Table of contents

For pt.2 see ibid., vol.33, p.1249-60 (1988). Sololuminescence can readily be seen when aerated is insonated with continuous wave therapeutic ultrasound at room temperature but is not easily observed when short pulses of diagnostic ultrasound are used. In this work on ultrasound generator, operating in the region of 1 MHz and capable of producing pulses of different length and repetition rate, was used for insonation. The pulse repetition rate of the ultrasound was fixed at 1 kHz since this is characteristic of diagnostic machines, and a series of thresholds for sonoluminescence was obtained for two transducers, one therapeutic and one diagnostic, as the number of cycles in each pulse was varied. Sonoluminescence was observed for pulses of a few cycles, but the ultrasound intensity threshold for onset increased sharply with decreasing pulse length. Under all conditions tested, sonoluminescence was more readily sustained than initiated. At about 20 cycles per pulse, peak negative pressures of about 400 kPa initiated sonoluminescence. These conditions are well within the range of some regimens for Doppler ultrasound and not far removed from the diagnostic situation.

The interaction of ultrasonic waves with individual cells has been modelled on the basis that the cells can be represented by viscous liquid spheres surround by a viscoelastic shell (the membrane) immersed in a viscous fluid. The computational model includes thermal waves and requires 22 input parameters. Many of the parameters are not available in the literature and a detailed discussion is given on the procedures by which the values used in the model calculations were chosen. In spite of the arbitrariness of the choice of many of the parameter values, the computations show surprisingly good agreement with experimental measurements of ultrasonic attenuation in animal cell suspensions. The model has been used here to investigate different aspects of the interaction of ultrasound with the cells. It is found that the membrane is important only between 0.5 and 30 MHz and contributes less than 15% to the attenuation. Absorption is shown to be an important feature to include, while the scattering contribution to the attenuation is less than 1% at 3 MHz. The thermal effects are important at frequencies below 1 MHz and contribute some 65% to the attenuation at 100 MHz.

The characteristics of 0.6 cm³ thimble-type Baldwin-Farmer (BF 2571) ionisation chambers for adsorbed dose determinations in-phantom at mammography installations are investigated. The most important aspects for in-phantom dosimetry in mammography concern the conversion from air kerma to absorbed dose in mammary gland tissue, the energy dependence of the sensitivity of the ionisation chamber and the displacement correction factor for measurements in-phantom. Due to the considerable uncertainties in the elemental composition of the mammary glands the conversion from air kerma to absorbed dose in the mammary gland tissue has an uncertainty of the order of +or-20%. The air kerma calibration factor of the BF-ionisation chamber is about 10% larger at mammography radiation qualities than at 300 kV X-rays or ¹³⁷Cs gamma rays. For depths in excess of about 15 mm a displacement correction factor of 0.69+or-0.06 is derived for measurements with the BF2571 chamber inside polymethylmethacrylate phantoms irradiated with 30 kV X-rays (first HVL: 0.29 mm Al).

Verification of electron beam treatment-planning algorithms in the presence of heterogeneities can be very difficult. Using controlled geometries to minimise physical uncertainties in geometric alignment and composition, a large number of measurements were made to test the performance of a 2D and 3D electron pencil beam algorithm. A Therados RFA-3 beam-scanning system interfaced to a microcomputer was used to measure the dose distributions. The geometric arrangement consisted of single and double rods 1 cm in diameter situated just below the surface of a unit density phantom. The electron densities (relative to water) of the rods ranged from 2.12 (aluminium) to 1.29 (soft bone analogue), and their length could be varied between 1 cm and 10 cm. Measured isodose distributions beyond the inhomogeneities were compared with those predicted theoretically. Calculations were performed on a VAX-11/780 using 2D and 3D implementations of the Hogstrom electron pencil beam algorithm. The authors report on the nature of this 3D implementation and assess the magnitude of discrepancies between calculation and measurement for 10 MeV and 18 MeV electron beams, and for the variety of phantom compositions and geometries identified above.

