by irradiating one test-tube to 10 Gy at dose maximum , and the other to 4.5 Gy at . The new calibration method is compared with a `standard method' where five identical test-tubes of gel were irradiated to different known doses between 2 and 10 Gy. The percentage uncertainties in the slope and intercept of the calibration fit are found to be lower with the new method by a factor of about 4 and 10 respectively, when compared with the standard method and with published values. The gel was found to respond linearly within the error bars up to doses of 7 Gy, with a slope of and an intercept of Gy. For higher doses, nonlinear behaviour was observed.

The use of an ionization chamber for absorbed dose determinations in a medium requires one to take into account perturbation corrections due to the presence of the chamber cavity in the medium. Evaluation of these corrections for perturbation and their variation with depth in the medium has been performed for a flat cylindrical and a cylindrical (thimble-type) ionization chamber placed in a graphite phantom irradiated by a gamma beam using Monte Carlo calculations (EGS4 system with correlated sampling variance reduction technique). The results of these calculations agree with published experimental and theoretical data to better than 0.18%, with a statistical uncertainty of less than 0.17%.

The accuracy and traceability of the calibration of radiotherapy dosimeters is of great concern to those involved in the delivery of radiotherapy. It has been proposed that calibration should be carried out directly in terms of absorbed dose to water, instead of using the conventional and widely applied quantity of air kerma. In this study, the faithfulness in disseminating standards of both air kerma and absorbed dose to water were evaluated, through comparison of both types of calibration for three types of commonly used radiotherapy dosimeters at ⁶⁰Co gamma beams at a few secondary and primary standard dosimetry laboratories (SSDLs and PSDLs). A supplementary aim was to demonstrate the impact which the change in the method of calibration would have on clinical dose measurements at the reference point. Within the estimated uncertainties, both the air kerma and absorbed dose to water calibration factors obtained at different laboratories were regarded as consistent. As might be expected, between the SSDLs traceable to the same PSDL the observed differences were smaller (less than 0.5%) than between PSDLs or SSDLs traceable to different PSDLs (up to 1.5%). This can mainly be attributed to the reported differences between the primary standards. The calibration factors obtained by the two methods differed by up to about 1.5% depending on the primary standards involved and on the parameters of calculation used for ⁶⁰Co gamma radiation. It is concluded that this discrepancy should be settled before the new method of calibration at ⁶⁰Co gamma beams in terms of absorbed dose to water is taken into routine use.

A new dosimetric quantity, the lateral build-up ratio (LBR), has been introduced to calculate depth dose distribution for any shaped field. Factors to account for change in incident fluence with collimation are applied separately. The LBR data for a small circular field are used to extract radial spread of the pencil beam, , as a function of depth and energy. By using the relationship between LBR, , energy and depth, a formalism is developed to calculate dose per monitor unit for any shaped field. Criteria for lateral scatter equilibrium are also developed which are useful in clinical dosimetry.

Small-field and stereotactic radiosurgery (SRS) dosimetry with radiation detectors, used for clinical practice, have often been questioned due to the lack of lateral electron equilibrium and uncertainty in beam energy. A dosimetry study was performed for a dedicated 6 MV SRS unit, capable of generating circular radiation fields with diameters of 1.25-5 cm at isocentre using the BEAM/EGS4 Monte Carlo code. With this code the accelerator was modelled for radiation fields with a diameter as small as 0.5 cm. The radiation fields and dosimetric characteristics (photon spectra, depth doses, lateral dose profiles and cone factors) in a water phantom were evaluated. The cone factor for a specific cone c at depth d is defined as , where is the reference cone. To verify the Monte Carlo calculations, measurements were performed with detectors commonly used in SRS such as small-volume ion chambers, a diamond detector, TLDs and films. Results show that beam energies vary with cone diameter. For a 6 MV beam, the mean energies in water at the point of maximum dose for a 0.5 cm cone and a 5 cm cone are 2.05 MeV and 1.65 MeV respectively. The values of obtained by the simulations are in good agreement with the results of the measurements for most detectors. When the lateral resolution of the detectors is taken into account, the results agree within a few per cent for most fields and detectors. The calculations showed a variation of with depth in the water. Based on calculated electron spectra in water, the validity of the assumption that measured dose ratios are equal to measured detector readings was verified.