A programme of work on soft errors of 64 K and 256 K DRAMs has begun in these laboratories, studying the error rates and modelling them. A feasibility study is being carried out of neutron dosimetry employing soft errors in 256 K DRAM. Thermal neutrons from a neutron howitzer and albedo neutrons from ²⁵²Cf neutrons scattered off a water phantom produced 'soft errors', through the action of alpha particles and tritons, generated by the (n, alpha ) reaction in ⁶LiF converters. The energy dependence of the soft error rate (SER) was studied by degrading 5.5. MeV alpha particles after passage through plastic absorbers. The dependence of SER on the operating bias of DRAM was also studied. For the study of thermal neutron sensitivity, the thickness of the ⁶LiF converter was optimised. A calibration curve of soft errors against dose equivalent was plotted, the dose equivalent having been calculated from thermal neutron flux measurements. A dose equivalent of 25 mu Sv corresponds to a soft error of about 1. There are potential advantages of a DRAM as a digital neutron dosemeter.

Methods for calculating multiple point T₁ relaxation images are described and compared. A robust line fitting method is presented and its relevance to high resolution relaxation imaging discussed. Selective removal of data points is demonstrated to reduce systematic errors in linear models. A non-linear least squares iterative method is also developed and implemented. The results include simulated data, clinical images and a phantom study. Several of the methods are substantial improvements on existing linear techniques. The increased quality and consistency of the calculated images make them particularly appropriate for automated pattern recognition.

One drawback of factor analysis of dynamic structures (FADS) is that only empirical methods are available to determine the number of independent factors which must be extracted from the dynamic sequence. The aim of this work is to propose a rigorous method to determine the number of statistically significant independent factors contained in a dynamic sequence of scintigraphic images. The proposed method uses the correspondence analysis (CA) of the dynamic series and the test of the likelihood ratio (LR). It is based on the fact that CA allows the minimisation of a distance between the data and the model close to the least-squares distance weighted by the value of the model. Since the noise has a Poisson distribution, the likelihood can be considered as maximised and the LR test can be used. The proposed method has been validated on a numerical phantom, and has been illustrated using actual scinigraphic data. The results show that the number of significant factors determined by the proposed test is in good agreement with the number of factors which have a physiological meaning.

Presents a general technique for the determination of eddy current distributions within an irregularly shaped conducting sample in the uniform field of an NMR RF coil. Also presented is a general expression for the sample-induced coil resistance. The technique is applied specifically to a conducting cylindrical solid. Unlike previous descriptions of cylindrical samples in solenoidal coils where the induction is parallel to the axis of the cylinder and eddy current streamlines are circular, the author treats the case where the induction is perpendicular to the cylindrical axis and where consequently eddy current streamlines take on an irregular shape.

A method has been developed for accurate correction of red blood cell count for coincidence in aperture-impedance electronic blood cell counters. It is based on extrapolation of the slope of regression of the counts which are obtained with sequential dilutions of the samples.

The authors have numerically modelled the thermal effects of yellow (577 or 578 nm) light illumination on the ectatic blood vessels of port wine stains. They investigated the effect of the laser treatment parameters of light irradiance and illumination time on the extent of coagulation (coagulation thickness) of the endothelial cells of the ectatic vessels. They assumed that this coagulation is dependent on heating the cells to a critical temperature (coagulation temperature). They iteratively adjusted the treatment parameters so that the model vessels has a maximum temperature that did not exceed the boiling point of blood. Given the likely range of variation of coagulation temperature, coagulation thickness and vessel size, coagulation temperature was found to have the greatest effect on the treatment parameters. It is concluded that, in order to achieve coagulation with these short laser pulses, the choice of irradiance is critical.