The implementation of intensity modulated radiotherapy by dynamic multileaf collimator control involves the use of interpreter software which creates leaf trajectory plans for each leaf pair. Interpreter software for use with an Elekta SL15 linear accelerator and dedicated multileaf collimator has been written and tested. In practice the ideal trajectory plans often predict contact between leaves from opposing leaf banks, but this is prohibited by control software on the Elekta system as it could lead to mechanical damage. If the modulation within the geometric limits of a shaped field is not to be compromised then strategies to avoid leaf contact result in additional unwanted doses outside the geometric edge. The magnitude of any such additional dose can be reduced to acceptable levels by a technique which we have called rectangular edge synchronization. The performance of interpreter software which incorporates rectangular edge synchronization has been compared with that of potentially more efficient software which does not. The option containing the rectangular edge synchronization algorithm was shown to work consistently well at high monitor unit rates, and without incurring leaf contacts, under a wide range of test conditions. It therefore provides a sound basis for using intensity modulation to replace mechanical wedges, to simulate customized patient shape compensators, or to implement the results of inverse treatment planning processes that require superimposed intensity modulated beams.

Highly conformal dose distributions can be generated by intensity-modulated radiotherapy. Intensity-modulated beams (IMBs) are generally determined by inverse-planning techniques designed to maximize conformality. Usually such techniques apply no constraints on the form of the IMBs which may then develop fine-scale modulation. In this paper we present a technique for generating smoother IMBs, which yields a dose distribution almost identical to that without the constraint on the form of the IMBs. The method applies various filters successively at intervals throughout the iterative inverse planning. It is shown that the IMBs so determined using a simple median window filter have desirable properties in terms of increasing the efficiency of delivery by the dynamic multileaf collimator method and may be `more like conventional beams' than unconstrained, highly modulated IMBs.

The gamma unit is used to irradiate a target within the brain. During such a treatment many parameters, including the number of shots, the coordinates, the collimator size and the weight associated with each shot, affect the amount of dose delivered to the target volume and to the surrounding normal tissues. Hence it is not easy to determine an appropriate set of these parameters by a trial and error method. For this reason, we present here an optimization method to determine mathematically those parameters. This method is composed of two steps: firstly, a quasi-Newton method is used to deal with the continuous variables such as position and weight of shots; the result obtained at the end of this step then serves as the initial configuration for the next step, in which a simulated annealing method is applied to optimize all the aforementioned parameters. Application of the proposed methods to two examples shows that our optimization algorithm runs in a satisfactory way.

Detailed quality control (QC) protocols are a necessity for modern radiotherapy departments. The established QC protocols for treatment planning systems (TPS) do not include recommendations on the advanced features of three-dimensional (3D) treatment planning, like the dose volume histograms (DVH). In this study, a test protocol for DVH characteristics was developed. The protocol assesses the consistency of the DVH computation to the dose distribution calculated by the same TPS by comparing DVH parameters with values obtained by the isodose distributions. The computation parameters (such as the dimension of the computation grid) that are applied to the TPS during the tests are not fixed but set by the user as if the test represents a typical clinical case. Six commercial TPS were examined with this protocol within the frame of the EC project Dynarad (Biomed I). The results of the intercomparison prove the consistency of the DVH results to the isodose values for most of the examined TPS. However, special attention should be paid when working with cases of adverse conditions such as high dose gradient regions. In these cases, higher errors are derived, especially when an insufficient number of dose calculation points are used for the DVH computation.

Earlier work on the dielectric properties of cryoprotectant agents is extended using a modified measuring cell and method of data reduction yielding results more efficiently over a temperature range from -75 to C at frequencies from 27 MHz to 3 GHz. Dimethyl sulphoxide and butanediol are investigated at lower aqueous concentrations than previously and the implications for electromagnetic rewarming of cryopreserved materials are discussed.

Recurrences of malignant tumours in the chest wall are proposed as a valuable model of tissue mainly perfused by small size vessels (the so-called `phase III' vessels). Invasive thermal measurements have been performed on two patients affected by cutaneous metastasis of malignant tumours during hyperthermic sessions. Thermal probes were inserted into catheters implanted into the tissue at different depths. In one of the catheters a probe connected with laser-Doppler equipment was inserted to assess blood perfusion in the tumour periphery. The perfusion was monitored throughout the sessions, and a noticeable temporal variability was observed. The effect of the perfusion on the thermal map in the tissue was evaluated locally and the `effective conductivity' of the perfused tissue was estimated by means of the numerical integration of the `bio-heat' equation.

The tumour temperature, at the site where the perfusion probe is located, can be predicted by the numerical model provided two free parameters, and , are evaluated with a fitting procedure. is related to the effective conductivity and to the SAR term of the bio-heat equation. The model aimed at estimating the `effective conductivity' of the perfused tissue, and average values of of in Patient 1 and of in Patient 2 were obtained throughout the treatment.

However, when the average temperature in a larger tumour volume is to be predicted but only a single, `local' measurement of the perfusion is available and is assumed to be representative for the whole region, the model results are far less satisfactory. This is probably due to the fact that changes of blood perfusion throughout hyperthermic sessions occur to different extents within the tumour volume, and the differences in perfusion cannot be ignored.

The above result suggests that, in addition to the `temperature map', also a `perfusion map' within the heated volume should be monitored routinely throughout hyperthermic sessions.

Phase contrast x-ray imaging is a powerful technique for the detection of low-contrast details in weakly absorbing objects. This method is of possible relevance in the field of diagnostic radiology. In fact, imaging low-contrast details within soft tissue does not give satisfactory results in conventional x-ray absorption radiology, mammography being a typical example. Nevertheless, up to now all applications of the phase contrast technique, carried out on thin samples, have required radiation doses substantially higher than those delivered in conventional radiological examinations. To demonstrate the applicability of the method to mammography we produced phase contrast images of objects a few centimetres thick while delivering radiation doses lower than or comparable to doses needed in standard mammographic examinations (typically mean glandular dose (MGD)).

We show images of a custom mammographic phantom and of two specimens of human breast tissue obtained at the SYRMEP bending magnet beamline at Elettra, the Trieste synchrotron radiation facility. The introduction of an intensifier screen enabled us to obtain phase contrast images of these thick samples with radiation doses comparable to those used in mammography. Low absorbing details such as thick nylon wires or thin calcium deposits within breast tissue, invisible with conventional techniques, are detected by means of the proposed method. We also find that the use of a bending magnet radiation source relaxes the previously reported requirements on source size for phase contrast imaging. Finally, the consistency of the results has been checked by theoretical simulations carried out for the purposes of this experiment.

A genetic algorithm (GA) based feature selection method was developed for the design of high-sensitivity classifiers, which were tailored to yield high sensitivity with high specificity. The fitness function of the GA was based on the receiver operating characteristic (ROC) partial area index, which is defined as the average specificity above a given sensitivity threshold. The designed GA evolved towards the selection of feature combinations which yielded high specificity in the high-sensitivity region of the ROC curve, regardless of the performance at low sensitivity. This is a desirable quality of a classifier used for breast lesion characterization, since the focus in breast lesion characterization is to diagnose correctly as many benign lesions as possible without missing malignancies. The high-sensitivity classifier, formulated as the Fisher's linear discriminant using GA-selected feature variables, was employed to classify 255 biopsy-proven mammographic masses as malignant or benign. The mammograms were digitized at a pixel size of mm, and regions of interest (ROIs) containing the biopsied masses were extracted by an experienced radiologist. A recently developed image transformation technique, referred to as the rubber-band straightening transform, was applied to the ROIs. Texture features extracted from the spatial grey-level dependence and run-length statistics matrices of the transformed ROIs were used to distinguish malignant and benign masses. The classification accuracy of the high-sensitivity classifier was compared with that of linear discriminant analysis with stepwise feature selection . With proper GA training, the ROC partial area of the high-sensitivity classifier above a true-positive fraction of 0.95 was significantly larger than that of , although the latter provided a higher total area under the ROC curve. By setting an appropriate decision threshold, the high-sensitivity classifier and correctly identified 61% and 34% of the benign masses respectively without missing any malignant masses. Our results show that the choice of the feature selection technique is important in computer-aided diagnosis, and that the GA may be a useful tool for designing classifiers for lesion characterization.

In order to assess the accuracy of peripheral QCT (Stratec XCT 960) we analysed scans of the European Forearm Phantom and another phantom consisting of HPO encased in aluminium tubes to simulate cortical walls. Additionally 14 cadaveric forearm specimen scans were compared to CT scans acquired on a GE9800Q. The accuracy for density assessment of the European Forearm Phantom was better than 3%. A small increase in density was observed with increasing thickness of the aluminium wall (10% for each mm). Density measurements within the wall were confounded by limited spatial resolution. For a thickness of less than 4 mm, the density within the wall was underestimated by up to 40%. The measurement of mineral content was not influenced by this effect and showed an accuracy error of less than 6%. The agreement of density measurements on the different CT systems was very strong (; RMSE<6.2%). Our findings suggest that the Stratec pQCT scanner very accurately measures volumetric trabecular and total bone mineral densities at the distal radius while the assessment of cortical density is associated with considerable inaccuracies due to limited spatial resolution.

This paper investigates 3D image reconstruction from truncated cone-beam (CB) projections acquired with a helical vertex path. First, we show that a rigorous derivation of Grangeat's formula for truncated projections leads to a small additional term compared with previously published similar formulations. This correction term is called the boundary term. Next, this result is used to develop a CB filtered-backprojection (FBP) algorithm for truncated helical projections. This new algorithm only requires the CB projections to be measured within the region that is bounded, in the detector, by the projections of the upper and lower turns of the helix. Finally, simulations with mathematical phantoms demonstrate that: (i) the boundary term is necessary to obtain high-quality imaging of low-contrast structures and (ii) good image quality is obtained even with large values of the pitch of the helix.