A method for measuring images retrieved from reflection holograms is presented. Holographic images of a well defined three-dimensional object are measured using the Reflex Metrograph. The holograms were reconstructed using both coherent and incoherent radiation. These measurements allow conclusions to be drawn about the degree of deformation or magnification inherent in reflection holograms. To illustrate this technique a reflection hologram of an orthodontic cast of teeth and gums has been made and measurements on it have been performed. This provides sufficient accuracy for orthodontic applications.

The performance characteristics of conventional and K-edge filters have been studied for the entire field of general diagnostic radiology. The problem of optimising the conflicting needs from patient dose, image contrast and exposure time was managed primarily by computer simulation. In comparison to conventional filters like iron or copper no significant advantages can be obtained with K-edge filters in practice except for the special circumstances of automatic fluoroscopy. The optimal choice of conventional filters is discussed. The different roles of backscattered radiation for skin dose and integral dose are demonstrated, and the necessity of correcting for increased backscatter following additional filtration is stressed. The potential for dose reductions by adding filtration has been found to be substantially smaller than the figures mostly reported. Finally some methodological problems of studies of this kind are pointed out.

The significance of pH as an indicator of the intracellular environment is well established and ³¹P NMR spectroscopy provides a non-invasive means of measuring pH (Gadian et al., 1979). The derivation of a pH calibration curve to be used with a whole body system for in vivo studies of pH is described. Different standard reference samples are considered, both internal and external, and the advantages of using an internal reference are established. It is concluded that PCr is the most suitable reference sample for this study, because it may be used dissolved in the solutions and because of its very narrow peak. Bo inhomogeneity and local susceptibility effects are thus minimised, and the experimental arrangement more closely resembles the in vivo situation.

The measurement of Pb in bone by X-ray fluorescence using ¹⁰⁹Cd gamma -ray to excite K-shell X-rays is considered. Improvements to the basic system performance were made by changing the design of the source collimator and by including data from the Pb K beta _1,3 photopeaks, having used pile-up rejection to reduce background in their region of the spectrum. The summed variances for eight sample measurements are given: two each of four plaster-of-Paris phantoms embedded in wax.

A ⁶Li-filled neutron detector, placed at the tumour site, which integrates over the slowing down spectrum of neutrons in the same manner as the ¹⁰B contained in the tumour, appears to provide a practical method to measure the equivalent thermal neutron fluence in boron neutron capture therapy. If the boron concentration of the blood, brain and tumour are known, then, with the detector measurement, the boron absorbed dose can be calculated for these tissues. A detector containing 0.01 mg of ⁶Li results in the production of an amount of ¹⁸F which produces a detected activity of 1.5*10^-9 counts per second per n/cm². This detector can measure the equivalent thermal neutron fluence with an accuracy of 1% for a single treatment fraction of 10¹² cm^-2, corresponding to one-tenth of the total fluence to the tumour. The depression of the boron absorbed dose by a 0.25 cm radius spherical detector with a ⁶Li mass of 0.01 mg is approximately 0.15%, immediately adjacent to the detector.

These Monte Carlo studies have shown that, if the distal end of range of the proton beam is to be tailored tightly to the target volume, sophisticated treatment planning programs, beam delivery and patient immobilisation techniques have to be developed and details, such as having the compensating bolus fit snugly to the patient's surface, have to be considered. Although the magnitudes of the effects reported here are specific to the components and beam quality in this Harvard Cyclotron Laboratory proton beam line, the principles are general and apply to all heavy charged particle beams. Thus, in the design of a proton beam delivery system, every effort should be made to allow the beam-modifying devices to get very close to the patient or to devise some other means to minimise the separation of patient and bolus.

The authors have developed an L-band ESR system with a loop-gap resonator and using this system they have successfully performed in vivo ESR measurements of nitroxides administered to rats. This system is quite suitable for direct observation of free radicals in living animals even though exogenous ones are described here.

The energy resolution index of NaI detectors for use in medical gamma -cameras has been calculated for a range of peak energies and has been shown to be independent of energy. It is concluded that measurement of this index provides a simple accurate method to measure energy resolution which can easily be incorporated into a regular quality assurance program.