A feasibility study of soft-tissue imaging based on x-ray wide-angle diffraction contrast has been performed at the medical beamline of the European Synchrotron Radiation Facility (ESRF). The technique employs computed-tomography algorithms to reconstruct from one data set the spatial distribution of several tissues differentiated by their diffraction properties. Radial diffraction profiles are measured in parallel projections from the sample and decomposed into material-selective weighting factors, which form the sinograms for the reconstructions. Attenuation effects - inherent in imaging techniques using scattered radiation - are efficiently corrected for by a ray-tracing method applied to the corresponding absorption image.

Images of 7 cm diameter samples composed of fat, bone and muscle were generated at 60 and 80 keV x-ray energy. The highest surface-absorbed dose was 24 mGy, but substantial contrast could still be obtained at 7 mGy, indicating potential applicability in medical imaging. The dominant noise contribution in the images stems from the detection system, pointing to a possible decrease in the surface-absorbed dose for an optimized system of more than a factor of 2.

Although electronic portal imaging devices (EPIDs) are efficient tools for radiation therapy verification, they only provide images of overlapped anatomic structures. We investigated using a fluorescent screen/CCD-based EPID, coupled with a novel multi-level scheme algebraic reconstruction technique (MLS-ART), for a feasibility study of portal computed tomography (CT) reconstructions. The CT images might be useful for radiation treatment planning and verification. We used an EPID, set it to work at the linear dynamic range and collimated 6 MV photons from a linear accelerator to a slit beam of 1 cm wide and 25 cm long. We performed scans under a total of monitor units (MUs) for several phantoms in which we varied the number of projections and MUs per projection. The reconstructed images demonstrated that using the new MLS-ART technique megavoltage portal CT with a total of 200 MUs can achieve a contrast detectibility of (object size mm) and a spatial resolution of 2.5 mm.

The temporal distribution of decay events recorded by a gamma camera in `list mode' differs from the Poisson distribution because of dead-time effects. We propose a new model for the dead-time behaviour of a gamma camera. The most important feature of our model is that the loss of events occurs in pairs or higher multiples due to the so-called `pile-up' effect. We analyse the consequences of pile-up for the temporal distribution of events recorded by a gamma camera. The probability distribution for the time intervals between events recorded by the camera is calculated from first principles.

We construct estimators for the parameter of the new distribution. We distinguish between the estimation of the total count rate and the estimation of a certain subset of the total count rate. Computer simulation confirms that our estimators are less influenced by dead-time effects than the standard estimator.

An electronic collimation technique is developed which utilizes the chi-square goodness-of-fit measure to filter scattered gammas incident upon a medical imaging detector. In this data mining technique, Compton kinematic expressions are used as the chi-square fitting templates for measured energy-deposition data involving multiple-interaction scatter sequences. Fit optimization is conducted using the Davidon variable metric minimization algorithm to simultaneously determine the best-fit gamma scatter angles and their associated uncertainties, with the uncertainty associated with the first scatter angle corresponding to the angular resolution precision for the source. The methodology requires no knowledge of materials and geometry. This pattern recognition application enhances the ability to select those gammas that will provide the best resolution for input to reconstruction software. Illustrative computational results are presented for a conceptual truncated-ellipsoid polystyrene position-sensitive fibre head-detector Monte Carlo model using a triple Compton scatter gamma sequence assessment for a Tc point source. A filtration rate of 94.3% is obtained, resulting in an estimated sensitivity approximately three orders of magnitude greater than a high-resolution mechanically collimated device. The technique improves the nominal single-scatter angular resolution by up to approximately 24 per cent as compared with the conventional analytic electronic collimation measure.

The widely applied single-interaction analytic expression characterizing the energy resolution component of the angular resolution precision for an electronically collimated point source is extended to include multiple-interaction Compton scatter sequences as well as sequences terminated by photoelectric absorption. The analytic formulation is developed using the statistical variance of the mean for components comprising composite, multivariate resolution precision estimators. It is demonstrated that enhanced resolution precision in the incident interaction scatter angle is attained when use is made of information from multiple interactions. An improvement in the resolution precision of up to approximately 40% is observed for triple Compton scatter. Comparison of the analytic estimates with Monte Carlo/chi-square results shows good agreement.

Mounting evidence indicates that scatter and attenuation are major confounds to objective diagnosis of brain disease by quantitative SPECT. There is considerable debate, however, as to the relative importance of scatter correction (SC) and attenuation correction (AC), and how they should be implemented.

The efficacy of SC and AC for Tc brain SPECT was evaluated using a two-compartment fully tissue-equivalent anthropomorphic head phantom. Four correction schemes were implemented: uniform broad-beam AC, non-uniform broad-beam AC, uniform SC+AC, and non-uniform SC+AC. SC was based on non-stationary deconvolution scatter subtraction, modified to incorporate a priori knowledge of either the head contour (uniform SC) or transmission map (non-uniform SC). The quantitative accuracy of the correction schemes was evaluated in terms of contrast recovery, relative quantification (cortical:cerebellar activity), uniformity ((coefficient of variation of 230 macro-voxels)), and bias (relative to a calibration scan).

Our results were: uniform broad-beam AC (the most popular correction): 71% contrast recovery, 112% relative quantification, 7.0% uniformity, +23% bias. Non-uniform broad-beam (soft tissue ) AC: 73%, 114%, 6.0%, +21%, respectively. Uniform SC+AC: 90%, 99%, 4.9%, +12%, respectively. Non-uniform SC+AC: 93%, 101%, 4.0%, +10%, respectively.

SC and AC achieved the best quantification; however, non-uniform corrections produce only small improvements over their uniform counterparts. SC+AC was found to be superior to AC; this advantage is distinct and consistent across all four quantification indices.

Cho et al have proposed a fast backprojection scheme for parallel beam geometries, the incremental algorithm, which performs backprojection on a ray-by-ray (beam-by-beam) basis as opposed to a pixel-by-pixel approach. We present an extension of this incremental, beamwise, backprojection algorithm to the case of volume reconstruction of data acquired by a cylindrical, multiring positron tomograph. Use is made of geometrical symmetries of the image volume. This method results in a twelve-fold reduction of execution time compared with a straightforward, voxel-driven implementation of the same interpolation equations.

A Monte Carlo model has been developed for optical coherence tomography (OCT). A geometrical optics implementation of the OCT probe with low-coherence interferometric detection was combined with three-dimensional stochastic Monte Carlo modelling of photon propagation in the homogeneous sample medium. Optical properties of the sample were selected to simulate intralipid and blood, representing moderately ( g = 0.7) and highly ( g = 0.99) anisotropic scattering respectively.

For shallow optical depths in simulated intralipid (<3 scattering mean free path (mfp) units), the number of detected backscattered photons followed the extinction-single-backscatter model, and OCT was found to detect only minimally scattered photons. Within this depth range the backscatter positions of detected photons corresponded well with the nominal focus position of the probe. For propagation to deeper positions in intralipid, localization of backscattering was quickly lost due to detection of stray photons, and the number of detected photons remained constant with increasing depth in the non-absorbing medium.

For strongly forward-directed scattering in simulated blood, the number of detected photons approached the extinction-single-backscatter model only for very shallow depths (<2 mfp units). However, backscattering positions for detected photons correlated well with the nominal focus position of the probe even for optical depths greater than 40 mfp units.

A Monte Carlo model has been developed for optical Doppler tomography (ODT) within the framework of a model for optical coherence tomography (OCT). A phantom situation represented by blood flowing in a horizontal m diameter vessel placed at m axial depth in 2% intralipid solution was implemented for the Monte Carlo simulation, and a similar configuration used for experimental ODT measurements in the laboratory.

Simulated depth profiles through the centre of the vessel of average Doppler frequency demonstrated an accuracy of 3-4% deviation in frequency values and position localization of flow borders, compared with true values.

Stochastic Doppler frequency noise was experimentally observed as a shadowing in regions underneath the vessel and also seen in simulated Doppler frequency depth profiles. By Monte Carlo simulation, this Doppler noise was shown to represent a nearly constant level over an investigated m interval of depth underneath the vessel. The noise level was essentially independent of the numerical aperture of the detector and angle between the flow velocity and the direction of observation, as long as this angle was larger than . Since this angle determines the magnitude of the Doppler frequency for backscattering from the flow region, this means that the signal-to-noise ratio between Doppler signal from the flow region to Doppler noise from regions underneath the flow is improved by decreasing the angle between the flow direction and direction of observation. Doppler noise values from Monte Carlo simulations were compared with values from statistical analysis.

We measured the optical properties of muscular tissue using several methods. Collimated transmission measurements of thin slabs showed spatial anisotropy of the scattering processes. Surface roughness of the sample disables the calculation of the extinction coefficient from these measurements. From angular intensity measurements we found a scattering asymmetry parameter g = 0.96. In fresh samples the optical diffusion constant D depends on the orientation with respect to the longitudinal direction of the muscular cells. From the D values we calculated perpendicular to the longitudinal direction as (at 543 nm), (at 594 nm) and (at 632 nm). The values for D which we measured from samples that were frozen and thawed did not show dependence on orientation. From spectral dependent reflectance measurements we found an oxygenation degree of 0.61 and a reduced scattering coefficient around 560 nm.

Time-domain potentials measured at 64 points on the surface of a large canine heart, considered comparable with those of a human heart, were used to calculate the electric fields and current densities within various organs of the human body. A heterogeneous volume conductor model of an adult male with a resolution of approximately and 30 segmented tissue types was used along with the admittance method and successive over-relaxation to calculate the voltage distribution throughout the torso and head as a function of time. From this time-domain voltage description, values of and were obtained, allowing for maximum values to be found within the given tissues of interest. Frequency analysis was then used to solve for and in the various organs, so that average, minimum and maximum values within specific bandwidths (0-40, 40-70 and 70-100 Hz) could be analysed. A comparison was made between the computed results and measured data from both EKG waveforms and isopotential surface maps for validation, with good agreement in both amplitude and shape between the computed and measured results. These computed endogenous fields were then compared with exogenous fields induced in the body from a 60 Hz high-voltage power line and a 60 Hz uniform magnetic field of 1 mT directed from the front to the back of the body. The high-voltage power line EMFs and 1 mT magnetic field were used as `bench' marks for comparison with several safety guidelines for power frequency (50/60 Hz) EMF exposures. The endogenous electric fields and current densities in most of the tissues (except for organs in close proximity to the heart, for example lungs, liver, etc) in the frequency band 40-70 Hz were found to be considerably smaller, between 5% and 10%, than those induced in the human body by the electric and magnetic fields generated by the 60 Hz sources described above.

We report findings from studies of water subjected to tension by pulsed dynamic stressing which may have significance in the context of assessing the safety of biomedical applications of low-frequency ultrasound.

When incident on bubbles near a free surface (or a flexible membrane), shockwaves generated by cavitation bubble collapse are found to lead to the production of liquid jets which are directed towards the surface. Tension pulses generated by reflection of these shockwaves at a free surface initiate, and propagate, the growth of further cavitation bubbles. Estimates of the tensile strength of water, , from measurements of the velocity of these pulses suggest that is far higher than has previously been reported for experiments involving the reflection of low-frequency ultrasound; and the estimate of reported here considerably exceeds the values of tension thought to be generated by ultrasound in many biomedical applications.

Monitoring the generation of cavitation is of great interest for diagnostic and therapeutic use of ultrasound in medicine, since cavitation is considered to play a major role in nonthermal ultrasound interactions with tissue. Important parameters are the number of cavitation events and the energy released during the bubble collapse. This energy is correlated to the maximum bubble radius which is related to the cavitation lifespan. The aim of this study was therefore to investigate the influence of the acoustic pressure amplitude and the pulse repetition frequency (PRF) in the field of a lithotripter (Lithostar, Siemens) on the number, size and lifespan of transient cavitation bubbles in water. We used scattered laser light recorded by a photodiode and stroboscopic photographs to monitor the cavitation activity. We found that PRF (range 0.5-5 Hz) had no influence on the cavitation bubble lifespan and size, whereas lifespan and size increased with the acoustic pressure amplitude. In contrast, the number of cavitation events strongly increased with PRF, whereas the pressure amplitude had no significant influence on the number of cavitation events. Thus, by varying the pressure amplitude and PRF, it might be possible to deliver a defined relative number of cavitations at a defined relative energy level in a defined volume. This seems to be relevant to further studies that address the biological effects of transient cavitation occurring in the fields of lithotripters.

The current induced in the outer circuit of a fast response ionization chamber exposed to pulsed radiation consists of two components, a fast one induced by free electrons and a slow one induced by ions. The fast electron component may be used for the representation of the shape of the ionizing pulse. In order to avoid interference with the slow ion current, the latter has to be removed from the signal. This is achieved by deriving a voltage course from the chamber signal which fits the shape of the ion component and subtracting this from the entire signal. The function of the electronic circuit used for this purpose is described. Some considerations about the time resolution of the chamber gas are to be found in the appendix.

A significant timer error has been found on a linear accelerator used for intra-operative radiation therapy. Since typical treatment doses range up to 20 Gy, calibrations that do not account for this effect can incur a large dosimetry error. The effect is dose rate dependent, but energy independent. For a dose rate of 900 cGy min, the error in dose calibration varies between 2 and 3% for an average of 2.5%. At dose rates of 600 and 300 cGy min, the error was 1.7% and 0.8% respectively. The average timer error at 900 cGy min was found to be monitor units. It is argued that even a small timer error has serious dosimetry consequences for the implementation of intensity modulated radiation therapy.

Conventional 3D dose calculations for stereotactic radiosurgery involve integration of individual static beams comprising a set of arcs. For iterative optimization of multiple isocentre treatment, which requires repetitive dose calculations at a large number of sample points, the conventional method is too slow. To overcome this problem spherically symmetric dose distributions are assumed. The authors describe a spherical dose model derived from a parametrized convolution of the collimator width and a dose spread kernel. The method is fast and easy to implement requiring just a single empirically derived value. Furthermore, the model is differentiable with respect to the parameters to be optimized. This property is useful when the optimization strategies rely on gradient information.

The characteristics of a prototype computer-assisted dynamic multileaf collimator (DMLC), specifically designed for small-field conformal radiotherapy, were evaluated at the Istituto Nazionale Tumori of Milan. The collimating device consists of two opposing banks of 16 pairs of 8 cm thick, 3.6 mm wide tungsten leaves and allows shaping of a radiation field up to a size of at the isocentre. The screening thickness of each leaf is 6.25 mm at the accelerator gantry isocentre. The leaves have a trapezoidal cross section and move along an arched path, thus providing a `double focused' collimation system. The DMLC was installed on the head of a Varian Clinac 2100C linear accelerator. Mechanical and dosimetric evaluations were performed to test the stability of the mechanical isocentre and to determine leaf leakage, penumbra width, accuracy of leaf positions and uniformity of leaf speed. Displacement of the mechanical isocentre was less than 1 mm at all gantry angles. Standard radiographic films exposed to 6 MV x-ray radiation were used for dosimetric evaluations. Leakage between leaves was less than 2.5%, and leakage through abutted leaves was less than 5.5%. The penumbra width between 20% and 80% isodose at different positions of leaf banks was 2.7 mm in the direction of the leaf motion and 3.1 mm along the side of the leaf with a standard deviation of 0.2 mm in both directions. Accuracy in the positioning of the leaf was 0.3 mm, whereas the maximum repositioning error was less than 0.2 mm. Finally, during movement of the leaves at the maximum speed of , the standard deviation of the leaf positioning error was 0.2 mm, proving an accurate uniformity of leaf speed.

The tongue and groove effect is an underdosing effect which can occur in certain applications of multileaf collimators. It results from the need to overlap adjacent leaves of a multileaf collimator in order to limit leakage between leaves. The applications in which the effect can occur are the abutment of fields where the beam edges are defined by the leaf edge and the production of intensity-modulated fields by dynamic collimation. The effect has been measured for the `worst case' when just two MLC fields are matched along leaf edges which have overlapping steps. Measurements of the dose have been made at and also at a more clinically relevant depth of 87 mm in Perspex for beam energies of 6 MV, 8 MV and 20 MV on two Philips SL series accelerators. Dose distributions were recorded on radiographic film which was subsequently digitized for analysis. The dose reduction of the tongue and groove effect was found to be 15-28% and spread over a width of 3.8 to 4.2 mm. This is somewhat shallower and wider than would be expected from a simple, idealized model of the effect which would predict a dose reduction of 80% over a width of 1 mm.

It has been pointed out by Dr Biao Chen, of the University of Rochester, that there is an error in the original version of above article. The error occurs in table 3, in which the parameters for a least-square curve fit are given for the linear attenuation coefficients of the phantom materials tested. Two of the coefficients (a₂ and a₃) were expressed with too few significant digits. Unfortunately, these coefficients are for the highest degree terms, and round-off errors can produce rather dramatic changes in the values calculated from the equation in table 3, especially at high energy. For example, using the original values for CIRS: Gland, {a₀, a₁, a₂, a₃}={23.84, -16.2, 3.48, -0.25}, the linear attenuation appears to be 0.697 which is 271% greater than the value determined from the basis material technique given in table 2 (0.188). Using the parameters in the table below which are expressed with greater precision, the linear attenuation is calculated to be 0.1886, yielding much better agreement (∼0.3%). In all cases, the linear attenuation coefficients calculated with the non-linear least-squares parameters from table 3 are within 4% of the coefficients determined in table 2.

The correct table 3 is presented in full below.

Table 3. Model equation and least-square curve fits of linear attenuation coefficients: µ(E) = exp[a₀ + a₁ (ln E) + a₂(ln E)² + a₃(ln E)³] (E = energy (kev), µ = linear attenuation (cm^-1)).

<PRE> a₀ a₁ a₂ a₃ RMI: Fat (AP6) 25.43 -18.52 4.269 -0.3341 1.26 RMI: Gland (MS11) 24.83 -16.97 3.677 -0.2707 1.20 </PRE>

CIRS: Fat 24.92 -18.16 4.190 -0.3285 1.20 CIRS: Gland 23.84 -16.20 3.482 -0.2541 1.21

Standard deviations on the estimates of the fitted parameters are less than 4%. The equation is valid only between E = 18 keV and E = 100 keV.

The data of the fourth column of table 3 (S(X) IAM), reported on pages 2556 and 2557, were erroneously scaled by a factor , with M the molecular mass of the compound. Therefore, the data of S(X) IAM for PMMA on page 2556 and the data of S(X) IAM for fat on page 2557 must be multiplied by a factor of 10.01 and 29.21 respectively.

Moreover, due to a mistake in Thompson expression evaluations, the data of Bradley et al (1989) reported in figure 1 had results enhanced beyond x = 2.5 mm^-1. After the proper computation, these discrepancies were removed. The results of molecular differential cross section and form factor for coherent scattering presented in the original version remain correct.

The authors' thanks are due to Dr Robert Leclair - from Carleton, Canada - and to Dr David Dance - from London, UK - for pointing out these errors which were due to the same computer program mistake in forming the final tabulations and graphic files.

A full list is published in January, April, July and October

1998 A practical and theoretical course in radiotherapy physics (Part A: Radiation dosimetry and treatment planning) 2-6 November, Sutton, Surrey, UK Contact: Dr Alan Nahum, The Joint Department of Physics, The Royal Marsden NHS Trust, Sutton, Surrey SM2 5PT, UK (Tel: +44 181 642 6011; E-mail: alan@icr.ac.uk)

Int. conf. on medical physics and ann. conf. on medical physics of association of medical physicists of India 6-9 November, New Delhi, India Contact: Dr M M Rehani, Medical Physics Unit, I R Cancer Hospital, AIIMS, New Delhi 110 029, India (Tel: +91 11 6594448; Fax: +91 11 6862663; E-mail: mrehani@medinst.ernet.in)

1999 A practical and theoretical course in radiotherapy physics (Part B: Brachytherapy, radiobiology and treatment machines) 1-5 March, Sutton, Surrey, UK Contact: Dr Alan Nahum, The Joint Department of Physics, The Royal Marsden NHS Trust, Sutton, Surrey SM2 5PT, UK (Tel: +44 181 642 6011; E-mail: alan@icr.ac.uk)

58th ann. mtg of the Japanese Radiological Society; 55th scientific assembly of the Japanese Society of Radiological Technology; Int. technical exhibition of medical imaging '99 6-8 April, Tokyo, Japan Contact: Secretariat of JMCP, Kitaotemachi Bldg, 1-7-6 Uchikanda, Chiyoda-ku, Tokyo 101 0047, Japan (Tel: 81 3 5281 0456; Fax: 81 3 5281 0457)

4th int. computer aided ergonomics and safety conf., CAES '99 19-21 May, Barcelona, Spain Contact: Mr Markku Leppänen, Secretary General CAES '99, Tampere University of Technology, Occupational Safety Engineering, PO Box 541, FIN-33101 Tampere, Finland (Fax: +358 3 365 2671; E-mail: mleppane@cc.tut.fi; http://www.caes99.org)

SRP int. symp. Achievements and challenges: Advancing radiation protection into the 21st century 14-18 June, Southport, UK Contact: SRP, Ramillies House, 1-9 Hills Place, London W1R 1AG, UK (Tel: +44 (0)171 287 4955, +44 (0)171 287 4906)

45th ann. scientific mtg of the Canadian Organization of Medical Physicists (COMP) 15-18 June, Sherbrooke, Quebec, Canada Contact: COMP Secretariat, 11328 88 Street, Edmonton, AB, Canada T5B 3P8 (E-mail: bmcgarry@compusmart.ab.ca)

AAPM summer school 1999 on digital imaging June 22-26 Contact: Mrs Andrea H DeGirolamo, Membership Manager, American Association of Physicists in Medicine, One Physics Ellipse, College Park, MD 20740 3846, USA (E-mail: andrea@aapm.org)

6th cong. of the International Society for Skin Imaging 4-6 July, London, UK Contact: Scientific and Administrative Secretariat, c/o HITEC (Hammersmith Hospital), Du Cane Road, London W12 0HS, UK (E-mail: hitec@rpms.ac.uk; Internet: http://www.icr.ac.uk/physics/conferences/issi.html)

AAPM 41st ann. mtg July 25-29, Nashville, TN, USA Contact: Mrs Andrea H DeGirolamo, Membership Manager, American Association of Physicists in Medicine, One Physics Ellipse, College Park, MD 20740 3846, USA (E-mail: andrea@aapm.org)